repo
stringlengths
2
99
file
stringlengths
13
225
code
stringlengths
0
18.3M
file_length
int64
0
18.3M
avg_line_length
float64
0
1.36M
max_line_length
int64
0
4.26M
extension_type
stringclasses
1 value
toulbar2
toulbar2-master/misc/script/wcsp2lp-support.py
#!/usr/bin/env python # -*- coding: utf-8 -*- import copy import os import sys import itertools import numpy assert len(sys.argv) == 3, "Please specify INPUT and OUTPUT filenames." # does the WCSP have one or more constant terms has_constant_term = False # classe pour écriture à largeur de texte contrôlée class WidthFile(file): maxcol = 80 def __init__(self, *x, **k): file.__init__(self, *x, **k) self.col = 0 def write(self, x): lines = x.splitlines() #print "outputting", lines if (self.col + len(lines[0])) >= 80: file.write(self, "\n") self.col = 0 map(lambda x: file.write(self, x + '\n'), lines[:-1]) file.write(self, lines[-1]) if len(lines) > 1: self.col = len(lines[-1]) else: self.col += len(lines[-1]) # le nom des variables pour l'encodage des valeurs des domaines. les # variables booléennes restent booléennes mais il faut compter les # littéraux négatifs par ailleurs (format lp ne gère pas (1-x)) def domain_var(n, v): return " d%i_%i " % (n, v) def mdomain_var(coeff, n, v): if (v == 1) and (domains[n] == 2): return "%+i d%i_0 " % (-coeff,n) else : return "%+i d%i_%i " % (coeff, n, v) # le nom des variables pour l'encodage des autres tuples def tuple_var(tvar,tval): tvarval = map(lambda var,val: (var,val),tvar,tval) # normalize tuple st = sorted(tvarval, key=lambda x: x[0]) name = "t" for x in st: name = name + ("_%i_%i" % x) return name #le produit cartésien des séquences (stockées dans une séquence vlist). def product(vlist): return apply(itertools.product,vlist) #enumerate all "tuples" on tvar (for var, if it appears in tvar, a #single value val is used instead of thh full domain) generating the #set of support tuples. def enum_tuples(tvar, var, val): return product(map(lambda ovar: [val] if (var == ovar) else xrange(domains[ovar]), tvar)) # reading numbers def read_num_vec(toks): return map(int, toks) def read_int_tok(tok_iter): return int(tok_iter(1)[0]) # lire une définition de cost function. The cost table is a tuple based dictionary def read_fun(tok_iter): n_var = read_int_tok(tok_iter) vars_ = read_num_vec(tok_iter(n_var)) defcost = read_int_tok(tok_iter) if defcost == -1: defcost = tok_iter(1)[0] n_spec = 1 else: n_spec = read_int_tok(tok_iter) tvo = sorted(map(lambda var,val: (var,val),vars_,range(len(vars_))),key=lambda x: x[0]) ovars = tuple(x[0] for x in tvo) varorder = tuple(x[1] for x in tvo) specs = dict() for i in xrange(n_spec): if defcost!='knapsackp': tc = read_num_vec(tok_iter(n_var + 1)) if isinstance(defcost, basestring): specs = tc else: specs[tuple(tc[i] for i in varorder)] = tc[-1] else : Weight=[] Weight.append(read_int_tok(tok_iter)) for j in range(n_var): nbval=read_int_tok(tok_iter) Weight.append(nbval) for k in range(nbval): Weight.append(read_int_tok(tok_iter)) Weight.append(read_int_tok(tok_iter)) specs=Weight if isinstance(defcost, basestring): return vars_, defcost, specs else: return ovars, defcost, specs # parcourir une cost function table def iter_fun(vars_, defcost, specs): vardom = [xrange(domains[v]) for v in vars_] for t in itertools.product(*vardom): if t in specs: yield t, specs[t] else: yield t, defcost # parcourir une cost function table en évitant les tuples d'un coût # donné si possible (defcost) def iter_funavoid(vars_, defcost, specs, avoid): if (defcost == avoid): for t in specs: yield t, specs[t] else: vardom = [xrange(domains[v]) for v in vars_] for t in itertools.product(*vardom): if t in specs: yield t, specs[t] else: yield t, defcost # ------------- MAIN --------------------- def token_iter(filename): for l in open(filename).xreadlines(): for stok in l.strip().split(" "): for ttok in stok.strip().split("\t"): if ttok: yield ttok tokens = token_iter(sys.argv[1]) def next_tokens(n): return [tokens.next() for i in xrange(n)] #line_iter = open(sys.argv[1]).xreadlines() output = WidthFile(sys.argv[2], 'w') print "File %s opened" % sys.argv[1] # reading parameters #params = (line_iter.next().strip().split(" ")) name = tokens.next() n_var, max_domain_size, n_fun, upper_bound = read_num_vec(next_tokens(4)) domains = read_num_vec(next_tokens(n_var)) n_fun = int(n_fun) ub = int(upper_bound) print >> output, "Minimize" all_fun = [read_fun(next_tokens) for i in xrange(n_fun)] print "\nCost functions read." # Output the criteria. Do not integrate zero or "infinite" cost # components here. Zero is useless, "infinite" will be handled as # linear constraints negative_litterals = 0 for vars_, defcost, specs in all_fun: if isinstance(defcost, basestring): continue n_vars = len(vars_) if (n_vars == 0): has_constant_term = 1 output.write(' +%i t ' % defcost) else: for t, cost in iter_funavoid(vars_, defcost, specs,0): if cost == 0 or cost >= ub: continue if n_vars == 1: output.write(mdomain_var(cost,vars_[0], t[0])) if (t[0] == 1 and domains[vars_[0]] <= 2): negative_litterals = negative_litterals + cost else: output.write(' +%i %s ' % (cost, tuple_var(vars_, t))) if negative_litterals: has_constant_term = 1 output.write(" +%i t" % negative_litterals) print "Criteria generated." output.write("\n\nSubject to\n\n") # Set of tuple vars that need not be used ub_tuplevars = set() # Hard constraints: for every value with cost >=ub, we forbid it # explicitely. Tuples variables are just removed. for vars_, defcost, specs in all_fun: if isinstance(defcost, basestring): if defcost == 'knapsack': for i, v in enumerate(vars_): if int(specs[i+1]>=0): output.write('+%i d%i_0' % (specs[i+1], v)) else: output.write('%i d%i_0' % (specs[i+1], v)) output.write(' <= %i\n\n' % (sum(specs) - 2*specs[0],)) elif defcost== 'knapsackp': last=1 tot=0 for i, v in enumerate(vars_): nbval=specs[last] if(domains[v]==2): val1=0 val0=0 for j in range(nbval): if int(specs[last+1+2*j])==1: val1=int(specs[last+2+2*j]) tot+=val1 else: val0=int(specs[last+2+2*j]) if val0-val1>=0: output.write('+%i d%i_0' % (val0-val1, v)) else: output.write('%i d%i_0' % (val0-val1, v)) else: for j in range(nbval): if int(specs[last+2+2*j]>=0): output.write('+%i d%i_%i' % (specs[last+2+2*j], v,specs[last+1+2*j])) else: output.write('%i d%i_%i' % (specs[last+2+2*j], v,specs[last+1+2*j])) last=last+nbval*2+1 output.write(' >= %i\n\n' % (specs[0]-tot,)) continue n_vars = len(vars_) for t, cost in iter_funavoid(vars_, defcost, specs,0): if cost < ub: continue if n_vars == 1: output.write('%s = %i\n\n' % (mdomain_var(1,vars_[0], t[0]), -(domains[vars_[0]] == 2 and t[0] == 1))) else: ub_tuplevars.add(tuple_var(vars_, t)) print "Hard constraints generated." # Direct encoding. Exactly one value constraint. Boolean variables are not included here. for i, dom in enumerate(domains): if (dom > 2) : map(lambda v: output.write(mdomain_var(1, i, v)), xrange(dom)) output.write(" = 1\n\n") if (dom == 1) : output.write("%s = 1\n\n" % domain_var(i,0)) print "Domain constraints generated." # marginal consistency: one value selected iff one associated tuple selected. # if several functions have the same cost, we need only to do it once scopes = set(vars_ for vars_, defcost, specs in (f for f in all_fun if len(f[0]) >= 2 and not isinstance(f[1], basestring))) print "%i different scopes detected." % len(scopes) for vars_ in scopes: for va in vars_: for a in xrange(domains[va]): map(lambda b: output.write("+1 %s " % tuple_var(vars_, b)) if tuple_var(vars_, b) not in ub_tuplevars else 0, enum_tuples(vars_,va,a)) output.write(" %s " % mdomain_var(-1, va, a)) output.write("= %i\n\n" % (domains[va] == 2 and a == 1)) print "Marginal consistency constraints generated." if has_constant_term : output.write("t = 1\n") print "Tuple bounds generated." output.write("\n\nBinary\n\n") # indicate 0/1 variables (direct encoding). for i, dom in enumerate(domains): if (dom > 2) : map(lambda v: output.write("%s " % domain_var(i, v)), xrange(dom)) else: output.write("%s " % domain_var(i, 0)) for vars_, defcost, specs in (f for f in all_fun if len(f[0]) >= 2): if isinstance(defcost, basestring): continue map(lambda b: output.write("%s " % tuple_var(vars_, b)) if tuple_var(vars_, b) not in ub_tuplevars else 0, enum_tuples(vars_,-1,-1)) if has_constant_term : output.write("t") output.write("\n\nEnd") print "Domain binaries generated." print "Finished."
9,939
31.913907
146
py
toulbar2
toulbar2-master/validation/bilevel/bilevel_mibs2.py
import pytoulbar2 as tb2 cfn = tb2.CFN(ubinit = 1000, verbose = 0) cfn.NoPreprocessing() cfn.Option.btdMode = 1 cfn.Option.hbfs = 0 # create restricted leader problem cfn.Option.bilevel = 1 cfn.AddVariable('x0',range(2)) cfn.AddVariable('x1',range(2)) cfn.AddVariable('x2',range(2)) cfn.AddLinearConstraint([7,5,2],['x0','x1','x2'],'<=',9) cfn.Option.initialLbBLP = cfn.Option.initialLbBLP + [cfn.CFN.wcsp.getLb()] cfn.CFN.wcsp.setLb(0) cfn.Option.negCostBLP = cfn.Option.negCostBLP + [cfn.CFN.wcsp.getNegativeLb()] cfn.CFN.wcsp.decreaseLb(-cfn.CFN.wcsp.getNegativeLb()) # create follower problem cfn.Option.bilevel = 2 cfn.AddVariable('C0',range(4)) cfn.AddVariable('C1',range(3)) cfn.AddVariable('C2',range(5)) cfn.AddFunction(['C0','C1','C2'], [(0 if (11 * v0 + 4 * v1 + 6 * v2 <= 50) else 1000000) for v0 in range(4) for v1 in range(3) for v2 in range(5)]) cfn.AddFunction(['x0','C0'], [(0 if v0 <= 3*(1-x0) else 1000000) for x0 in range(2) for v0 in range(4)]) cfn.AddFunction(['x1','C1'], [(0 if v1 <= 2*(1-x1) else 1000000) for x1 in range(2) for v1 in range(3)]) cfn.AddFunction(['x2','C2'], [(0 if v2 <= 4*(1-x2) else 1000000) for x2 in range(2) for v2 in range(5)]) cfn.AddFunction(['C0'], [-8 * v0 for v0 in range(4)]) cfn.AddFunction(['C1'], [-12 * v1 for v1 in range(3)]) cfn.AddFunction(['C2'], [-3 * v2 for v2 in range(5)]) cfn.Option.initialLbBLP = cfn.Option.initialLbBLP + [cfn.CFN.wcsp.getLb()] cfn.CFN.wcsp.setLb(0) cfn.Option.negCostBLP = cfn.Option.negCostBLP + [cfn.CFN.wcsp.getNegativeLb()] cfn.CFN.wcsp.decreaseLb(-cfn.CFN.wcsp.getNegativeLb()) # create negative form of follower problem cfn.Option.bilevel = 3 cfn.AddVariable('C0neg',range(4)) cfn.AddVariable('C1neg',range(3)) cfn.AddVariable('C2neg',range(5)) cfn.AddFunction(['C0neg','C1neg','C2neg'], [(8 * v0 + 12 * v1 + 3 * v2 if (11 * v0 + 4 * v1 + 6 * v2 <= 50) else 1000000) for v0 in range(4) for v1 in range(3) for v2 in range(5)]) cfn.AddFunction(['x0','C0neg'], [(0 if v0 <= 3*(1-x0) else 1000000) for x0 in range(2) for v0 in range(4)]) cfn.AddFunction(['x1','C1neg'], [(0 if v1 <= 2*(1-x1) else 1000000) for x1 in range(2) for v1 in range(3)]) cfn.AddFunction(['x2','C2neg'], [(0 if v2 <= 4*(1-x2) else 1000000) for x2 in range(2) for v2 in range(5)]) cfn.Option.initialLbBLP = cfn.Option.initialLbBLP + [cfn.CFN.wcsp.getLb()] cfn.CFN.wcsp.setLb(0) cfn.Option.negCostBLP = cfn.Option.negCostBLP + [cfn.CFN.wcsp.getNegativeLb()] cfn.CFN.wcsp.decreaseLb(-cfn.CFN.wcsp.getNegativeLb()) cfn.Option.bilevel = 4 cfn.Option.decimalPointBLP = [0,0,0] cfn.Option.costMultiplierBLP = [1.,1.,-1.] cfn.Option.initialUbBLP = [tb2.tb2.MAX_COST,tb2.tb2.MAX_COST,tb2.tb2.MAX_COST] print(cfn.Option.negCostBLP) print(cfn.Option.initialLbBLP) cfn.CFN.wcsp.setLb(cfn.Option.initialLbBLP[0] + cfn.Option.initialLbBLP[2]) cfn.CFN.wcsp.decreaseLb(cfn.Option.negCostBLP[0] + cfn.Option.negCostBLP[2]) cfn.Option.setVarOrder('0 -1 0 1 2\n1 0 0 1 2\n2 0 0 1 2 3 4 5\n3 0 0 1 2 6 7 8\n') cfn.Solve(showSolutions=3)
2,994
43.044118
181
py
toulbar2
toulbar2-master/validation/bilevel/bilevel_mibs0.py
import pytoulbar2 as tb2 cfn = tb2.CFN(ubinit = 1000, verbose = 0) cfn.NoPreprocessing() cfn.Option.btdMode = 1 cfn.Option.hbfs = 0 # create restricted leader problem cfn.Option.bilevel = 1 cfn.AddVariable('x',range(9)) cfn.AddFunction(['x'],[-vx for vx in range(9)]) cfn.Option.initialLbBLP = cfn.Option.initialLbBLP + [cfn.CFN.wcsp.getLb()] cfn.CFN.wcsp.setLb(0) cfn.Option.negCostBLP = cfn.Option.negCostBLP + [cfn.CFN.wcsp.getNegativeLb()] cfn.CFN.wcsp.decreaseLb(-cfn.CFN.wcsp.getNegativeLb()) # create follower problem cfn.Option.bilevel = 2 cfn.AddVariable('y',range(6)) cfn.AddFunction(['x','y'], [(10 * vy if ((-25 * vx + 20 * vy <= 30) and (1 * vx + 2 * vy <= 10) and (2 * vx - 1 * vy <= 15) and (2 * vx + 10 * vy >= 15)) else 1000000) for vx in range(9) for vy in range(6)]) cfn.Option.initialLbBLP = cfn.Option.initialLbBLP + [cfn.CFN.wcsp.getLb()] cfn.CFN.wcsp.setLb(0) cfn.Option.negCostBLP = cfn.Option.negCostBLP + [cfn.CFN.wcsp.getNegativeLb()] cfn.CFN.wcsp.decreaseLb(-cfn.CFN.wcsp.getNegativeLb()) # create negative form of follower problem cfn.Option.bilevel = 3 cfn.AddVariable('yneg',range(6)) cfn.AddFunction(['x','yneg'], [(-10 * vy if ((-25 * vx + 20 * vy <= 30) and (1 * vx + 2 * vy <= 10) and (2 * vx - 1 * vy <= 15) and (2 * vx + 10 * vy >= 15)) else 1000000) for vx in range(9) for vy in range(6)]) cfn.Option.initialLbBLP = cfn.Option.initialLbBLP + [cfn.CFN.wcsp.getLb()] cfn.CFN.wcsp.setLb(0) cfn.Option.negCostBLP = cfn.Option.negCostBLP + [cfn.CFN.wcsp.getNegativeLb()] cfn.CFN.wcsp.decreaseLb(-cfn.CFN.wcsp.getNegativeLb()) cfn.Option.bilevel = 4 cfn.Option.decimalPointBLP = [0,0,0] cfn.Option.costMultiplierBLP = [1.,1.,-1.] cfn.Option.initialUbBLP = [tb2.tb2.MAX_COST,tb2.tb2.MAX_COST,tb2.tb2.MAX_COST] print(cfn.Option.negCostBLP) print(cfn.Option.initialLbBLP) cfn.CFN.wcsp.setLb(cfn.Option.initialLbBLP[0] + cfn.Option.initialLbBLP[2]) cfn.CFN.wcsp.decreaseLb(cfn.Option.negCostBLP[0] + cfn.Option.negCostBLP[2]) cfn.Option.setVarOrder('0 -1 0\n1 0 0\n2 0 0 1\n3 0 0 2\n') cfn.Solve(showSolutions=3)
2,055
37.792453
211
py
toulbar2
toulbar2-master/validation/bilevel/bilevel_mibs1.py
import pytoulbar2 as tb2 cfn = tb2.CFN(ubinit = 1000, verbose = 0) cfn.NoPreprocessing() cfn.Option.btdMode = 1 cfn.Option.hbfs = 0 # create restricted leader problem cfn.Option.bilevel = 1 cfn.AddVariable('C0',range(11)) cfn.AddFunction(['C0'],[-v for v in range(11)]) cfn.Option.initialLbBLP = cfn.Option.initialLbBLP + [cfn.CFN.wcsp.getLb()] cfn.CFN.wcsp.setLb(0) cfn.Option.negCostBLP = cfn.Option.negCostBLP + [cfn.CFN.wcsp.getNegativeLb()] cfn.CFN.wcsp.decreaseLb(-cfn.CFN.wcsp.getNegativeLb()) # create follower problem cfn.Option.bilevel = 2 cfn.AddVariable('C1',range(6)) cfn.AddFunction(['C0','C1'], [(7 * v1 if ((-3 * v0 + 2 * v1 <= 12) and (1 * v0 + 2 * v1 <= 20) and (2 * v0 - 1 * v1 <= 7) and (-2 * v0 + 4 * v1 <= 16)) else 1000000) for v0 in range(11) for v1 in range(6)]) # all cost functions and constraints on the same scope must be merged cfn.Option.initialLbBLP = cfn.Option.initialLbBLP + [cfn.CFN.wcsp.getLb()] cfn.CFN.wcsp.setLb(0) cfn.Option.negCostBLP = cfn.Option.negCostBLP + [cfn.CFN.wcsp.getNegativeLb()] cfn.CFN.wcsp.decreaseLb(-cfn.CFN.wcsp.getNegativeLb()) # create negative form of follower problem cfn.Option.bilevel = 3 cfn.AddVariable('C1neg',range(6)) cfn.AddFunction(['C0','C1neg'], [(-7 * v1 if ((-3 * v0 + 2 * v1 <= 12) and (1 * v0 + 2 * v1 <= 20) and (2 * v0 - 1 * v1 <= 7) and (-2 * v0 + 4 * v1 <= 16)) else 1000000) for v0 in range(11) for v1 in range(6)]) # all cost functions and constraints on the same scope must be merged cfn.Option.initialLbBLP = cfn.Option.initialLbBLP + [cfn.CFN.wcsp.getLb()] cfn.CFN.wcsp.setLb(0) cfn.Option.negCostBLP = cfn.Option.negCostBLP + [cfn.CFN.wcsp.getNegativeLb()] cfn.CFN.wcsp.decreaseLb(-cfn.CFN.wcsp.getNegativeLb()) cfn.Option.bilevel = 4 cfn.Option.decimalPointBLP = [0,0,0] cfn.Option.costMultiplierBLP = [1.,1.,-1.] cfn.Option.initialUbBLP = [tb2.tb2.MAX_COST,tb2.tb2.MAX_COST,tb2.tb2.MAX_COST] print(cfn.Option.negCostBLP) print(cfn.Option.initialLbBLP) cfn.CFN.wcsp.setLb(cfn.Option.initialLbBLP[0] + cfn.Option.initialLbBLP[2]) cfn.CFN.wcsp.decreaseLb(cfn.Option.negCostBLP[0] + cfn.Option.negCostBLP[2]) cfn.Option.setVarOrder('0 -1 0\n1 0 0\n2 0 0 1\n3 0 0 2\n') cfn.Solve(showSolutions=3)
2,196
40.45283
280
py
toulbar2
toulbar2-master/validation/default/weightedcspconstraint.py
VERBOSE=0 PROBLEM1="../../validation/bilevel/bilevel1b.cfn" PROBLEM2="../../validation/bilevel/bilevel2.cfn" LB=11 UB=20 import pytoulbar2 as tb2 cfn1 = tb2.CFN(verbose = VERBOSE) cfn1.Read(PROBLEM1) cfn2 = tb2.CFN(verbose = VERBOSE) cfn2.Read(PROBLEM2) cfn1.AddWeightedCSPConstraint(cfn2, LB, UB, True) cfn1.Solve(showSolutions=3, allSolutions=100) # find 18 solutions
375
19.888889
65
py
toulbar2
toulbar2-master/validation/default/clique.py
import pytoulbar2 as tb2 m = tb2.CFN(1, verbose=0) w=m.AddVariable('w', range(2)) x=m.AddVariable('x', range(2)) y=m.AddVariable('y', range(2)) z=m.AddVariable('z', range(2)) m.CFN.wcsp.postCliqueConstraint([x,y,z,w],'1 1 1 1 1 1 1 1 1') for u in [w,x,y,z]: for v in [w,x,y,z]: if u<v: m.AddFunction([u,v],[0, 0, 0, 1000]) m.Solve(showSolutions=1, allSolutions=16)
377
18.894737
62
py
toulbar2
toulbar2-master/validation/default/sregular.py
import pytoulbar2 as tb2 m = tb2.CFN(12, verbose=0) v1=m.AddVariable('v1', range(2)) v2=m.AddVariable('v2', range(2)) v3=m.AddVariable('v3', range(2)) v4=m.AddVariable('v4', range(2)) m.AddFunction([v1], [2, 0]) m.AddFunction([v4], [0, 3]) m.CFN.wcsp.postWRegular([v1,v2,v3,v4],'var','DAG', 12, 2, [tb2.tb2.WeightedObjInt(0,0)], [tb2.tb2.WeightedObjInt(0,0), tb2.tb2.WeightedObjInt(1,0)], [tb2.tb2.DFATransition(0,0,0,0),tb2.tb2.DFATransition (0,1,1,0),tb2.tb2.DFATransition(1,1,1,0)]) m.Solve(showSolutions=1, allSolutions=16)
534
28.722222
202
py
ns_lattice
ns_lattice-master/ns_lattice/setup.py
''' Use of this source code is governed by a MIT-style license that can be found in the LICENSE file. Created on Jan 27, 2017 @author: Niels Lubbes https://python-packaging.readthedocs.io/en/latest/minimal.html https://pypi.python.org/pypi?%3Aaction=list_classifiers ''' from setuptools import setup setup( name = 'ns_lattice', version = '4', description = 'Algorithms for computing in Neron-Severi lattice', classifiers = [ 'Development Status :: 3 - Alpha', 'License :: OSI Approved :: MIT License', 'Programming Language :: Python :: 2', 'Programming Language :: Python :: 3', 'Topic :: Scientific/Engineering :: Mathematics', ], keywords = 'Neron-Severi-lattice', url = 'http://github.com/niels-lubbes/ns_lattice', author = 'Niels Lubbes', license = 'MIT', package_dir = {'ns_lattice': 'src/ns_lattice'}, packages = ['ns_lattice'], package_data = {'ns_lattice': ['ns_tools.sobj']}, # include_package_data = True, install_requires = ['linear_series'], test_suite = 'nose.collector', tests_require = ['nose'], entry_points = { 'console_scripts': ['run-lattice=ns_lattice.__main__:main'], }, zip_safe = False )
1,250
29.512195
97
py
ns_lattice
ns_lattice-master/ns_lattice/src/tests/test_class_ns_tools.py
''' Use of this source code is governed by a MIT-style license that can be found in the LICENSE file. Created on Feb 13, 2017 @author: Niels Lubbes ''' from ns_lattice.class_ns_tools import NSTools class TestClassNSTools: def test__p( self ): NSTools.filter( None ) assert NSTools.p( 'Hello world!' ) != None NSTools.filter( ['another_class.py'] ) assert NSTools.p( 'No output since called from another class.' ) == None NSTools.filter_unset() assert NSTools.p( 'Filter is disabled so output this string.' ) != None NSTools.filter_reset() assert NSTools.p( 'Filter is enabled again so do not output.' ) == None NSTools.filter( ['test_class_ns_tools.py'] ) assert NSTools.p( 'Only output if called from this class' ) != None def test__tool_dct( self ): nt = NSTools() nt2 = NSTools() # watch out to not use the default file name # otherwise it might take long to load the data test_fname = 'test_tools' key = 'test__tool_dct' dct = nt.get_tool_dct( fname = test_fname ) dct[key] = True nt.save_tool_dct( fname = test_fname ) assert key in nt.get_tool_dct( fname = test_fname ) assert key in nt2.get_tool_dct( fname = test_fname ) nt.set_enable_tool_dct( False ) assert key not in nt.get_tool_dct( fname = test_fname ) assert key not in nt2.get_tool_dct( fname = test_fname ) nt.set_enable_tool_dct( True ) assert key in nt.get_tool_dct( fname = test_fname ) assert key in nt2.get_tool_dct( fname = test_fname )
1,654
28.035088
97
py
ns_lattice
ns_lattice-master/ns_lattice/src/tests/test_class_div.py
''' Use of this source code is governed by a MIT-style license that can be found in the LICENSE file. Created on Feb 8, 2017 @author: Niels Lubbes ''' from ns_lattice.sage_interface import sage_ZZ from ns_lattice.sage_interface import sage_matrix from ns_lattice.class_div import Div from ns_lattice.class_ns_tools import NSTools class TestClassDiv: def test__new( self ): assert Div.new( '3e0+e1+5e5-e6' ).e_lst == [3, 1, 0, 0, 0, 5, -1, 0, 0] assert Div.new( 'e1-e2' ).e_lst == [0, 1, -1, 0, 0, 0, 0, 0, 0] assert Div.new( '-e1+e2' ).e_lst == [0, -1, 1, 0, 0, 0, 0, 0, 0] assert Div.new( '-3e0' ).e_lst == [-3, 0, 0, 0, 0, 0, 0, 0, 0] assert Div.new( '-e3' ).e_lst == [0, 0, 0, -1, 0, 0, 0, 0, 0] assert Div.new( '12' ).e_lst == [0, 1, -1, 0, 0, 0, 0, 0, 0] assert Div.new( '-12' ).e_lst == [0, -1, 1, 0, 0, 0, 0, 0, 0] assert Div.new( '1245' ).e_lst == [1, 0, -1, 0, -1, -1, 0, 0, 0] assert Div.new( '214' ).e_lst == [2, 0, -1, -1, 0, -1, -1, -1, -1] assert Div.new( '306' ).e_lst == [3, -1, -1, -1, -1, -1, -2, -1, -1] assert Div.new( '-308' ).e_lst == [-3, 1, 1, 1, 1, 1, 1, 1, 2] def test__get_label__True( self ): assert Div( [3, 1, 0, 0, 0, 5, -1, 0, 0] ).get_label( True ) == '3e0+e1+5e5-e6' assert Div( [0, 1, -1, 0, 0, 0, 0, 0, 0] ).get_label( True ) == '12' assert Div( [0, -1, 1, 0, 0, 0, 0, 0, 0] ).get_label( True ) == '-12' assert Div( [-3, 0, 0, 0, 0, 0, 0, 0, 0] ).get_label( True ) == '-3e0' assert Div( [0, 0, 0, -1, 0, 0, 0, 0, 0] ).get_label( True ) == '-e3' assert Div( [0, 1, -1, 0, 0, 0, 0, 0, 0] ).get_label( True ) == '12' assert Div( [0, -1, 1, 0, 0, 0, 0, 0, 0] ).get_label( True ) == '-12' assert Div( [1, 0, -1, 0, -1, -1, 0, 0, 0] ).get_label( True ) == '1245' assert Div( [2, 0, -1, -1, 0, -1, -1, -1, -1] ).get_label( True ) == '214' assert Div( [3, -1, -1, -1, -1, -1, -2, -1, -1] ).get_label( True ) == '306' assert Div( [-3, 1, 1, 1, 1, 1, 1, 1, 2] ).get_label( True ) == '-308' def test__get_abbr_label( self ): assert Div.new( 'e1' ).get_abbr_label() == 'e1' assert Div.new( 'e1-e2' ).get_abbr_label() == 'e12' assert Div.new( '2e0-e1-e2-e4-e5' ).get_abbr_label() == '2e1245' assert Div.new( 'e0-e1' ).get_abbr_label() == '1e1' def test__lt( self ): assert Div.new( '1123' ) < Div.new( '1124' ) assert Div.new( '12' ) < Div.new( '1123' ) assert Div.new( '12' ) < Div.new( '13' ) assert Div.new( '12' ) < Div.new( '34' ) def test__get_basis_change( self ): B = sage_matrix( sage_ZZ, [( 1, -1, 0, 0, 0, 0 ), ( 1, 0, -1, 0, 0, 0 ), ( 1, -1, -1, 0, 0, 0 ), ( 0, 0, 0, 1, 0, 0 ), ( 0, 0, 0, 0, 1, 0 ), ( 0, 0, 0, 0, 0, 1 )] ) # (-2)-classes assert Div.new( '1123', 6 ).get_basis_change( B ).get_label() == 'e2-e3' assert Div.new( '1345', 6 ).get_basis_change( B ).get_label() == 'e0+e1-e2-e3-e4-e5' assert Div.new( '12', 6 ).get_basis_change( B ).get_label() == '-e0+e1' assert Div.new( '13', 6 ).get_basis_change( B ).get_label() == 'e1-e2-e3' assert Div.new( '23', 6 ).get_basis_change( B ).get_label() == 'e0-e2-e3' # (-1)-classes assert Div.new( 'e1', 6 ).get_basis_change( B ).get_label() == 'e1-e2' assert Div.new( 'e2', 6 ).get_basis_change( B ).get_label() == 'e0-e2' assert Div.new( 'e3', 6 ).get_basis_change( B ).get_label() == 'e3' assert Div.new( '2e0-e1-e2-e3-e4-e5', 6 ).get_basis_change( B ).get_label() == 'e0+e1-e3-e4-e5' # classes of conical families assert Div.new( 'e0-e1', 6 ).get_basis_change( B ).get_label() == 'e0' assert Div.new( 'e0-e2', 6 ).get_basis_change( B ).get_label() == 'e1' assert Div.new( 'e0-e3', 6 ).get_basis_change( B ).get_label() == 'e0+e1-e2-e3' assert Div.new( '2e0-e1-e3-e4-e5', 6 ).get_basis_change( B ).get_label() == '2e0+e1-e2-e3-e4-e5' assert Div.new( '2e0-e2-e3-e4-e5', 6 ).get_basis_change( B ).get_label() == 'e0+2e1-e2-e3-e4-e5' def test__is_positive( self ): assert Div.new( 'e0-e1', 6 ).is_positive() assert Div.new( 'e1-e2', 6 ).is_positive() assert not Div.new( '-e2+e1', 6 ).is_positive() assert not Div.new( '-e0+e1+e2', 6 ).is_positive() if __name__ == '__main__': NSTools.filter( None ) TestClassDiv().test__is_positive() pass
4,686
41.225225
104
py
ns_lattice
ns_lattice-master/ns_lattice/src/tests/test_class_dp_lattice.py
''' Use of this source code is governed by a MIT-style license that can be found in the LICENSE file. Created on Nov 7, 2017 @author: Niels Lubbes ''' from ns_lattice.sage_interface import sage_QQ from ns_lattice.sage_interface import sage_identity_matrix from ns_lattice.sage_interface import sage_matrix from ns_lattice.class_div import Div from ns_lattice.dp_involutions import complete_basis from ns_lattice.sage_interface import sage_vector from ns_lattice.div_in_lattice import get_divs from ns_lattice.div_in_lattice import get_ak from ns_lattice.class_eta import ETA from ns_lattice.sage_interface import sage_ZZ from ns_lattice.dp_root_bases import get_dynkin_type from ns_lattice.class_ns_tools import NSTools from ns_lattice.class_dp_lattice import DPLattice class TestClassDPLattice(): def test__eq( self ): NSTools.set_enable_tool_dct( False ) Md_lst = [] M = sage_identity_matrix( sage_QQ, 4 ) dpl23 = DPLattice( [Div.new( '23', 4 )], Md_lst, M ) dpl1123 = DPLattice( [Div.new( '1123', 4 )], Md_lst, M ) dpl12 = DPLattice( [Div.new( '12', 4 )], Md_lst, M ) assert dpl23 != dpl1123 assert dpl23 == dpl12 NSTools.set_enable_tool_dct( True ) def test__get_marked_Mtype( self ): NSTools.set_enable_tool_dct( False ) # (2A1, 4A1) Neron-Severi lattice of ring torus rank = 6 d_lst = [ 'e2-e4', 'e3-e5', 'e0-e1-e2-e4', 'e0-e1-e3-e5'] Md_lst = ['e4-e5', 'e0-e1-e2-e3'] M = [( 2, 1, 1, 1, 0, 0 ), ( -1, 0, -1, -1, 0, 0 ), ( -1, -1, 0, -1, 0, 0 ), ( -1, -1, -1, 0, 0, 0 ), ( 0, 0, 0, 0, 0, 1 ), ( 0, 0, 0, 0, 1, 0 )] d_lst = [ Div.new( d, rank ) for d in d_lst ] Md_lst = [ Div.new( Md, rank ) for Md in Md_lst ] M = sage_matrix( M ) dpl = DPLattice( d_lst, Md_lst, M ) print( dpl.get_marked_Mtype() ) print( dpl.Mtype ) assert dpl.get_marked_Mtype() == "2A1'" NSTools.set_enable_tool_dct( True ) def test__get_bas_lst__rank_3( self ): NSTools.set_enable_tool_dct( False ) bas_lst = DPLattice.get_bas_lst( 3 ) assert len( bas_lst ) == 2 for bas in bas_lst: print( bas ) print( len( bas_lst ) ) NSTools.set_enable_tool_dct( True ) def test__get_bas_lst__rank_4( self ): NSTools.set_enable_tool_dct( False ) bas_lst = DPLattice.get_bas_lst( 4 ) for bas in bas_lst: print( bas ) print( len( bas_lst ) ) assert len( bas_lst ) == 6 type_lst = [] for bas in bas_lst: type_lst += [( bas.Mtype, bas.type )] print( type_lst ) assert str( type_lst ) == "[('A0', 'A0'), ('A0', 'A1'), ('A0', 'A1'), ('A0', '2A1'), ('A0', 'A2'), ('A0', 'A1+A2')]" NSTools.set_enable_tool_dct( True ) def test__get_inv_lst__rank_4( self ): NSTools.set_enable_tool_dct( False ) rank = 4 inv_lst = DPLattice.get_inv_lst( rank ) print( len( inv_lst ) ) for inv in inv_lst: inv.set_attributes( 8 ) type_lst = [] for inv in inv_lst: type_lst += [( inv.Mtype, inv.type )] print( type_lst[-1] ) assert len( inv_lst ) == 4 assert str( type_lst ) == "[('A0', 'A0'), ('A1', 'A0'), ('A1', 'A0'), ('2A1', 'A0')]" NSTools.set_enable_tool_dct( True ) def test__get_cls_slow__rank_3( self ): NSTools.set_enable_tool_dct( False ) rank = 3 dpl_lst = DPLattice.get_cls_slow( rank ) for dpl in dpl_lst: dpl.set_attributes( 8 ) type_lst = [] for dpl in dpl_lst: type_lst += [( dpl.Mtype, dpl.type )] print( type_lst[-1] ) print( type_lst ) assert str( type_lst ) == "[('A0', 'A0'), ('A0', 'A1'), ('A1', 'A0')]" NSTools.set_enable_tool_dct( True ) def test__get_cls_slow__rank_4( self ): NSTools.set_enable_tool_dct( False ) rank = 4 dpl_lst = DPLattice.get_cls_slow( rank ) for dpl in dpl_lst: dpl.set_attributes( 8 ) type_lst = [] for dpl in dpl_lst: type_lst += [( dpl.Mtype, dpl.type )] print( type_lst[-1] ) print( type_lst ) assert str( type_lst ) == "[('A0', 'A0'), ('A0', 'A1'), ('A0', 'A1'), ('A0', '2A1'), ('A0', 'A2'), ('A0', 'A1+A2'), ('A1', 'A0'), ('A1', 'A1'), ('A1', 'A0'), ('A1', 'A1'), ('A1', 'A2'), ('2A1', 'A0')]" NSTools.set_enable_tool_dct( True ) def test__get_num_types( self ): NSTools.set_enable_tool_dct( False ) bas_lst = DPLattice.get_bas_lst( 4 ) inv_lst = DPLattice.get_inv_lst( 4 ) bas = bas_lst[1] inv = inv_lst[-1] assert inv.Mtype == '2A1' assert bas.type == 'A1' assert DPLattice.get_num_types( inv, bas, bas_lst ) == 0 bas = bas_lst[1] inv = inv_lst[2] assert inv.Mtype == 'A1' assert bas.type == 'A1' assert DPLattice.get_num_types( inv, bas, bas_lst ) == -1 NSTools.set_enable_tool_dct( True ) def test__get_part_roots( self ): NSTools.set_enable_tool_dct( False ) inv_lst = DPLattice.get_inv_lst( 4 ) inv = inv_lst[1] assert inv.Mtype == 'A1' s_lst, q_lst = DPLattice.get_part_roots( inv ) assert len( s_lst ) == 1 assert q_lst == [] NSTools.set_enable_tool_dct( True ) def test__seek_bases( self ): NSTools.set_enable_tool_dct( False ) bas = DPLattice.get_bas_lst( 4 )[-1] assert bas.type == 'A1+A2' inv = DPLattice.get_inv_lst( 4 )[0] assert inv.Mtype == 'A0' r_lst = get_divs( get_ak( bas.get_rank() ), 0, -2, True ) dpl_lst = DPLattice.seek_bases( inv, bas.d_lst, r_lst ) for dpl in dpl_lst: dpl.set_attributes() print( dpl.Mtype, dpl.type, dpl.d_lst ) assert len( dpl_lst ) == 1 NSTools.set_enable_tool_dct( True ) def test__get_cls__rank_3( self ): NSTools.set_enable_tool_dct( False ) dpl_lst = DPLattice.get_cls( 3 ) type_lst = [] for dpl in dpl_lst: type_lst += [( dpl.Mtype, dpl.type )] print( type_lst ) assert str( type_lst ) == "[('A0', 'A0'), ('A0', 'A1'), ('A1', 'A0')]" NSTools.set_enable_tool_dct( True ) def test__import_cls( self ): NSTools.set_enable_tool_dct( False ) dpl_lst = DPLattice.get_cls( 3 ) type_lst = [( dpl.Mtype, dpl.type ) for dpl in dpl_lst ] assert str( type_lst ) == "[('A0', 'A0'), ('A0', 'A1'), ('A1', 'A0')]" inv = DPLattice.get_inv_lst( 4 )[1] assert inv.Mtype == 'A1' out_lst = DPLattice.import_cls( dpl_lst, inv ) assert len( out_lst ) == 1 assert out_lst[0].get_rank() == 4 assert out_lst[0].Mtype == 'A1' assert out_lst[0].type == 'A0' NSTools.set_enable_tool_dct( True ) def test__get_cls__rank_4( self ): NSTools.set_enable_tool_dct( False ) dpl_lst = DPLattice.get_cls( 4 ) type_lst = [] for dpl in dpl_lst: type_lst += [( dpl.Mtype, dpl.type )] print( dpl.get_marked_Mtype(), dpl.type ) print( type_lst ) assert str( type_lst ) == "[('A0', 'A0'), ('A0', 'A1'), ('A0', 'A1'), ('A0', '2A1'), ('A0', 'A2'), ('A0', 'A1+A2'), ('A1', 'A0'), ('A1', 'A1'), ('A1', 'A0'), ('A1', 'A1'), ('A1', 'A2'), ('2A1', 'A0')]" NSTools.set_enable_tool_dct( True ) def test__get_real_type( self ): NSTools.set_enable_tool_dct( False ) dpl_lst = DPLattice.get_cls_slow( 4 ) type_lst = [] for dpl in dpl_lst: type_lst += [( dpl.get_marked_Mtype(), dpl.get_real_type() )] out = '' for type in type_lst: print( type[0] + ', ' + type[1] ) out += type[0] + ',' + type[1] + '; ' print( out ) assert out.strip() == "A0,A0; A0,{A1}; A0,{A1}; A0,2{A1}; A0,{A2}; A0,{A1}+{A2}; A1,A0; A1,{A1}; A1',A0; A1',{A1}; A1',{A2}; 2A1,A0;" NSTools.set_enable_tool_dct( True ) def test__get_SG( self ): NSTools.set_enable_tool_dct( False ) dpl_lst = DPLattice.get_cls( 4 ) out_lst = [] for dpl in dpl_lst: SG, SG_data = dpl.get_SG() out_lst += [[ dpl.Mtype, dpl.get_real_type()] + SG_data] for out in out_lst: print( out ) print( out_lst ) assert str( out_lst ) == "[['A0', 'A0', 3, 0, [0], [], False, False, True, True], ['A0', '{A1}', 2, 0, [0], [], False, False, True, True], ['A0', '{A1}', 3, 0, [0], [], False, False, True, True], ['A0', '2{A1}', 2, 0, [0], [], False, False, True, True], ['A0', '{A2}', 1, 0, [0], [], True, True, True, True], ['A0', '{A1}+{A2}', 1, 0, [0], [], True, True, True, True], ['A1', 'A0', 1, 0, [0], [], True, True, True, True], ['A1', '{A1}', 1, 0, [0], [], True, True, True, True], ['A1', 'A0', 3, 0, [0], [], False, False, True, True], ['A1', '{A1}', 2, 0, [0], [], False, False, True, True], ['A1', '{A2}', 1, 0, [0], [], True, True, True, True], ['2A1', 'A0', 1, 0, [0], [], True, True, True, True]]" NSTools.set_enable_tool_dct( True ) def test__are_root_bases( self ): NSTools.set_enable_tool_dct( False ) bas_lst = DPLattice.get_bas_lst( 4 ) for bas in bas_lst: if bas.d_lst == []: continue mat = complete_basis( bas.d_lst ) r_lst = get_divs( get_ak( bas.get_rank() ), 0, -2, True ) print( bas.type, bas.d_lst, 10 * '=' ) for r in r_lst: vec = ~mat * sage_vector( r.e_lst ) print( r.e_lst, vec, r, list( mat ) ) in_span = set( vec[len( bas.d_lst ):] ) == {0} zz_coef = set( [elt in sage_ZZ for elt in vec ] ) == {True} pos_coef = set( [elt >= 0 for elt in vec] ) == {True} if in_span and zz_coef: assert pos_coef NSTools.set_enable_tool_dct( True ) if __name__ == '__main__': NSTools.filter( None ) # NSTools.filter( ['class_dp_lattice.py', 'class_eta.py'] ) # TestClassDPLattice().test__eq() # TestClassDPLattice().test__get_marked_Mtype() # TestClassDPLattice().test__get_bas_lst__rank_3() # TestClassDPLattice().test__get_bas_lst__rank_4() # TestClassDPLattice().test__get_inv_lst__rank_4() # TestClassDPLattice().test__get_cls_slow__rank_3() # TestClassDPLattice().test__get_cls_slow__rank_4() # TestClassDPLattice().test__get_num_types() # TestClassDPLattice().test__get_part_roots() # TestClassDPLattice().test__seek_bases() # TestClassDPLattice().test__import_cls() # TestClassDPLattice().test__get_cls__rank_3() # TestClassDPLattice().test__get_cls__rank_4() # TestClassDPLattice().test__get_real_type() TestClassDPLattice().test__get_SG() # TestClassDPLattice().test__are_root_bases() pass
11,122
31.054755
706
py
ns_lattice
ns_lattice-master/ns_lattice/src/tests/test_div_in_lattice.py
''' Use of this source code is governed by a MIT-style license that can be found in the LICENSE file. Created on Feb 8, 2017 @author: Niels Lubbes ''' from ns_lattice.class_ns_tools import NSTools from ns_lattice.class_div import Div from ns_lattice.div_in_lattice import get_divs from ns_lattice.div_in_lattice import get_indecomp_divs from ns_lattice.div_in_lattice import get_ak class TestDivInLattice: def test__get_divs_2_2( self ): NSTools.set_enable_tool_dct( False ) d = Div.new( '2e0-e1-e2' ) dc = 2 cc = 2 c_lst = get_divs( d, dc, cc, True ) assert [c.get_label() for c in c_lst ] == [ '2e0-e1-e2' ] NSTools.set_enable_tool_dct( True ) def test__get_divs__minus_1_classes__rank_4( self ): NSTools.set_enable_tool_dct( False ) chk_lst = ['e1', 'e0-e1-e2'] out_lst = [] for div in get_divs( get_ak( 4 ), 1, -1, False ): out_lst += [ div.get_label() ] assert out_lst == chk_lst NSTools.set_enable_tool_dct( True ) def test__get_divs__minus_1_classes__rank_5( self ): NSTools.set_enable_tool_dct( False ) chk_lst = ['e1', 'e2', 'e3', 'e4', 'e0-e1-e2', 'e0-e1-e3', 'e0-e2-e3', 'e0-e1-e4', 'e0-e2-e4', 'e0-e3-e4'] out_lst = [] for div in get_divs( get_ak( 5 ), 1, -1, True ): out_lst += [ div.get_label() ] assert out_lst == chk_lst NSTools.set_enable_tool_dct( True ) def test__get_divs__minus_1_classes__rank_9( self ): NSTools.set_enable_tool_dct( False ) chk_lst = [ 'e1', 'e0-e1-e2', '2e0-e1-e2-e3-e4-e5', '3e0-2e1-e2-e3-e4-e5-e6-e7', '4e0-2e1-2e2-2e3-e4-e5-e6-e7-e8', '5e0-2e1-2e2-2e3-2e4-2e5-2e6-e7-e8', '6e0-3e1-2e2-2e3-2e4-2e5-2e6-2e7-2e8' ] out_lst = [] for div in get_divs( get_ak( 9 ), 1, -1, False ): out_lst += [ div.get_label() ] assert out_lst == chk_lst NSTools.set_enable_tool_dct( True ) def test__get_divs__minus_2_classes__rank_5__perm_true( self ): NSTools.set_enable_tool_dct( False ) chk_lst = [12, 23, 13, 34, 24, 14, 1123, 1124, 1134, 1234] out_lst = [] for div in get_divs( get_ak( 5 ), 0, -2, True ): out_lst += [ int( div.get_label( True ) ) ] print( out_lst ) assert out_lst == chk_lst NSTools.set_enable_tool_dct( True ) def test__get_divs__roman_surface( self ): NSTools.set_enable_tool_dct( False ) h = Div.new( '4e0-e1-e2-e3-e4-e5-e6-e7-e8' ) out_lst = get_divs( h, 2, -2, False ) out_lst += get_divs( h, 2, -1, False ) print( out_lst ) assert str( out_lst ) == '[2e0-e1-e2-e3-e4-e5-e6, e0-e1-e2]' NSTools.set_enable_tool_dct( True ) def test__get_divs__fam_classes__rank_6__perm_false( self ): NSTools.set_enable_tool_dct( False ) chk_lst = ['e0-e1', '2e0-e1-e2-e3-e4'] out_lst = [] for div in get_divs( get_ak( 6 ), 2, 0, False ): out_lst += [ div.get_label() ] assert out_lst == chk_lst NSTools.set_enable_tool_dct( True ) def test__get_divs__fam_classes__rank_6__perm_true( self ): NSTools.set_enable_tool_dct( False ) chk_lst = ['e0-e1', 'e0-e2', 'e0-e3', 'e0-e4', 'e0-e5', '2e0-e1-e2-e3-e4', '2e0-e1-e2-e3-e5', '2e0-e1-e2-e4-e5', '2e0-e1-e3-e4-e5', '2e0-e2-e3-e4-e5'] out_lst = [] for div in get_divs( get_ak( 6 ), 2, 0, True ): out_lst += [ div.get_label() ] print( out_lst ) assert out_lst == chk_lst NSTools.set_enable_tool_dct( True ) def test__get_indecomp_divs( self ): NSTools.set_enable_tool_dct( False ) c_lst = ['e0-e1', 'e0-e2', 'e0-e3', 'e0-e4', 'e0-e5', '2e0-e1-e2-e3-e4', '2e0-e1-e2-e3-e5', '2e0-e1-e2-e4-e5', '2e0-e1-e3-e4-e5', '2e0-e2-e3-e4-e5'] c_lst = [ Div.new( c ) for c in c_lst ] d_lst = [ 12, 1123 ] d_lst = [ Div.new( str( d ) ) for d in d_lst ] chk_lst = ['e0-e1', 'e0-e3', 'e0-e4', 'e0-e5', '2e0-e1-e2-e4-e5', '2e0-e1-e3-e4-e5'] out_lst = [] for div in get_indecomp_divs( c_lst, d_lst ): out_lst += [ div.get_label() ] print( out_lst ) assert out_lst == chk_lst NSTools.set_enable_tool_dct( True ) if __name__ == '__main__': NSTools.filter( None ) # TestDivInLattice().test__get_divs__fam_classes__rank_6__perm_true() # TestDivInLattice().test__get_divs__minus_2_classes__rank_5__perm_true() # TestDivInLattice().test__get_divs__minus_1_classes__rank_9() # TestDivInLattice().test__get_divs__roman_surface()
5,065
32.773333
97
py
ns_lattice
ns_lattice-master/ns_lattice/src/tests/test_class_eta.py
''' Use of this source code is governed by a MIT-style license that can be found in the LICENSE file. Created on Jan 27, 2018 @author: Niels Lubbes ''' from ns_lattice.class_eta import ETA from ns_lattice.class_ns_tools import NSTools class TestClassETA(): def test__update( self ): eta = ETA( 10, 2 ) for i in range( 10 ): assert eta.counter == i eta.update( '*test*', 3 ) assert eta.counter == i + 1 if __name__ == '__main__': NSTools.filter( ['class_eta.py'] ) NSTools.filter( None ) TestClassETA().test__update()
592
21.807692
97
py
ns_lattice
ns_lattice-master/ns_lattice/src/tests/test_dp_root_basis.py
''' Use of this source code is governed by a MIT-style license that can be found in the LICENSE file. Created on Feb 13, 2017 @author: Niels Lubbes ''' from ns_lattice.sage_interface import sage_QQ from ns_lattice.sage_interface import sage_identity_matrix from ns_lattice.sage_interface import sage_Graph from ns_lattice.class_ns_tools import NSTools from ns_lattice.class_div import Div from ns_lattice.dp_root_bases import is_root_basis from ns_lattice.dp_root_bases import get_graph from ns_lattice.dp_root_bases import get_ext_graph from ns_lattice.dp_root_bases import get_dynkin_type from ns_lattice.dp_root_bases import convert_type from ns_lattice.dp_root_bases import get_root_bases_orbit class TestDPRootBasis(): def test__is_root_basis( self ): assert is_root_basis( [] ) bas_lst = [1123 ] assert is_root_basis( [Div.new( str( bas ), 4 ) for bas in bas_lst] ) bas_lst = [1123, 23 ] assert is_root_basis( [Div.new( str( bas ), 4 ) for bas in bas_lst] ) bas_lst = [1123, 1123 ] assert not is_root_basis( [Div.new( str( bas ), 4 ) for bas in bas_lst] ) bas_lst = [12, -23 ] assert not is_root_basis( [Div.new( str( bas ), 4 ) for bas in bas_lst] ) def test__get_graph( self ): bas_lst = [12, 23, 34 ] d_lst = [Div.new( str( bas ), 5 ) for bas in bas_lst] G = get_graph( d_lst ) test_G = sage_Graph() test_G.add_vertices( [0, 1, 2] ) test_G.add_edge( 0, 1, 1 ) test_G.add_edge( 1, 2, 1 ) assert G == test_G def test__get_ext_graph( self ): NSTools.set_enable_tool_dct( False ) # # example for Neron-Severi lattice of sextic weak del Pezzo surface # The A1 root sub-systems [23] and [1123] are not equivalent. # We use as invariant a graph. # M = sage_identity_matrix( sage_QQ, 4 ) # real structure is the identity e_lst = [ 'e1', 'e0-e1-e2', 'e2', 'e0-e2-e3', 'e3', 'e0-e1-e3' ] # (-1)-classes d_lst1 = [Div.new( s, 4 ) for s in e_lst + ['23'] ] G1 = get_ext_graph( d_lst1, M ) d_lst2 = [Div.new( s, 4 ) for s in e_lst + ['1123'] ] G2 = get_ext_graph( d_lst2, M ) assert not G1.is_isomorphic( G2, edge_labels = True ) NSTools.set_enable_tool_dct( True ) def test__get_dynkin_type( self ): NSTools.set_enable_tool_dct( False ) bas_lst = [12, 23, 34 ] d_lst = [Div.new( str( bas ), 5 ) for bas in bas_lst] print( d_lst ) assert get_dynkin_type( d_lst ) == 'A3' NSTools.set_enable_tool_dct( True ) def test__convert_type( self ): NSTools.set_enable_tool_dct( False ) assert convert_type( '2A1+D4' ) == ['A1', 'A1', 'D4'] assert convert_type( '2A1+A2+A3' ) == ['A1', 'A1', 'A2', 'A3'] assert convert_type( 'A0+2A1+3A1+D4+A0' ) == 5 * ['A1'] + ['D4'] NSTools.set_enable_tool_dct( True ) def test__get_root_bases_orbit__rank_3( self ): NSTools.set_enable_tool_dct( False ) d_lst = [12] d_lst = [Div.new( str( d ), 3 ) for d in d_lst] d_lst_lst = get_root_bases_orbit( d_lst, False ) print( d_lst_lst ) assert str( d_lst_lst ) == '[[e1-e2], [-e1+e2]]' d_lst_lst = get_root_bases_orbit( d_lst, True ) print( d_lst_lst ) assert str( d_lst_lst ) == '[[e1-e2]]' NSTools.set_enable_tool_dct( True ) def test__get_root_bases_orbit__rank_4( self ): NSTools.set_enable_tool_dct( False ) d_lst = [12] d_lst = [Div.new( str( d ), 4 ) for d in d_lst] d_lst_lst = get_root_bases_orbit( d_lst, False ) print( d_lst_lst ) assert str( d_lst_lst ) == '[[e1-e2], [-e1+e2], [e1-e3], [-e2+e3], [-e1+e3], [e2-e3]]' d_lst_lst = get_root_bases_orbit( d_lst, True ) print( d_lst_lst ) assert str( d_lst_lst ) == '[[e1-e2], [e1-e3], [e2-e3]]' NSTools.set_enable_tool_dct( True ) if __name__ == '__main__': NSTools.filter( None ) # TestDPRootBasis().test__get_ext_graph() # TestDPRootBasis().test__get_root_bases_orbit__rank_3() # TestDPRootBasis().test__get_root_bases_orbit__rank_4() TestDPRootBasis().test__convert_type()
4,288
31.007463
97
py
ns_lattice
ns_lattice-master/ns_lattice/src/tests/test_ns_basis.py
''' Use of this source code is governed by a MIT-style license that can be found in the LICENSE file. Created on Feb 9, 2017 @author: Niels Lubbes ''' import sys from ns_lattice.sage_interface import sage_identity_matrix from ns_lattice.sage_interface import sage_matrix from ns_lattice.sage_interface import sage_ZZ from ns_lattice.sage_interface import sage_QQ from ns_lattice.sage_interface import sage_register_unpickle_override from ns_lattice.class_div import Div from ns_lattice.class_ns_tools import NSTools from ns_lattice.class_dp_lattice import DPLattice from ns_lattice.div_in_lattice import get_divs from ns_lattice.div_in_lattice import get_ak from ns_lattice.ns_basis import get_bases_lst from ns_lattice.ns_basis import get_webs from ns_lattice.ns_basis import contains_perm from ns_lattice.ns_basis import triples class TestNSBasis( object ): def test__get_basis_lst__rank_4__False( self ): NSTools.set_enable_tool_dct( False ) rank = 4 # construct DPLattice d_lst = [] Md_lst = [] M = sage_identity_matrix( rank ) dpl = DPLattice( d_lst, Md_lst, M ) # change basis a_lst = [ 'e0-e1', 'e0-e2'] a_lst = [ Div.new( a, rank ) for a in a_lst ] m1_lst = get_divs( get_ak( rank ), 1, -1, True ) d_tup_lst = get_bases_lst( a_lst, M, d_lst, m1_lst, False ) B = sage_matrix( sage_ZZ, [ d.e_lst for d in d_tup_lst[0] ] ) dplB = dpl.get_basis_change( B ) int_mat = list( dplB.m1_lst[0].int_mat ) print( dplB ) print( str( d_tup_lst ) ) assert str( d_tup_lst ) == '[(e0-e1, e0-e2, e3, e0-e1-e2)]' print( list( B ) ) assert str( list( B ) ) == '[(1, -1, 0, 0), (1, 0, -1, 0), (0, 0, 0, 1), (1, -1, -1, 0)]' print( str( int_mat ) ) assert str( int_mat ) == '[(0, 1, 0, 0), (1, 0, 0, 0), (0, 0, -1, 0), (0, 0, 0, -1)]' NSTools.set_enable_tool_dct( True ) def test__get_basis_lst__rank_4__True( self ): NSTools.set_enable_tool_dct( False ) rank = 4 # construct DPLattice d_lst = [] Md_lst = [] M = sage_identity_matrix( rank ) dpl = DPLattice( d_lst, Md_lst, M ) # change basis a_lst = [ 'e0-e1', 'e0-e2'] a_lst = [ Div.new( a, rank ) for a in a_lst ] m1_lst = get_divs( get_ak( rank ), 1, -1, True ) d_tup_lst = get_bases_lst( a_lst, M, d_lst, m1_lst, True ) print( d_tup_lst ) assert str( d_tup_lst ) == '[(e0-e1, e0-e2, e3, e0-e1-e2), (e0-e1, e0-e2, e0-e1-e2, e3)]' for d_tup in d_tup_lst: B = sage_matrix( sage_ZZ, [ d.e_lst for d in d_tup ] ) dplB = dpl.get_basis_change( B ) int_mat = list( dplB.m1_lst[0].int_mat ) print( str( int_mat ) ) assert str( int_mat ) == '[(0, 1, 0, 0), (1, 0, 0, 0), (0, 0, -1, 0), (0, 0, 0, -1)]' NSTools.set_enable_tool_dct( True ) def test__get_webs__rank_4( self ): NSTools.set_enable_tool_dct( False ) # sage_register_unpickle_override( 'class_div', 'Div', Div ) # sage_register_unpickle_override( 'class_dp_lattice', 'DPLattice', DPLattice ) d_lst = [] Md_lst = [] M = sage_identity_matrix( 4 ) dpl = DPLattice( d_lst, Md_lst, M ) fam_lst_lst = get_webs( dpl ) for fam_lst in fam_lst_lst: print( fam_lst ) NSTools.set_enable_tool_dct( True ) def test__contains_perm__rank6( self ): f_lst_lst = [['e0-e1', '2e0-e2-e3-e4-e5'], ['e0-e5', 'e0', 'e1']] c_lst = ['e0-e2', '2e0-e1-e3-e4-e5'] rank = 6 nf_lst_lst = [] for f_lst in f_lst_lst: nf_lst_lst += [[ Div.new( f, rank ) for f in f_lst ]] f_lst_lst = nf_lst_lst c_lst = [ Div.new( c, rank ) for c in c_lst ] assert contains_perm( f_lst_lst, c_lst ) def test__triples( self ): NSTools.set_enable_tool_dct( False ) rank = 6 # (2A1, 4A1) d_lst = [ 'e2-e4', 'e3-e5', 'e0-e1-e2-e4', 'e0-e1-e3-e5'] Md_lst = ['e4-e5', 'e0-e1-e2-e3'] M = [( 2, 1, 1, 1, 0, 0 ), ( -1, 0, -1, -1, 0, 0 ), ( -1, -1, 0, -1, 0, 0 ), ( -1, -1, -1, 0, 0, 0 ), ( 0, 0, 0, 0, 0, 1 ), ( 0, 0, 0, 0, 1, 0 )] d_lst = [ Div.new( d, rank ) for d in d_lst ] Md_lst = [ Div.new( Md, rank ) for Md in Md_lst ] M = sage_matrix( M ) dpl = DPLattice( d_lst, Md_lst, M ) t_lst = triples( dpl, 2 ) print( t_lst ) assert str( t_lst ) == '[[e0-e1, e0-e2, 2e0-e2-e3-e4-e5]]' NSTools.set_enable_tool_dct( True ) if __name__ == '__main__': # NSTools.filter( 'ns_basis.py' ) NSTools.filter( None ) # TestNSBasis().test__get_basis_lst__rank_4__False() # TestNSBasis().test__get_basis_lst__rank_4__True() # TestNSBasis().test__get_webs__rank_4() # TestNSBasis().test__contains_perm__rank6() # TestNSBasis().test__triples() pass
5,021
29.071856
153
py
ns_lattice
ns_lattice-master/ns_lattice/src/tests/test_convert_to_tex.py
''' Use of this source code is governed by a MIT-style license that can be found in the LICENSE file. Created on Feb 8, 2017 @author: Niels Lubbes ''' from ns_lattice.class_ns_tools import NSTools from ns_lattice.convert_to_tex import cls_to_tex class TestConvertToTex: def test__cls_to_tex( self ): if 'get_cls_9' not in NSTools.get_tool_dct(): return out = cls_to_tex() print( out ) if __name__ == '__main__': NSTools.filter( None ) TestConvertToTex().test__cls_to_tex() pass
543
16.548387
97
py
ns_lattice
ns_lattice-master/ns_lattice/src/tests/__init__.py
0
0
0
py
ns_lattice
ns_lattice-master/ns_lattice/src/tests/test_dp_involutions.py
''' Use of this source code is governed by a MIT-style license that can be found in the LICENSE file. Created on Feb 8, 2017 @author: Niels Lubbes ''' from ns_lattice.sage_interface import sage_vector from ns_lattice.sage_interface import sage_matrix from ns_lattice.sage_interface import sage_identity_matrix from ns_lattice.sage_interface import sage_diagonal_matrix from ns_lattice.dp_involutions import complete_basis from ns_lattice.dp_involutions import is_integral_involution from ns_lattice.dp_involutions import basis_to_involution from ns_lattice.class_div import Div from ns_lattice.class_ns_tools import NSTools class TestDPInvolutions(): def test__complete_basis__34_45_rank6( self ): d_lst = [ 34, 45] rank = 6 d_lst = [ Div.new( str( d ), rank ) for d in d_lst ] mat = complete_basis( d_lst ) assert mat == sage_matrix( [( 0, 0, -1, 0, 0, 0 ), ( 0, 0, 0, 1, 0, 0 ), ( 0, 0, 0, 0, 1, 0 ), ( 1, 0, 0, 0, 0, 1 ), ( -1, 1, 0, 0, 0, 1 ), ( 0, -1, 0, 0, 0, 1 )] ) def test__complete_basis__23_34_45_rank6( self ): d_lst = [ 23, 34, 45 ] rank = 6 d_lst = [ Div.new( str( d ), rank ) for d in d_lst ] mat = complete_basis( d_lst ) assert mat == sage_matrix( [( 0, 0, 0, -1, 0, 0 ), ( 0, 0, 0, 0, 1, 0 ), ( 1, 0, 0, 0, 0, 1 ), ( -1, 1, 0, 0, 0, 1 ), ( 0, -1, 1, 0, 0, 1 ), ( 0, 0, -1, 0, 0, 1 )] ) def test__complete_basis__1123_12_23_45_rank6( self ): # 4A1 d_lst = [ 1123, 12, 23, 45 ] rank = 6 d_lst = [ Div.new( str( d ), rank ) for d in d_lst ] mat = complete_basis( d_lst ) print( mat ) assert mat == sage_matrix( [( 0, 0, 0, 1, -3, 0 ), ( 1, 0, 0, -1, 1, 0 ), ( -1, 1, 0, -1, 1, 0 ), ( 0, -1, 0, -1, 1, 0 ), ( 0, 0, 1, 0, 0, 1 ), ( 0, 0, -1, 0, 0, 1 ) ] ) def test__complete_basis__1145_23_rank6( self ): d_lst = [ 1145, 23 ] rank = 6 d_lst = [ Div.new( str( d ), rank ) for d in d_lst ] mat = complete_basis( d_lst ) print( mat ) assert mat == sage_matrix( [( 0, 1, -1, 0, 0, 0 ), ( 0, -1, 0, 1, 0, 0 ), ( 1, 0, 0, 0, 1, 0 ), ( -1, 0, 0, 0, 1, 0 ), ( 0, -1, 0, 0, 0, 1 ), ( 0, -1, 1, -1, 0, -1 ) ] ) def test__complete_basis__12_23_rank4( self ): d_lst = [ 12, 23 ] rank = 4 d_lst = [ Div.new( str( d ), rank ) for d in d_lst ] V = complete_basis( d_lst ) D = sage_diagonal_matrix( [-1, -1, 1, 1] ) J = sage_diagonal_matrix( [1, -1, -1, -1] ) M = V * D * ~V assert str( list( M ) ) == "[(1, 0, 0, 0), (0, -1/3, 2/3, 2/3), (0, 2/3, -1/3, 2/3), (0, 2/3, 2/3, -1/3)]" assert M * M == sage_identity_matrix( 4 ) assert M.T * J * M == J assert is_integral_involution( M ) == False def test__complete_basis__1123_12_23_rank4( self ): d_lst = [ 1123, 12, 23 ] rank = 4 d_lst = [ Div.new( str( d ), rank ) for d in d_lst ] V = complete_basis( d_lst ) D = sage_diagonal_matrix( [-1, -1, -1, 1] ) J = sage_diagonal_matrix( [1, -1, -1, -1] ) M = V * D * ~V print( ~V * sage_vector( [1, -1, -1, -1] ) ) print( ~V * sage_vector( [0, 1, 0, -1] ) ) print( ~V * sage_vector( [0, 0, 1, -1] ) ) print( V ) assert M == basis_to_involution( d_lst, rank ) assert str( list( V ) ) == "[(0, 0, 1, -3), (1, 0, -1, 1), (-1, 1, -1, 1), (0, -1, -1, 1)]" assert str( list( ~V ) ) == "[(0, 2/3, -1/3, -1/3), (0, 1/3, 1/3, -2/3), (-1/2, -1/2, -1/2, -1/2), (-1/2, -1/6, -1/6, -1/6)]" assert str( list( M ) ) == "[(2, 1, 1, 1), (-1, -4/3, -1/3, -1/3), (-1, -1/3, -4/3, -1/3), (-1, -1/3, -1/3, -4/3)]" assert M * M == sage_identity_matrix( 4 ) assert M.T * J * M == J assert is_integral_involution( M ) == False def test__complete_basis__12__rank4( self ): d_lst = [ 12 ] rank = 4 d_lst = [ Div.new( str( d ), rank ) for d in d_lst ] V = complete_basis( d_lst ) D = sage_diagonal_matrix( [-1, 1, 1, 1] ) J = sage_diagonal_matrix( [1, -1, -1, -1] ) M = V * D * ~V assert M == basis_to_involution( d_lst, rank ) assert str( list( M ) ) == "[(1, 0, 0, 0), (0, 0, 1, 0), (0, 1, 0, 0), (0, 0, 0, 1)]" assert M * M == sage_identity_matrix( 4 ) assert M.T * J * M == J assert is_integral_involution( M ) == True def test__complete_basis__1123__rank4( self ): d_lst = [ 1123 ] rank = 4 d_lst = [ Div.new( str( d ), rank ) for d in d_lst ] V = complete_basis( d_lst ) D = sage_diagonal_matrix( [-1, 1, 1, 1] ) J = sage_diagonal_matrix( [1, -1, -1, -1] ) M = V * D * ~V assert M == basis_to_involution( d_lst, rank ) assert str( list( M ) ) == "[(2, 1, 1, 1), (-1, 0, -1, -1), (-1, -1, 0, -1), (-1, -1, -1, 0)]" assert M * M == sage_identity_matrix( 4 ) assert M.T * J * M == J assert is_integral_involution( M ) == True def test__complete_basis__1123_12__rank4( self ): d_lst = [ 1123, 12 ] rank = 4 d_lst = [ Div.new( str( d ), rank ) for d in d_lst ] V = complete_basis( d_lst ) D = sage_diagonal_matrix( [-1, -1, 1, 1] ) J = sage_diagonal_matrix( [1, -1, -1, -1] ) M = V * D * ~V assert M == basis_to_involution( d_lst, rank ) assert str( list( M ) ) == "[(2, 1, 1, 1), (-1, -1, 0, -1), (-1, 0, -1, -1), (-1, -1, -1, 0)]" assert M * M == sage_identity_matrix( 4 ) assert M.T * J * M == J assert is_integral_involution( M ) == True if __name__ == '__main__': NSTools.filter( None ) # TestDPInvolutions().test__complete_basis__12_23_rank4() # TestDPInvolutions().test__complete_basis__1123_12_23_rank4() # TestDPInvolutions().test__complete_basis__12__rank4() # TestDPInvolutions().test__complete_basis__1123__rank4() # TestDPInvolutions().test__complete_basis__1123_12__rank4() pass
6,780
37.310734
133
py
ns_lattice
ns_lattice-master/ns_lattice/src/ns_lattice/class_eta.py
''' Use of this source code is governed by a MIT-style license that can be found in the LICENSE file. Created on Jan 27, 2018 @author: Niels Lubbes ''' from ns_lattice.sage_interface import sage_n from ns_lattice.class_ns_tools import NSTools import time class ETA( object ): ''' For estimating the time it takes for a loop in a program to terminate (ETA=estimated time of arrival). During the loop feedback is printed. ''' def __init__( self, total, ival ): ''' Should be called before a loop in starts. Parameters ---------- total: int Number of times the loop needs to be traced. ival : int Nonzero number of traced loops in program until feedback about etimated end time is printed ''' # total number of loops self.total = total # number of loops after which eta is updated self.ival = 1 if ival > 0: self.ival = ival # loop counter self.counter = 0 # times self.ini_time = self.time() # time when method was called self.prv_time = self.ini_time # time which is updated after ival loops. self.eta_time = 0 # estimated time of arrival in minutes def time( self ): return time.time() def update( self, *info_lst ): ''' Should be called inside a loop. Prints an estimation for the time it takes for a program to terminate (ETA for short). We refer to the program termination as arrival. Parameters ---------- *info_lst : string Variable length argument list consisting of additional information that is printed together with ETA. ''' if self.counter % self.ival == 0: cur_time = self.time() ival_time = ( cur_time - self.prv_time ) / ( 60 * self.ival ) passed_time = sage_n( ( cur_time - self.ini_time ) / 60, digits = 5 ) self.eta_time = sage_n( ival_time * ( self.total - self.counter ), digits = 5 ) s = '' for info in info_lst: s += str( info ) + ' ' NSTools.p( 'ETA =', self.eta_time, 'm,', 'counter =', self.counter, '/', self.total, ',', 'time =', passed_time, 'm,', 'info =', s ) # update previous time self.prv_time = cur_time # increase counter self.counter += 1
2,603
26.410526
97
py
ns_lattice
ns_lattice-master/ns_lattice/src/ns_lattice/dp_root_bases.py
''' Created on Aug 11, 2016 @author: Niels See [http://arxiv.org/abs/1302.6678] for more info. Classification of root subsystems of root systems of type either A1, A1+A2, A4, D5, E6, E7 or E8. ''' import time from ns_lattice.sage_interface import sage_VectorSpace from ns_lattice.sage_interface import sage_vector from ns_lattice.sage_interface import sage_QQ from ns_lattice.sage_interface import sage_identity_matrix from ns_lattice.sage_interface import sage_Graph from ns_lattice.sage_interface import sage_Partitions from ns_lattice.sage_interface import sage_RootSystem from ns_lattice.sage_interface import sage_Subsets from ns_lattice.sage_interface import sage_Combinations from ns_lattice.sage_interface import sage_Permutations from ns_lattice.class_ns_tools import NSTools from ns_lattice.class_div import Div from ns_lattice.div_in_lattice import get_divs from ns_lattice.div_in_lattice import get_indecomp_divs from ns_lattice.div_in_lattice import get_ak def is_root_basis( d_lst ): ''' Parameters ---------- d_lst : list<Div> A list of lists of "Div" objects "d", such that d*d=-2 and d*(-3h+e1+...+er)=0 where r=rank-1 and rank in [3,...,7]. Returns ------- bool True if input is the empty list or if divisors in "d_lst" are linear independent as vectors and their pairwise product is either -2, 0 or 1. ''' # check empty-list if d_lst == []: return True # check pairwise inner product for i in range( len( d_lst ) ): for j in range( len( d_lst ) ): if d_lst[i] * d_lst[j] not in [0, 1, -2]: return False # check linear independence # Linear independent vectors with pairwise positive intersection product # must form a root basis. Thus vectors of positive roots in the corresponding # root system are all positive # V = sage_VectorSpace( sage_QQ, d_lst[0].rank() ) W = V.subspace( [d.e_lst for d in d_lst] ) return W.rank() == len( d_lst ) def get_graph( d_lst ): ''' Parameters ---------- d_lst : list<Div> A list of "Div" objects. Returns ------- sage_Graph A labeled "Graph()" where the elements of "d_lst" are the vertices. Different vertices are connected if their corresponding intersection product is non-zero and the edge is labeled with the intersection product. ''' G = sage_Graph() G.add_vertices( range( len( d_lst ) ) ); for i in range( len( d_lst ) ): for j in range( len( d_lst ) ): if d_lst[i] * d_lst[j] > 0 and i != j: G.add_edge( i, j, d_lst[i] * d_lst[j] ) return G def get_ext_graph( d_lst, M ): ''' Parameters ---------- d_lst : list<Div> A list of "Div" objects of equal rank. M : sage_matrix<sage_ZZ> A square matrix with integral coefficients of rank "d_lst[0].rank()" Returns ------- A labeled "sage_Graph()" where the elements of "d_lst" are the vertices. A pair of non-orthogonal vertices are connected by and edge labeled with their non-zero intersection product. Two vertices which are related via M are connected with an edge labeled 1000. Labeled self-loops are also included. ''' NSTools.p( 'd_lst =', len( d_lst ), d_lst, ', M =', list( M ) ) G = sage_Graph() G.add_vertices( range( len( d_lst ) ) ) for i in range( len( d_lst ) ): for j in range( len( d_lst ) ): if d_lst[i] * d_lst[j] != 0: G.add_edge( i, j, d_lst[i] * d_lst[j] ) for i in range( len( d_lst ) ): j = d_lst.index( d_lst[i].mat_mul( M ) ) G.add_edge( i, j, 1000 ) return G def get_dynkin_type( d_lst ): ''' Parameters ---------- d_lst : list<Div> A list of lists of "Div" objects "d" of the same rank, such that d*d=-2 and d*(-3h+e1+...+er)=0 where r=rank-1 and rank in [3,...,9]. We assume that "is_root_basis(d_lst)==True": linear independent, self intersection number -2 and pairwise product either 0 or 1. Returns ------- string Returns a string denoting the Dynkin type of a root system with basis "d_lst". Returns 'A0' if "d_lst==[]". Note ---- For example: [<1145>, <1123>, <23>, <45>, <56>, <78>] --> '3A1+A3' where <1145> is shorthand for "Div.new('1145')". Raises ------ ValueError If the Dynkin type of d_lst cannot be recognized. ''' if d_lst == []: return 'A0' # check whether values are cached # construct_dynkin_types = True max_r = d_lst[0].rank() - 1 key = 'get_dynkin_type_' + str( max_r ) for r in range( max_r, 8 + 1 ): if 'get_dynkin_type_' + str( r ) in NSTools.get_tool_dct(): key = 'get_dynkin_type_' + str( r ) construct_dynkin_types = False # construct list of dynkin types if values are not cached # if construct_dynkin_types: NSTools.p( 'Constructing list of Dynkin types... max_r =', max_r ) ade_lst = [] for comb_lst in sage_Combinations( max_r * ['A', 'D', 'E'], max_r ): for perm_lst in sage_Permutations( comb_lst ): ade_lst += [perm_lst] # # "ade_lst" contains all combinations of 'A', 'D', 'E' # and looks as follows: # # ade_lst[0] = ['A', 'A', 'A', 'A', 'A', 'A', 'A', 'A'] # ade_lst[1] = ['A', 'A', 'A', 'A', 'A', 'A', 'A', 'D'] # ade_lst[2] = ['A', 'A', 'A', 'A', 'A', 'A', 'D', 'A'] # ... # ade_lst[?] = ['A', 'D', 'A', 'D', 'A', 'D', 'E', 'A'] # ... # ade_lst[-1]= ['E', 'E', 'E', 'E', 'E', 'E', 'E', 'E'] # type_lst = [] ts_lst = [] for ade in ade_lst: for r in range( 1, max_r + 1 ): for p_lst in sage_Partitions( r + max_r, length = max_r ): # obtain type list t_lst = [( ade[i], p_lst[i] - 1 ) for i in range( max_r ) if p_lst[i] != 1] t_lst.sort() # obtain Root system # or continue if invalid Cartan/Dynkin type if ( 'D', 2 ) in t_lst or ( 'D', 3 ) in t_lst: continue try: rs = sage_RootSystem( t_lst ) except ValueError as err: continue # not a valid Cartan type # obtain graph G mat = list( -1 * rs.cartan_matrix() ) G = sage_Graph() G.add_vertices( range( len( mat ) ) ); for i in range( len( mat ) ): for j in range( len( mat[0] ) ): if mat[i][j] == 1: G.add_edge( i, j ) # obtain string for type # Example: [(A,1),(A,1),(A,1),(A,3)] ---> '3A1+A3' tmp_lst = [t for t in t_lst] ts = '' while len( tmp_lst ) > 0: t = tmp_lst[0] c = tmp_lst.count( t ) while t in tmp_lst: tmp_lst.remove( t ) if ts != '': ts += '+' if c > 1: ts += str( c ) ts += t[0] + str( t[1] ) # add to type_lst if new if ts not in ts_lst: type_lst += [( G, ts, t_lst )] ts_lst += [ts] NSTools.p( 'added to list: ', ts, '\t\t...please wait...' ) NSTools.p( 'Finished constructing list of Dynkin types.' ) # cache the constructed "type_lst" NSTools.get_tool_dct()[key] = type_lst NSTools.save_tool_dct() # end if else: type_lst = NSTools.get_tool_dct()[key] G1 = get_graph( d_lst ) # loop through all types and check equivalence for ( G2, ts, t_lst ) in type_lst: if G1.is_isomorphic( G2 ): return ts raise ValueError( 'Could not recognize Dynkin type: ', d_lst ) def convert_type( type ): ''' Converts a Dynkin type string to a sorted list of irreducible Dynkin types. For example if type is '2A1+D4', then the output is ['A1','A1','D4']. If the type is '2A1+A2+A3', then the output is ['A1','A1','A2','A3']. We exclude elements that are equal to 'A0'. Parameters ---------- type: string A string representing a Dynkin type. We assume that an irreducible rootsystem occurs with multiplicity at most 9. For example '10A1' is not allowed, but '9A1' is allowed. Returns ------- list<string> A list of string representing the Dynkin type of an irreducible root system. ''' t_lst = type.split( '+' ) out_lst = [] for t in t_lst: if t[0] not in ['A', 'D', 'E']: mult, subtype = int( t[0] ), t[1:] else: mult, subtype = 1, t out_lst += mult * [ subtype ] out_lst = [out for out in out_lst if out != 'A0'] return sorted( out_lst ) def get_root_bases_orbit( d_lst, positive = True ): ''' Computes the orbit of a root base under the Weyl group. Parameters ---------- d_lst : list<Div> A list of lists of "Div" objects "d" of the same rank or the empty list. positive : bool Returns ------- list<list<Div>> A list of distinct lists of "Div" objects "d" of the same rank. such that d*d=-2 and d*(-3h+e1+...+er)=0 where r=rank-1. If "d_lst" is the empty list, then "[]" is returned. Otherwise we return a list of root bases such that each root basis is obtained as follows from a root "s" such that s*s=-2 and s*(-3h+e1+...+er)=0: [ d + (d*s)d for d in d_lst ] We do this for all possible roots in [s1,s2,s3,...]: [ [ d + (d*s1)d for d in d_lst ], [ d + (d*s2)d for d in d_lst ], ... ] Mathematically, this means that we consider the Weyl group of the root system with Dynkin type determined by the rank of elements in "d_lst". The Dynkin type is either A1, A1+A2, A4, D5, E6, E7 or E8. We return the orbit of the elements in "d_lst" under the action of the Weyl group. If "positive==True" then the roots in the basis are all positive and thus of the form <ij>, <1ijk>, <2ij>, <30i> with i<j<k. For example '15' and '1124' but not '-15' or '-1124'. See "Div.get_label()" for the notation. ''' if d_lst == []: return [[]] rank = d_lst[0].rank() # in cache? key = 'get_root_bases_orbit_' + str( d_lst ) + '_' + str( rank ) if key in NSTools.get_tool_dct(): return NSTools.get_tool_dct()[key] # obtain list of all positive (-2)-classes m2_lst = get_divs( get_ak( rank ), 0, -2, True ) # m2_lst += [ m2.int_mul( -1 ) for m2 in m2_lst] NSTools.p( 'd_lst =', len( d_lst ), d_lst, ', m2_lst =', len( m2_lst ), m2_lst ) # data for ETA computation counter = 0 total = len( m2_lst ) ival = 5000 d_lst.sort() d_lst_lst = [d_lst] for cd_lst in d_lst_lst: total = len( m2_lst ) * len( d_lst_lst ) for m2 in m2_lst: # ETA if counter % ival == 0: start = time.time() counter += 1 if counter % ival == 0: passed_time = time.time() - start NSTools.p( 'ETA in minutes =', passed_time * ( total - counter ) / ( ival * 60 ), ', len(d_lst_lst) =', len( d_lst_lst ), ', total =', total ) # # The action of roots on a root base is by reflection: # cd - 2(cd*m2/m2*m2)m2 # Notice that m2*m2==-2. # od_lst = [ cd + m2.int_mul( cd * m2 ) for cd in cd_lst] # print( 'm2 =', m2, ', od_lst =', od_lst, ', cd_lst =', cd_lst, ', d_lst_lst =', d_lst_lst, ' positive =', positive ) od_lst.sort() if od_lst not in d_lst_lst: d_lst_lst += [od_lst] # select positive roots if positive==True pd_lst_lst = [] for d_lst in d_lst_lst: if positive and '-' in [ d.get_label( True )[0] for d in d_lst ]: continue # continue with for loop since a negative root in basis pd_lst_lst += [d_lst] # cache output NSTools.get_tool_dct()[key] = pd_lst_lst NSTools.save_tool_dct() NSTools.p( '#orbit(' + str( d_lst ) + ') =', len( pd_lst_lst ) ) return pd_lst_lst
13,321
30.56872
158
py
ns_lattice
ns_lattice-master/ns_lattice/src/ns_lattice/dp_involutions.py
''' Created on Aug 11, 2016 @author: Niels Lubbes Classification of unimodular involutions of Neron-Severi lattice of weak del Pezzo surfaces. ''' from ns_lattice.sage_interface import sage_ZZ from ns_lattice.sage_interface import sage_QQ from ns_lattice.sage_interface import sage_matrix from ns_lattice.sage_interface import sage_identity_matrix from ns_lattice.sage_interface import sage_diagonal_matrix from ns_lattice.class_ns_tools import NSTools from ns_lattice.class_div import Div from ns_lattice.div_in_lattice import get_ak def complete_basis( d_lst ): ''' Parameters ---------- d_lst : list<Div> A list of "Div" objects. Returns ------- sage_matrix<sage_QQ> Returns a square matrix over QQ of full rank. The first columns correspond to the elements in d_lst (where d_lst is sorted). The appended columns are orthogonal to the first "len(d_lst)" columns. Examples -------- We explain with 3 examples where the dotted (:) vertical lines denote which columns are appended. | 0 0 : 1 0 0 0 | | 0 0 0 : 1 0 0 | | 0 0 0 1 : -3 0 | | 0 0 : 0 -1 0 0 | | 0 0 0 : 0 -1 0 | | 1 0 0 -1 : 1 0 | | 0 0 : 0 0 -1 0 | | 1 0 0 : 0 0 -1 | |-1 1 0 -1 : 1 0 | | 1 0 : 0 0 0 -1 | |-1 1 0 : 0 0 -1 | | 0 -1 0 -1 : 1 0 | |-1 1 : 0 0 0 -1 | | 0 -1 1 : 0 0 -1 | | 0 0 1 0 : 0 1 | | 0 -1 : 0 0 0 -1 | | 0 0 -1 : 0 0 -1 | | 0 0 -1 0 : 0 1 | ''' # sort d_lst = [ d for d in d_lst] d_lst.sort() # extend with orthogonal vectors row_lst = [ d.e_lst for d in d_lst] ext_lst = [] for v_lst in sage_matrix( sage_ZZ, row_lst ).right_kernel().basis(): ext_lst += [ [-v_lst[0]] + list( v_lst[1:] ) ] # accounts for signature (1,rank-1) mat = sage_matrix( sage_QQ, row_lst + ext_lst ).transpose() # verify output if mat.rank() < d_lst[0].rank(): raise Error( 'Matrix expected to have full rank: ', d_lst, '\n' + str( mat ) ) de_lst = [ Div( ext ) for ext in ext_lst ] for de in de_lst: for d in d_lst: if d * de != 0: raise Error( 'Extended columns are expected to be orthogonal: ', de, d, de_lst, d_lst, list( mat ) ) return mat def is_integral_involution( M ): ''' Parameters ---------- M : sage_matrix A matrix M. Returns ------- bool Returns True if the matrix is an involution, preserves inner product with signature (1,r) and has integral coefficients. ''' nrows, ncols = M.dimensions() # check whether involution if M * M != sage_identity_matrix( nrows ): return False # check whether inner product is preserved S = sage_diagonal_matrix( [1] + ( ncols - 1 ) * [-1] ) if M.transpose() * S * M != S: return False # check whether coefficients are integral for r in range( nrows ): for c in range( ncols ): if M[r][c] not in sage_ZZ: return False # check whether canonical class is preserved ak = get_ak( nrows ) if ak.mat_mul( M ) != ak: return False return True def basis_to_involution( d_lst, rank ): ''' Parameters ---------- d_lst : list<Div> A list of "Div" objects of rank "rank". rank : int An integer in [3,...,9]. Returns ------- sage_MATRIX<sage_QQ> Returns matrix over QQ that correspond to an involution of ZZ<h,e1,...,er> here r=rank-1. The first columns correspond to the elements in d_lst (where d_lst is sorted). The appended columns are orthogonal to the first "len(d_lst)" columns. ''' if d_lst == []: return sage_identity_matrix( sage_QQ, rank ) l = len( d_lst ) V = complete_basis( d_lst ) D = sage_diagonal_matrix( l * [-1] + ( rank - l ) * [1] ) M = V * D * V.inverse() # MV=VD return M
4,093
28.035461
116
py
ns_lattice
ns_lattice-master/ns_lattice/src/ns_lattice/__main__.py
''' Use of this source code is governed by a MIT-style license that can be found in the LICENSE file. Created on Aug 11, 2016 @author: Niels Lubbes ''' import sys import os from ns_lattice.sage_interface import sage_matrix from ns_lattice.sage_interface import sage_ZZ from ns_lattice.sage_interface import sage_identity_matrix from ns_lattice.sage_interface import sage_Subsets from ns_lattice.sage_interface import sage_Permutations from ns_lattice.sage_interface import sage_Combinations from ns_lattice.sage_interface import sage_Graph from ns_lattice.sage_interface import sage_gcd from ns_lattice.sage_interface import sage_factor from ns_lattice.class_ns_tools import NSTools from ns_lattice.class_div import Div from ns_lattice.div_in_lattice import get_divs from ns_lattice.div_in_lattice import get_ak from ns_lattice.class_dp_lattice import DPLattice from ns_lattice.ns_basis import get_bases_lst from linear_series.class_poly_ring import PolyRing from linear_series.class_base_points import BasePointTree from linear_series.class_linear_series import LinearSeries def usecase__get_cls( max_rank ): ''' Classification of root bases in root system of rank at most "max_rank". See "DPLattice.get_cls_root_bases()". Parameters ---------- max_rank : int Maximal rank. ''' row_format = '{:>6}{:>5}{:>8}{:>16}{:>5}{:>5}{:>5}{:>5}{:>6}{:>7}{:>70}{:>135}{:>340}' rownr = 0 for rank in range( 3, max_rank + 1 ): dpl_lst = DPLattice.get_cls( rank ) row_lst = [['rownr', 'rank', 'Mtype', 'type', '#-2', '#-1', '#fam', '#-2R', '#-1R', '#famR', 'Md_lst', 'd_lst', 'M']] for dpl in sorted( dpl_lst ): row_lst += [ [rownr, rank, dpl.get_marked_Mtype(), dpl.get_real_type() ] + list( dpl.get_numbers() ) + [str( dpl.Md_lst )] + [str( dpl.d_lst )] + [str( list( dpl.M ) )] ] rownr += 1 s = '' for row in row_lst: s += row_format.format( *row ) + '\n' NSTools.p( 'Classification of root bases:\n' + s ) NSTools.p( 'rank =', rank, ', len =', len( dpl_lst ) ) NSTools.p( 80 * '#' ) for rank in range( 3, max_rank + 1 ): NSTools.p( 'rank =', rank, ', len =', len( DPLattice.get_cls( rank ) ) ) NSTools.p( 80 * '#' ) def usecase__get_classes_dp1( rank ): ''' Computes classes in the Neron-Severi lattice with predefined self-intersection and intersection with the canonical class. Parameters ---------- rank : int ''' # canonical class d = get_ak( rank ) # basis change a_lst = [ 'e0-e1', 'e0-e2'] a_lst = [ Div.new( a, rank ) for a in a_lst ] m1_lst = get_divs( d, 1, -1, True ) print( d ) M = sage_identity_matrix( rank ) d_lst = [] d_tup_lst = get_bases_lst( a_lst, M, d_lst, m1_lst, False ) B = sage_matrix( sage_ZZ, [ dt.e_lst for dt in d_tup_lst[0] ] ) # list the classes for ( dc, cc ) in [( 2, 0 ), ( 1, -1 ), ( 0, -2 ), ( 2, 2 ), ( 2, 4 ), ( 3, 1 )]: NSTools.p( '(dc, cc) =', ( dc, cc ) ) c_lst = get_divs( d, dc, cc, False ) for c in c_lst: NSTools.p( '\t\t', c, '\t\t', c.get_basis_change( B ) ) def usecase__graphs( max_rank ): ''' Lists attributes of simple family graphs. Parameters ---------- max_rank : int Maximal rank of DPLattice objects that are considered. ''' row_format = '{:<6}{:<5}{:<8}{:<16}{:<7}{:<10}{:<95}{:<30}{:<15}{:<15}{:<15}{:<15}' already_in_cache = True dpl_lst = [] rownr = 0 row_lst = [['rownr', 'deg', 'Mtype', 'type', '#vert', '#edges', 'degrees', 'labels', 'complete', 'connected', 'vert-xfer', 'edge-xfer']] for rank in range( 3, max_rank + 1 ): NSTools.p( 'rank =', rank ) for dpl in DPLattice.get_cls( rank ): already_in_cache = already_in_cache and ( dpl.SG != None ) dpl_lst += [dpl] SG, SG_data = dpl.get_SG() row_lst += [ [rownr, 10 - rank, dpl.get_marked_Mtype(), dpl.get_real_type() ] + SG_data ] rownr += 1 if rank == 9 and ( rownr <= 390 or rownr % 100 == 0 ): NSTools.p( '\t\trownr =', rownr ) s = '' for row in row_lst: s += row_format.format( *row ) + '\n' NSTools.p( 'Classification of simple family graphs:\n' + s ) if not already_in_cache: NSTools.p( 'Saving data for simple family graphs...' ) NSTools.save_tool_dct() # example for how to plot a simple family graph # NSTools.p( 'Plotting a simple family graph...' ) SG, SG_data = DPLattice.get_cls( 6 )[0].get_SG() P = SG.graphplot( vertex_size = 1, vertex_labels = True, edge_labels = True, color_by_label = False, layout = 'circular' ).plot() P.save( os.environ['OUTPUT_PATH'] + 'graph.png' ) NSTools.p( '#components =', SG.connected_components_number() ) def usecase__analyze_graphs( max_rank ): ''' We analyze the graphs of DPLattice objects in the output of DPLattice.get_cls(). Parameters ---------- max_rank : int Maximal rank of DPLattice objects that are considered. ''' # Examine which of the graphs associated to DPLattices # are isomorphic to one of the constructed graphs. # NSTools.p( '\t Compare contructed graphs with classified graphs...' ) rownr = -1 max_verts = 0 for rank in range( 3, max_rank + 1 ): NSTools.p( '\t ---' ) for dpl in DPLattice.get_cls( rank ): rownr += 1 # retrieve the graph SG for analysis SG, SG_data = dpl.get_SG() if SG.num_verts() <= 3: continue # check if each edge label is in [2,4] if [e for e in SG_data[3] if e not in [2, 4]] != []: continue # initialize string s = '' s += str( rownr ) + ' ' + 'rank=' + str( rank ) + ' ' # Initialize G_lst which is a list of tuples (G,G_str) # where G is a constructed graph and G_str is its string identifier. # The identifiers are according Theorem 1 in arXiv:1807.05881v2. # G_lst = [] for nv1 in range( 1, SG.num_verts() + 1 ): for nv2 in range( 1, SG.num_verts() + 1 ): # determine list c_lst of 2-element subsets of [1,...,m] # so that m is minimal under the condition that len(c_lst)>=nv1 c_lst = [] for i in range( 2 * nv1 ): c_lst = list( sage_Combinations( i, 2 ) ) if len( c_lst ) >= nv1: break # construct graphs # Gd = sage_Graph() Gd.add_vertices( range( nv1 ) ) G_lst += [( Gd, 'Gd:' + str( nv1 ) )] Ge = sage_Graph() Ge.add_vertices( range( nv1 ) ) for i in Ge.vertices(): for j in Ge.vertices(): Ge.add_edge( i, j, 2 ) G_lst += [( Ge, 'Ge:' + str( nv1 ) )] Gf = sage_Graph() Gf.add_vertices( range( len( c_lst ) ) ) for i in Gf.vertices(): for j in Gf.vertices(): q = len( [ c for c in c_lst[i] if c in c_lst[j] ] ) Gf.add_edge( i, j, 4 - 2 * q ) G_lst += [( Gf, 'Gf:' + str( Gf.num_verts() ) )] Gg = sage_Graph() Gg.add_vertices( range( len( c_lst ) ) ) for i in Gg.vertices(): for j in Gg.vertices(): q = len( [ c for c in c_lst[i] if c in c_lst[j] ] ) if q > 0: Gg.add_edge( i, j, 2 ) G_lst += [( Gg, 'Gg:' + str( Gg.num_verts() ) )] # construct combined graphs # if nv1 + nv2 > SG.num_verts(): continue Gd2 = sage_Graph() Gd2.add_vertices( range( nv2 ) ) Ge2 = sage_Graph() Ge2.add_vertices( range( nv2 ) ) for i in Ge2.vertices(): for j in Ge2.vertices(): Ge2.add_edge( i, j, 2 ) if nv1 + nv2 == SG.num_verts(): if ( Gd.num_verts(), Ge2.num_verts() ) != ( 1, 1 ): Gde = sage_Graph() Gde.add_vertices( Ge2.vertices() ) Gde.add_edges( Ge2.edges() ) for i in range( Ge2.num_verts() - 1, -1, -1 ): Gde.relabel( {i:i + Gd.num_verts()} ) Gde.add_vertices( Gd.vertices() ) Gde.add_edges( Gd.edges() ) for i in range( Gd.num_verts() ): for j in range( Gd.num_verts(), Gde.num_verts() ): Gde.add_edge( i, j, 2 ) G_lst += [( Gde, 'Gde:' + str( Gd.num_verts() ) + '+' + str( Ge2.num_verts() ) )] if len( c_lst ) + nv2 == SG.num_verts(): Gfd = sage_Graph() Gfd.add_vertices( Gd2.vertices() ) Gfd.add_edges( Gd2.edges() ) for i in range( Gd2.num_verts() - 1, -1, -1 ): Gfd.relabel( {i:i + Gf.num_verts()} ) Gfd.add_vertices( Gf.vertices() ) Gfd.add_edges( Gf.edges() ) for i in range( Gf.num_verts() ): for j in range( Gf.num_verts(), Gfd.num_verts() ): Gfd.add_edge( i, j, 2 ) G_lst += [( Gfd, 'Gfd:' + str( Gf.num_verts() ) + '+' + str( Gd2.num_verts() ) )] Gge = sage_Graph() Gge.add_vertices( Ge2.vertices() ) Gge.add_edges( Ge2.edges() ) for i in range( Ge2.num_verts() - 1, -1, -1 ): Gge.relabel( {i:i + Gg.num_verts()} ) Gge.add_vertices( Gg.vertices() ) Gge.add_edges( Gg.edges() ) for i in range( Gg.num_verts() ): for j in range( Gg.num_verts(), Gge.num_verts() ): Gge.add_edge( i, j, 2 ) G_lst += [( Gge, 'Gge:' + str( Gg.num_verts() ) + '+' + str( Ge2.num_verts() ) )] # check for each of the constructed graphs whether # it is isomorphic to dpl.get_SG() # for ( G, G_str ) in G_lst: if SG.is_isomorphic( G, edge_labels = True ): max_verts = max( max_verts, G.num_verts() ) if G_str not in s: s += G_str + ' ' if ':' in s: NSTools.p( '\t', s ) NSTools.p( 'max_verts =', max_verts ) def usecase__construct_surfaces(): ''' We construct a surface parametrization and its Neron-Severi lattice. Requires the linear_series package. ''' # Blowup of projective plane in 3 colinear points # and 2 infinitely near points. The image of the # map associated to the linear series is a quartic # del Pezzo surface with 5 families of conics. # Moreover the surface contains 8 straight lines. # ring = PolyRing( 'x,y,z', True ) p1 = ( -1, 0 ) p2 = ( 0, 0 ) p3 = ( 1, 0 ) p4 = ( 0, 1 ) p5 = ( 2, 0 ) bp_tree = BasePointTree() bp_tree.add( 'z', p1, 1 ) bp_tree.add( 'z', p2, 1 ) bp_tree.add( 'z', p3, 1 ) bp = bp_tree.add( 'z', p4, 1 ) bp.add( 't', p5, 1 ) ls = LinearSeries.get( [3], bp_tree ) NSTools.p( ls.get_bp_tree() ) NSTools.p( 'implicit equation =\n\t', ls.get_implicit_image() ) # construct NS-lattice where p1=e1,...,p5=e5 rank = 6 d_lst = [ 'e0-e1-e2-e3', 'e4-e5' ] # basepoint p5 is infinitely near to p4 Md_lst = [] M = sage_identity_matrix( 6 ) d_lst = [ Div.new( d, rank ) for d in d_lst ] Md_lst = [ Div.new( Md, rank ) for Md in Md_lst ] M = sage_matrix( M ) dpl = DPLattice( d_lst, Md_lst, M ) NSTools.p( 'Neron-Severi lattice =', dpl ) # search representative for the equivalence class in classification assert dpl in DPLattice.get_cls( rank ) def usecase__roman_circles(): ''' We compute circles on a Roman surface. ''' # parametrization of the Roman surface # p_lst = '[ z^2+x^2+y^2, -z*x, -x*y, z*y ]' # we consider the stereographic projection from # S^3 = { x in P^4 | -x0^2+x1^2+x2^2+x3^2+x4^2 = 0 } # where the center of projection is (1:0:0:0:1): # (x0:x1:x2:x3:x4) |---> (x0-x4:x1:x2:x3) # inverse stereographic projection into 3-sphere # s_lst = '[ y0^2+y1^2+y2^2+y3^2, 2*y0*y1, 2*y0*y2, 2*y0*y3, -y0^2+y1^2+y2^2+y3^2 ]' # compose p_lst with s_lst # ring = PolyRing( 'x,y,z,y0,y1,y2,y3' ) x, y, z, y0, y1, y2, y3 = ring.gens() p_lst = ring.coerce( p_lst ) s_lst = ring.coerce( s_lst ) dct = { y0:p_lst[0], y1:p_lst[1], y2:p_lst[2], y3:p_lst[3] } sp_lst = [ s.subs( dct ) for s in s_lst ] NSTools.p( 'sp_lst =' ) for sp in sp_lst: NSTools.p( '\t\t', sage_factor( sp ) ) NSTools.p( 'gcd(sp_lst) =', sage_gcd( sp_lst ) ) # determine base points # ring = PolyRing( 'x,y,z', True ) sp_lst = ring.coerce( sp_lst ) ls = LinearSeries( sp_lst, ring ) NSTools.p( ls.get_bp_tree() ) # We expect that the basepoints come from the intersection # of the Roman surface with the absolute conic: # A = { (y0:y1:y2:y3) in P^3 | y0=y1^2+y2^2+y3^2 = 0 } # # Circles are the image via p_lst of lines that pass through # complex conjugate points. # ring = PolyRing( 'x,y,z', False ) # reinitialize ring with updated numberfield a0, a1, a2, a3 = ring.root_gens() # a0=(1-I*sqrt(3)) with conjugate a0-1 and minimal polynomial t^2-t+1 # we compute candidate classes of circles # h = Div.new( '4e0-e1-e2-e3-e4-e5-e6-e7-e8' ) div_lst = get_divs( h, 2, -2, False ) + get_divs( h, 2, -1, False ) NSTools.p( 'Classes of circles up to permutation:' ) for c in div_lst: NSTools.p( '\t\t', c ) # We recover the preimages of circles in the Roman surface # under the map p_lst, by constructing for each candidate # class the corresponding linear series. # 2e0-e1-e2-e3-e4-e5-e6-e7-e8 b = [( a0 - 1, -a0 ), ( -a0, a0 - 1 )] b += [( -a0 + 1, a0 ), ( a0, -a0 + 1 )] b += [ ( a0 - 1, a0 ), ( -a0, -a0 + 1 )] b += [( -a0 + 1, -a0 ), ( a0, a0 - 1 )] bp_tree = BasePointTree() for i in range( 6 ): bp_tree.add( 'z', b[i], 1 ) NSTools.p( 'basepoints =', b ) NSTools.p( LinearSeries.get( [2], bp_tree ) ) # e0-e1-e2 b = [( a0 - 1, -a0 ), ( -a0, a0 - 1 )] bp_tree = BasePointTree() bp = bp_tree.add( 'z', b[0], 1 ) bp = bp_tree.add( 'z', b[1] , 1 ) NSTools.p( 'basepoints =', b ) NSTools.p( LinearSeries.get( [1], bp_tree ) ) # e0-e3-e4 b = [( -a0 + 1, a0 ), ( a0, -a0 + 1 )] bp_tree = BasePointTree() bp = bp_tree.add( 'z', b[0], 1 ) bp = bp_tree.add( 'z', b[1] , 1 ) NSTools.p( 'basepoints =', b ) NSTools.p( LinearSeries.get( [1], bp_tree ) ) # e0-e5-e6 b = [ ( a0 - 1, a0 ), ( -a0, -a0 + 1 )] bp_tree = BasePointTree() bp = bp_tree.add( 'z', b[0], 1 ) bp = bp_tree.add( 'z', b[1] , 1 ) NSTools.p( 'basepoints =', b ) NSTools.p( LinearSeries.get( [1], bp_tree ) ) # e0-e7-e8 b = [( -a0 + 1, -a0 ), ( a0, a0 - 1 )] bp_tree = BasePointTree() bp = bp_tree.add( 'z', b[0], 1 ) bp = bp_tree.add( 'z', b[1] , 1 ) NSTools.p( 'basepoints =', b ) NSTools.p( LinearSeries.get( [1], bp_tree ) ) return if __name__ == '__main__': # Debug output settings # mod_lst = [] mod_lst += ['__main__.py'] # mod_lst += ['class_dp_lattice.py'] # mod_lst += ['class_eta.py'] NSTools.filter( mod_lst ) # output only from specified modules NSTools.filter( None ) # print all verbose output, comment to disable. # NSTools.get_tool_dct().clear() # uncomment to remove all cache! if 'OUTPUT_PATH' not in os.environ: os.environ['OUTPUT_PATH'] = './' NSTools.start_timer() # # Should be between 3 and 9. # computes classifications up to rank "max_rank". # max_rank = 9 ######################################### # # # (Un)comment one or more use cases # # # ######################################### usecase__get_cls( max_rank ) usecase__get_classes_dp1( max_rank ) usecase__graphs( max_rank ) usecase__analyze_graphs( max_rank ) usecase__construct_surfaces() # usecase__roman_circles() # takes about 3 minutes ######################################### # # # End of list of use case methods. # # # ######################################### NSTools.end_timer() NSTools.p( 'The End' )
17,992
33.601923
140
py
ns_lattice
ns_lattice-master/ns_lattice/src/ns_lattice/convert_to_tex.py
''' Use of this source code is governed by a MIT-style license that can be found in the LICENSE file. Created on Dec 14, 2017 @author: Niels Lubbes ''' from ns_lattice.class_div import Div from ns_lattice.class_dp_lattice import DPLattice from ns_lattice.class_ns_tools import NSTools from ns_lattice.sage_interface import sage_identity_matrix def cls_to_tex(): ''' Create tex code for the output of DPLattice.get_cls() Returns ------- string A string representing a table of tables in Tex format. The table represent the classification of Neron-Severi lattice of weak del Pezzo surfaces. ''' # create a list of occuring divisors # div_lst = [] for rank in range( 3, 9 + 1 ): for dpl in DPLattice.get_cls( rank ): # construct list for involution (e0,...,er)|-->(i0,...,ir) i_lst = [Div( row ).mat_mul( dpl.M ) for row in sage_identity_matrix( rank ) ] # add each divisor that occurs to div_lst for elt in i_lst + dpl.d_lst: div_lst += [ Div( elt.e_lst + ( 9 - len( elt.e_lst ) ) * [0] ) ] div_lst = list( set( div_lst ) ) div_lst.sort() e0 = Div( [1, 0, 0, 0, 0, 0, 0, 0, 0 ] ) div_lst.remove( e0 ) div_lst = [e0] + div_lst # create dictionary characters for elements in div_lst # abc = 'abcdefghijklmnopqrstuvwxyz' ch_lst = [] ch_lst += [ '\\frac{' + ch1 + '}{' + ch2 + '}\\!' for ch1 in '0123456789' for ch2 in '0123456789' ] ch_lst += [ '\\frac{' + ch1 + '}{' + ch2 + '}\\!' for ch1 in '0123456789' for ch2 in 'abcdef' ] NSTools.p( '(len(ch_lst), len(div_lst)) =', ( len( ch_lst ), len( div_lst ) ) ) assert len( ch_lst ) >= len( div_lst ) # create legend and dictionary # lgd_lst = [] sym_dct = {} for i in range( len( div_lst ) ): sym_dct.update( {str( div_lst[i] ):ch_lst[i]} ) lgd_lst += [['$' + ch_lst[i] + '$ :', ( '$' + str( div_lst[i] ) + '$' ).replace( 'e', 'e_' ) ]] while len( lgd_lst ) % 3 != 0: lgd_lst += [['', '']] nnrows = len( lgd_lst ) / 3 # create tex for legend # tex_lgd = '' tex_lgd += '\\begin{table}\n' tex_lgd += '\\setstretch{1.4}\n' tex_lgd += '\\tiny\n' tex_lgd += '\\caption{Classification of Neron-Severi lattices of weak del Pezzo surfaces (see THM{nsl})}\n' tex_lgd += '\\label{tab:nsl}\n' tex_lgd += 'A dictionary for symbols in the columns $\\sigma_A$ and $B$:\n\\\\\n' tex_lgd += '\\begin{tabular}{@{}l@{}l@{~~~~}l@{}l@{~~~~}l@{}l@{}}\n' for idx in range( nnrows ): c1, c2, c3, c4, c5, c6 = lgd_lst[idx] + lgd_lst[idx + nnrows] + lgd_lst[idx + 2 * nnrows] tex_lgd += c1 + ' & ' + c2 + ' & ' + c3 + ' & ' + c4 + ' & ' + c5 + ' & ' + c6 tex_lgd += '\\\\\n' tex_lgd += '\\end{tabular}\n' tex_lgd += '\\end{table}\n\n' # number of rows of the big table # nrows = 57 # dictionary for replacing string symbols # rep_dct = {'A':'A_', 'D':'D_', 'E':'E_', '{':'\\underline{', '[':'\\udot{', ']':'}'} # create classification table # tab_lst = [] # rank 1 and 2 tab9 = [['i ', '$9$', "$A_0 $", '$A_0$', '$0$', '$1$', '']] tab8 = [['ii ', '$8$', "$A_0 $", '$A_0$', '$0$', '$2$', ''], ['iii', '$8$', "$A_0 $", '$A_0$', '$0$', '$1$', ''], ['iv ', '$8$', "$A_0 $", '$A_0$', '$1$', '$1$', ''], ['v ', '$8$', "$A_0 $", '$A_1$', '$0$', '$1$', '']] tab_lst += [ tab9, tab8 ] # rank 3,4,5,6,7,8 and 9 idx = 0 Mtype_lst = ['A1', '4A1'] # for breaking up table for degree 2 case for rank in range( 3, 9 + 1 ): tab = [] for dpl in DPLattice.get_cls( rank ): col1 = '$' + str( idx ) + '$' col2 = '$' + str( dpl.get_degree() ) + '$' col3 = '$' + str( dpl.get_marked_Mtype() ) + '$' for key in rep_dct: col3 = str( col3 ).replace( key, rep_dct[key] ) col4 = '$' + str( dpl.get_real_type() ) + '$' for key in rep_dct: col4 = str( col4 ).replace( key, rep_dct[key] ) col5 = '$' + str( dpl.get_numbers()[4] ) + '$' col6 = '$' + str( dpl.get_numbers()[5] ) + '$' i_lst = [ str( Div( rw ).mat_mul( dpl.M ) ) for rw in sage_identity_matrix( rank ) ] col7 = '' for i in i_lst: col7 += sym_dct[i] if col7 in ['012', '0123', '01234', '012345', '0123456', '01234567', '012345678']: col7 = '' col8 = '' for d in dpl.d_lst: col8 += sym_dct[str( d )] # these subroot systems cannot be realized as weak del Pezzo surfaces if col4 in ['$7\underline{A_1}$', '$8\underline{A_1}$', '$4\underline{A_1}+\underline{D_4}$']: col1 = '$\\times$' # break (sometimes) the table for degree 2 according to Mtype if dpl.get_degree() == 2 and dpl.Mtype in Mtype_lst: nheaders = len( tab ) / nrows # each header shifts the row number while len( tab ) % nrows != nrows - 1 - nheaders: # add rows until end of table tab += [7 * ['']] Mtype_lst.remove( dpl.Mtype ) # add row tab += [[col1, col2, col3, col4, col5, col6, '$' + col7 + '||\!' + col8 + '$' ]] idx += 1 tab_lst += [ tab ] # reformat table # # i d A B E G Ac%Bc hl = '@{~}l@{~~~}l@{~~~}l@{~~}l@{~~}l@{~~}l@{~~}l@{}' hrow = ['', 'd', '$D(A)$', '$D(B)$', '$\#E$', '$\#G$', '$\sigma_A||B$'] etab_lst = [] etab = [hrow] tab_idx = 0 for tab in tab_lst: for row in tab: if len( etab ) >= nrows: etab_lst += [etab] etab = [hrow] etab += [row] if len( etab ) < nrows and tab_idx <= 3: etab += [7 * [''], 7 * ['']] # add two empty rows to separate tables with different rank else: while len( etab ) < nrows: etab += [7 * ['']] # add empty rows to fill up table etab_lst += [etab] etab = [hrow] tab_idx += 1 NSTools.p( 'etab_lst: ', [len( etab ) for etab in etab_lst] ) # create tex for main classification table # tex_tab = '' tab_idx = 0 for etab in etab_lst: if tab_idx % 2 == 0: tex_tab += '\\begin{table}\n' tex_tab += '\\setstretch{1.6}\n' tex_tab += '\\centering\\tiny\n' tex_tab += '\\begin{tabular}{@{}l@{\\hspace{1cm}}l@{}}\n' elif tab_idx % 2 == 1: tex_tab += '&\n' tex_tab += '\\begin{tabular}{' + hl + '}\n' for row in etab: col1, col2, col3, col4, col5, col6, col78 = row tex_tab += col1 + ' & ' + col2 + ' & ' + col3 + ' & ' + col4 + ' & ' tex_tab += col5 + ' & ' + col6 + ' & ' + col78 tex_tab += ' \\\\\n' if row == hrow: tex_tab += '\\hline\n' tex_tab += '\\end{tabular}\n' if tab_idx % 2 == 1: tex_tab += '\\end{tabular}\n' tex_tab += '\\end{table}\n\n' tab_idx += 1 if tab_idx % 2 == 1: tex_tab += '&\n' tex_tab += '\\end{tabular}\n\n' # creating tex for commands tex_cmd = '' tex_cmd += '\\newcommand{\\udot}[1]{\\tikz[baseline=(todotted.base)]{\\node[inner sep=1pt,outer sep=0pt] (todotted) {$#1$};\\draw[densely dotted] (todotted.south west) -- (todotted.south east);}}' tex_cmd += '\n' tex_cmd += '\\newcommand{\\udash}[1]{\\tikz[baseline=(todotted.base)]{\\node[inner sep=1pt,outer sep=0pt] (todotted) {$#1$};\\draw[densely dashed] (todotted.south west) -- (todotted.south east);}}' tex_cmd += '\n\n' out = tex_cmd + tex_lgd + tex_tab return out
7,948
32.682203
201
py
ns_lattice
ns_lattice-master/ns_lattice/src/ns_lattice/class_div.py
''' Created on Aug 11, 2016 @author: Niels Lubbes ''' from ns_lattice.sage_interface import sage_ZZ from ns_lattice.sage_interface import sage_diagonal_matrix from ns_lattice.sage_interface import sage_vector class Div: '''Element in Neron-Severi lattice. The class represents a divisor class in the Neron-Severi lattice with respect to the standard basis: <e0,e1,e2,...> Attributes ---------- e_lst : list<sage_ZZ> List describes a divisor in terms of the standard basis. int_mat : sage_matrix<sage_ZZ> A matrix over ZZ of rank "len(e_lst)" represents the unimodular intersection product for the divisor. ''' # static variable # short_output = True # static list of intersection matrices # int_mat_lst = [] def __init__( self, e_lst = 9 * [0], int_mat = None ): ''' Return ------ Div Constructor (called when instantiating object). If "int_mat==None" then the default diagonal matrix has signature (+-...-). This matrix determines the intersection product of divisors. ''' self.e_lst = list( e_lst ) # # equal "self.int_mat" for each instantiated Div object references # to a unique matrix, so that no new matrix is instantiated for each # Div object. Maybe this is already ensured by Sage library, but just # to be on the safe side. # if int_mat == None: int_mat = sage_diagonal_matrix( sage_ZZ, [1] + ( self.rank() - 1 ) * [-1] ) if int_mat not in Div.int_mat_lst: Div.int_mat_lst += [int_mat] idx = Div.int_mat_lst.index( int_mat ) self.int_mat = Div.int_mat_lst[idx] @staticmethod def new( lbl, rank = 9 ): ''' Parameters ---------- lbl : string A string with format as output of "self.get_label()". rank : int Integer representing rank of Neron-Severi lattice in which "Div" lives. Returns ------- The "Div" corresponding to the label such that "len(self.e_lst)>=rank". ''' c = Div( rank * [0] ) # zero divisor class if 'e' in lbl: s = lbl if 'e0' in s: # cases: 'e0...', '-e0...', '3e0...' or '-2e0...' if s[0:2] == 'e0': c.e_lst = [1] s = s[2:] elif s[0:3] == '-e0': c.e_lst = [-1] s = s[3:] else: # '3e0...' or '-2e0...' c.e_lst = [ int( s.split( 'e0' )[0] ) ] # [4] if lbl='4h+3e...' s = s.split( 'e0' )[1] # for example '+3e2-2e5+6e7+e8' else: c.e_lst = [0] s = lbl coef_e = '' idx = 0 last_i = 0 # ei while idx < len( s ): if s[idx] != 'e': coef_e += s[idx] idx += 1 elif s[idx] == 'e': coef_i = '' idx += 1 while idx < len( s ) and s[idx] not in ['+', '-']: coef_i += s[idx] idx += 1 i = int( coef_i ) if coef_e == '-': coef_e = '-1' if coef_e in ['+', '']: coef_e = '1' c.e_lst += ( i - last_i - 1 ) * [0] + [int( coef_e )] coef_e = '' last_i = i else: # label of (-2)-class if rank > 9: raise ValueError( 'For (-2)-classes we expect the rank to be at most 9: ', rank ) # check whether the label is negative if lbl[0] == '-': neg = True lbl = lbl[1:] else: neg = False # '12' ---> e1-e2 if len( lbl ) == 2: c.e_lst[ int( lbl[0] ) ] = 1 c.e_lst[ int( lbl[1] ) ] = -1 # '1123' ---> e0-e1-e2-e3 elif len( lbl ) == 4 and lbl[0] == '1': c.e_lst[0] = int( lbl[0] ) c.e_lst[ int( lbl[1] ) ] = -1 c.e_lst[ int( lbl[2] ) ] = -1 c.e_lst[ int( lbl[3] ) ] = -1 # '212' ---> 2e0-e3-e4-...-e8 elif len( lbl ) == 3 and lbl[0] == '2': c.e_lst = 9 * [-1] c.e_lst[0] = int( lbl[0] ) c.e_lst[ int( lbl[1] ) ] = 0 c.e_lst[ int( lbl[2] ) ] = 0 if rank != 9 and set( c.e_lst[rank:] ) != set( [0] ): raise ValueError( 'Rank too low for label: ', rank, lbl ) c.e_lst = c.e_lst[:rank] # '308' ---> 3e0-e1-e2-...-e7-2e8 elif len( lbl ) == 3 and lbl[0] == '3' and lbl[1] == '0': c.e_lst = 9 * [-1] c.e_lst[0] = int( lbl[0] ) c.e_lst[ int( lbl[2] ) ] = -2 else: # unknown label raise ValueError( 'Label has incorrect format: ', lbl ) # for example '-12'=[0,-1,1,0,0,...] if neg: c.e_lst = [ -e for e in c.e_lst ] # end handling label of (-2)-class # update rank c.e_lst = c.e_lst + ( rank - len( c.e_lst ) ) * [0] return c def rank( self ): return len( self.e_lst ) def is_positive( self ): ''' Returns ------- bool Return True iff the first nonzero entry of the self.e_lst is positive. The zero divisor is also positive. ''' for e in self.e_lst: if e != 0: return e > 0 return True @staticmethod def get_min_rank( lbl ): ''' Parameters ---------- lbl : string A string with format as output of "self.get_label()". Returns ------- int The minimal rank of the "Div" object with a given label. Examples -------- >>> get_min_rank('78') 9 >>> get_min_rank('301') 9 >>> get_min_rank('12') 3 ''' d = Div.new( lbl ) lst = [ e for e in d.e_lst ] while lst[-1] == 0 and lst != []: lst.pop() return len( lst ) def get_basis_change( self, B ): ''' Parameters ---------- B : sage_matrix A matrix whose rows correspond to generators of a new basis. We assume that the intersection matrix for this basis is the default diagonal matrix with diagonal (1,-1,...,-1). Returns ------- Div A new "Div" object, which represents the current divisor with respect to a new basis. ''' new_int_mat = B * self.int_mat * B.T new_e_lst = self.mat_mul( ~( B.T ) ).e_lst return Div( new_e_lst, new_int_mat ) def __get_minus_two_label( self ): ''' Private helper method for "get_label()" Parameters ---------- self : Div self*self==-2 and self.rank<=9. Returns ------- string See output documents for self.get_label() ''' if self * self != -2 or self.rank() > 9: raise ValueError( 'Unexpected input for __get_mt_label: ', self.e_lst ) # first non-zero coefficient negative? neg = [e < 0 for e in self.e_lst if e != 0][0] # check whether the label should start with minus symbol if neg: tmp = [-e for e in self.e_lst] else: tmp = self.e_lst # set of non-zero coefficients for ei. oset = set( [ e for e in tmp[1:] if e != 0 ] ) # e1-e2 ---> '12' if tmp[0] == 0 and oset == set( [1, -1] ): lbl = '' for i in range( 1, len( tmp ) ): if tmp[i] != 0: lbl += str( i ) # e0-e1-e2-e3 ---> '1123' elif tmp[0] == 1 and oset == set( 3 * [-1] ): lbl = '1' for i in range( 1, len( tmp ) ): if tmp[i] != 0: lbl += str( i ) # 2e0-e3-e4-...-e8 ---> '212' elif tmp[0] == 2 and oset == set( 6 * [-1 ] ): lbl = '2' for i in range( 1, len( tmp ) ): if tmp[i] == 0: lbl += str( i ) # 3e0-e1-e2-...-e7-2e8 ---> '308' elif tmp[0] == 3 and oset == set( 7 * [-1 ] + [-2] ): lbl = '30' for i in range( 1, len( tmp ) ): if tmp[i] == -2: lbl += str( i ) if neg: lbl = '-' + lbl # for example: 12 --> -12 return lbl def get_abbr_label( self ): ''' Returns ------- string We describe the output label in terms of examples. > e1 ---> 'e1' > e1-e2 ---> 'e12' > 2e0-e1-e2-e4-e5 ---> '2e1245' > e0-e1 ---> '1e1' This options only works for special cases. The cases which are covered are (-1)- and (-2)-classes, and classes of conical families on weak Del Pezzo surfaces, with respect to the basis with intersection product defined by the diagonal matrix with diagonal (1,-1,...,-1). ''' np1 = len( [e for e in self.e_lst[1:] if e == 1] ) nm1 = len( [e for e in self.e_lst[1:] if e == -1] ) n01 = len( [e for e in self.e_lst[1:] if e > 1 or e < -1] ) if n01 == 0 and self[0] in range( 0, 10 ): # e1 if self[0] == 0 and np1 == 1 and nm1 == 0: return 'e' + str( self.e_lst.index( 1 ) ) # e1-e2 if self[0] == 0 and np1 == 1 and nm1 == 1: return 'e' + str( self.e_lst.index( 1 ) ) + str( self.e_lst.index( -1 ) ) # 2h-e1-e2-e3-e4-e5 or h-e1 if self[0] in range( 0, 10 ) and np1 == 0 and nm1 > 0: lbl = str( self[0] ) + 'e' for i in range( 1, len( self.e_lst ) ): if self[i] != 0: lbl += str( i ) return lbl raise ValueError( 'Input is not treated by this function (use get_label() instead):', self.e_lst ) def get_label( self, abbr = False ): ''' Parameters ---------- abbr : boolean Returns ------- string We describe the output label in terms of examples. If "abbr==True" and self*self==-2 and self.rank()<=9: > e1-e2 ---> '12' > -e1+e2 ---> '-12' > e0-e1-e2-e3 ---> '1123' > 2e0-e3-e4-...-e8 ---> '212' > 3e0-e1-e2-...-e7-2e8 ---> '308' > -3e0+e1+e2+...+e7+2e8 ---> '-308' For the remaining cases not treated above: > 3e0-2e1-13e2-4e3 ---> '3e0-2e1-13e2-4e3' ''' divK = Div( [-3] + ( self.rank() - 1 ) * [1] ) # treat cases for (-2)-label # if abbr and self * self == -2 and self.rank() <= 9 and self * divK == 0: return self.__get_minus_two_label() # from this point on we treat the general case # lbl = '' for i in range( 0, len( self.e_lst ) ): val = self[i] if val != 0: if val == 1: if lbl != '': lbl += '+' elif val == -1: lbl += '-' else: if val > 1 and lbl != '': lbl += '+' lbl += str( val ) lbl += 'e' + str( i ) return lbl def mat_mul( self, M ): ''' Parameters ---------- M : sage_matrix A matrix with self.rank() columns. Returns ------- Div Returns a "Div" object that is a result of applying the linear transformation corresponding to "M" to itself. ''' v = sage_vector( self.e_lst ).column() return Div( ( M * v ).list() ) def int_mul( self, n ): ''' Parameters ---------- n : int Returns ------- Div Returns a "Div" object that is a result of multiplying with the scalar "n". ''' return self.mat_mul( sage_diagonal_matrix( self.rank() * [n] ) ) # operator overloading for == def __eq__( self, other ): return self.e_lst == other.e_lst # operator overloading for != def __ne__( self, other ): return not self.__eq__( other ) # operator overloading for < # Used for sorting lists of "Div"-objects: # <http://stackoverflow.com/questions/1227121/compare-object-instances-for-equality-by-their-attributes-in-python> def __lt__( self, other ): ''' Parameters ---------- other : Div Returns ------- bool Here are some examples to explain the ordering we use for div classes e1 < e2 e0-e1-e2 < e0-e1-e3 1123 < 308 1123 < 1124 12 < 1123 12 < 13 12 < 34 ''' if self.rank() != other.rank(): return self.rank() < other.rank() a = self.e_lst b = other.e_lst if sum( a ) == sum( b ) == 1 and set( a ) == set( b ) == {0, 1}: return b < a # e1 < e2 a = [a[0]] + [ -elt for elt in reversed( a[1:] )] b = [b[0]] + [ -elt for elt in reversed( b[1:] )] return a < b # lexicographic order # operator overloading for * def __mul__( self, div ): ''' Parameters ---------- div : Div Returns ------- Div The intersection product of "self" and "div" wrt. to matrix "self.int_mat". ''' row_vec = sage_vector( sage_ZZ, self.e_lst ).row() col_vec = sage_vector( sage_ZZ, div.e_lst ).column() mat = self.int_mat v = row_vec * mat * col_vec return v[0][0] # operator overload for + def __add__( self, div ): v = sage_vector( sage_ZZ, self.e_lst ) + sage_vector( sage_ZZ, div.e_lst ) return Div( list( v ) ) # operator overload for - def __sub__( self, div ): v = sage_vector( sage_ZZ, self.e_lst ) - sage_vector( sage_ZZ, div.e_lst ) return Div( list( v ) ) # operator overloading for [] def __getitem__( self, index ): return self.e_lst[index] # operator overloading for [] def __setitem__( self, index, item ): self.e_lst[index] = item # overloading for str(.): human readable string representation of object def __str__( self ): if Div.short_output: return self.get_label() else: return str( self.e_lst ) # overloading "__repr__()" as well, since python call this for Div objects in a list def __repr__( self ): return self.__str__() # so that lists of this object can be used with set() def __hash__( self ): return hash( self.__str__() + '__' + str( self.rank() ) )
16,148
27.683837
122
py
ns_lattice
ns_lattice-master/ns_lattice/src/ns_lattice/class_dp_lattice.py
''' Created on Aug 15, 2016 @author: Niels Lubbes This module is for classifying real structures and singularities of weak Del Pezzo surfaces of degree between 1 and 7. ''' import time from ns_lattice.sage_interface import sage_identity_matrix from ns_lattice.sage_interface import sage_ZZ from ns_lattice.sage_interface import sage_QQ from ns_lattice.sage_interface import sage_Subsets from ns_lattice.sage_interface import sage_VectorSpace from ns_lattice.sage_interface import sage_vector from ns_lattice.sage_interface import sage_Graph from ns_lattice.div_in_lattice import get_divs from ns_lattice.div_in_lattice import get_indecomp_divs from ns_lattice.div_in_lattice import get_ak from ns_lattice.dp_root_bases import get_graph from ns_lattice.dp_root_bases import get_ext_graph from ns_lattice.dp_root_bases import get_dynkin_type from ns_lattice.dp_root_bases import convert_type from ns_lattice.dp_root_bases import get_root_bases_orbit from ns_lattice.dp_root_bases import is_root_basis from ns_lattice.dp_involutions import basis_to_involution from ns_lattice.dp_involutions import is_integral_involution from ns_lattice.class_ns_tools import NSTools from ns_lattice.class_div import Div from ns_lattice.class_eta import ETA class DPLattice: ''' Represents an equivalence class of the Neron-Severi lattice of a real weak del Pezzo surface, together with an involution "M" and a set of effective (-2)-classes "d_lst". The effective (-2)-classes form the basis of a root system. ( ZZ<e0,e1,...,er>, M, d_lst ) From these objects it is possible to compute the remaining attributes of this class. If <e0,e1,...,er> is a basis for the Neron-Severi lattice of the projective plane P^2 blown up in r points then the the canonical class k equals k=-3e0+e1+...+er. The intersection product is in this case -e0^2=e1^2=...=er^2=-1 with remaining intersections zero. Otherwise if <e0,e1,...,er> is a basis for the Neron-Severi lattice of the P^1xP^1 blown up in r points then the the canonical class k equals k=-2*(e0+e1). The intersection product is in this case -h*e1=e2^2=...=er^2=-1 with remaining intersections zero. Attributes ---------- M : sage_matrix<sage_ZZ> A matrix which correspond to an involution of the lattice <e0,e1,...,er> with r=rank-1 and 2 <= r <= 8. Md_lst : list<Div> A list of "Div" objects that correspond to the eigenvectors of eigenvalue 1 of M. These "Div" objects form a basis of a root subsystem. Mtype : string A String that denotes the Dynkin type of "Md_lst". d_lst : list<Div> A list of "Div" objects d such that d*d==-2 and d*k=0 where k denotes the canonical class. These elements represent effective (-2)-classes. type : string A String that denotes the Dynkin type of "d_lst". m1_lst : list<Div> A list of "Div" objects "m" such that m*m==-1==m*k and m*d>=0 for all d in d_lst, where k denotes the canonical class. These elements represent (-1)-classes that cannot be written as the sum of two effective classes. In other words, the classes are indecomposable. fam_lst : list<Div> A list of "Div" objects "f" such that f*f==0, f*(-k)==2 and m*d>=0 for all d in d_lst, where k denotes the canonical class. real_d_lst : list<Div> A list "Div" objects that represent indecomposable and real (-2)-classes. Thus these classes are send to itself by M. Geometrically these classes correspond to real isolated singularities. real_m1_lst : list<Div> A list "Div" objects that represent indecomposable and real (-1)-classes. Thus these classes are send to itself by M. Geometrically these classes correspond to real lines. real_fam_lst : list<Div> A list "Div" objects that represent real classes in "self.fam_lst". Thus these classes are send to itself by M. Geometrically these classes correspond to a real families of conics. or_lst : list<Div> A list of "Div" objects that represents roots that are orthogonal to "self.d_lst". sr_lst : list<Div> A list of "Div" objects that represents roots that are contained in the subspace spanned by "self.d_lst". G : sage_GRAPH The Cremona invariant for the current lattice. SG : sage_GRAPH Simple family graph (see self.get_SG()). SG_data : [int, int, list<int>, list<int>, bool, bool, bool, bool ] A list of of data that characterizes the simple family graph (see self.get_SG()). ''' def __init__( self, d_lst, Md_lst, M ): ''' Constructor. Returns ------- DPLattice A DPLattice class whose attributes are set according to input: * DPLattice.M * DPLattice.Md_lst * DPLattice.d_lst The remaining attributes of DPLattice can be computed from these attributes. In order for this object to make sense, it is required that the involution "M" preserves "d_lst" as a set. Geometrically this means that the involution sends isolated singularities to isolated singularities. ''' self.d_lst = d_lst self.Md_lst = Md_lst self.M = M self.m1_lst = None self.fam_lst = None self.real_d_lst = None self.real_m1_lst = None self.real_fam_lst = None self.Mtype = None self.type = None self.or_lst = None self.sr_lst = None self.G = None self.SG = None self.SG_data = None def set_attributes( self, level = 9 ): ''' Sets attributes of this object, depending on the input level. For constructing a classification we instantiate many DPLattice objects. This method allows us to minimize the number of attributes that computed (thus we use lazy evaluation). Parameter --------- self: DPLattice At least self.M, self.Md_lst and self.d_lst should be initialized. level : int A non-negative number. ''' # M, Md_lst and d_lst are set. if self.m1_lst == None: all_m1_lst = get_divs( get_ak( self.get_rank() ), 1, -1, True ) self.m1_lst = get_indecomp_divs( all_m1_lst, self.d_lst ) if level < 1: return if self.fam_lst == None: all_fam_lst = get_divs( get_ak( self.get_rank() ), 2, 0, True ) self.fam_lst = get_indecomp_divs( all_fam_lst, self.d_lst ) if level < 2: return if self.real_d_lst == None: self.real_d_lst = [ d for d in self.d_lst if d.mat_mul( self.M ) == d ] if level < 3: return if self.real_m1_lst == None: self.real_m1_lst = [ m1 for m1 in self.m1_lst if m1.mat_mul( self.M ) == m1 ] if level < 4: return if self.real_fam_lst == None: self.real_fam_lst = [ f for f in self.fam_lst if f.mat_mul( self.M ) == f ] if level < 5: return if self.or_lst == None: self.or_lst = [] for m2 in get_divs( get_ak( self.get_rank() ), 0, -2, True ): if [m2 * d for d in self.d_lst] == len( self.d_lst ) * [0]: self.or_lst += [m2] if level < 6: return if self.sr_lst == None: V = sage_VectorSpace( sage_QQ, self.get_rank() ) W = V.subspace( [d.e_lst for d in self.d_lst] ) self.sr_lst = [] for m2 in get_divs( get_ak( self.get_rank() ), 0, -2, True ): if sage_vector( m2.e_lst ) in W: self.sr_lst += [ m2 ] if level < 7: return if self.type == None: self.type = get_dynkin_type( self.d_lst ) if level < 8: return if self.Mtype == None: self.Mtype = get_dynkin_type( self.Md_lst ) if level < 9: return if self.G == None: self.G = get_ext_graph( self.d_lst + self.m1_lst, self.M ) def get_rank( self ): ''' Parameters ---------- self : DPLattice We expect self.M != None. Returns ------- int Integer denoting rank of lattice. ''' return self.M.dimensions()[0] def get_degree( self ): ''' Parameters ---------- self : DPLattice We expect self.M != None. Returns ------- int Integer denoting the degree of weak del Pezzo surface with "self" its corresponding Neron-Severi lattice. ''' return 10 - self.get_rank() def get_numbers( self ): ''' Parameters ---------- self : DPLattice Returns ------- list<int> List of 6 integers: 0: #indecomposable (-2)-classes 1: #indecomposable (-1)-classes 2: #families of conics 3: #real effective (-2)-classes 4: #real indecomposable (-1)-classes 5: #real families of conics where # stands for number of. Note that a divisor class is indecomposable if it is effective and cannot be written as the sum of two effective classes. ''' self.set_attributes( 6 ) return ( len( self.d_lst ), len( self.m1_lst ), len( self.fam_lst ), len( self.real_d_lst ), len( self.real_m1_lst ), len( self.real_fam_lst ) ) def contains_fam_pair( self ): ''' Parameters ---------- self : DPLattice Returns ------- bool True if self.real_fam_lst contains two Div classes with intersection one. Geometrically this means that a weak del Pezzo surface with a Neron-Severi lattice that is isomorphic to this one, must be birational to P1xP1 (ie. fiber product of the projective line with itself). ''' self.set_attributes( 6 ) for f1 in self.real_fam_lst: for f2 in self.real_fam_lst: if f1 * f2 == 1: return True return False def is_real_minimal( self ): ''' Parameters ---------- self : DPLattice Returns ------- bool True if self.m1_lst does not contain classes u and v such that either * u.mat_mul( self.M ) == v and u*v==0, or * u.mat_mul( self.M ) == u. This means that self is the DPLattice of a real-minimal weak del Pezzo surface. Thus no disjoint complex conjugate exceptional curves or real exceptional curves can be contracted. ''' self.set_attributes( 0 ) for u in self.m1_lst: v = u.mat_mul( self.M ) if v * u == 0 or v == u: return False return True def get_marked_Mtype( self ): ''' We mark Mtype with a '-symbol to distinguish between real structures of the same Dynkin type that are not conjugate. ''' if self.get_degree() not in [6, 4, 2]: return self.Mtype self.set_attributes( 8 ) if ( self.get_degree(), self.Mtype ) not in [ ( 6, 'A1' ), ( 4, '2A1' ), ( 2, '3A1' ) ]: return self.Mtype mark = '' if list( self.M.T[0] ) != [1] + ( self.get_rank() - 1 ) * [0]: # in this case e0 is not send to e0 by the involution self.M mark = "'" return self.Mtype + mark def get_real_type( self ): ''' Gets the Dynkin type (self.type) of self.d_lst. The components of the Dynkin diagram that are preserved by the involution induced by the real structure are marked. For example, {A2} means that the elements in the root bases for the A2 root systems are preserved elementwise by the involution. We write [A2] if the root bases is preserved by the involution as a whole but not element wise. We write 2A2 if the two A2 root bases are interchanged by the involution. Instead of 3A2 we may write for example [A2]+{A2}+2A2. Returns ------- string Dynkin types of components ''' comp_lst = get_graph( self.d_lst ).connected_components() comp_lst.reverse() # smaller components first if comp_lst == []: return 'A0' # construct list of types type_lst = [] for comp in comp_lst: c_lst = [ self.d_lst[i] for i in comp ] mc_lst = [] elementwise = True for c in c_lst: mc = c.mat_mul( self.M ) mc_lst += [mc] if c != mc: elementwise = False mc_lst.sort() type = get_dynkin_type( c_lst ) if set( mc_lst ) == set( c_lst ) and c_lst != []: if elementwise: type_lst += ['{' + type + '}'] else: type_lst += ['[' + type + ']'] else: type_lst += [type] # construct string out = '' while type_lst != []: type = type_lst[0] num = type_lst.count( type ) if num != 1: out += str( num ) out += type + '+' type_lst = [ elt for elt in type_lst if elt != type ] out = out[:-1] # remove last plus return out def get_basis_change( self, B ): ''' Parameters ---------- self : DPLattice B : sage_matrix<sage_ZZ> A matrix whose rows correspond to generators of a new basis. We assume that the intersection matrix for this basis is the default diagonal matrix with diagonal (1,-1,...,-1). Returns ------- DPLattice A new "DPLattice" object, which represents the current lattice with respect to a new basis. ''' self.set_attributes( 6 ) d_lst_B = [ d.get_basis_change( B ) for d in self.d_lst ] Md_lst_B = [ Md.get_basis_change( B ) for Md in self.Md_lst ] M_B = ~( B.T ) * self.M * ( B.T ) # ~B is inverse of B, new involution after coordinate change dpl = DPLattice( d_lst_B, Md_lst_B, M_B ) dpl.Mtype = self.Mtype dpl.type = self.type dpl.m1_lst = [ m1.get_basis_change( B ) for m1 in self.m1_lst ] dpl.fam_lst = [ fam.get_basis_change( B ) for fam in self.fam_lst ] dpl.real_d_lst = [ d.get_basis_change( B ) for d in self.real_d_lst ] dpl.real_m1_lst = [ m1.get_basis_change( B ) for m1 in self.real_m1_lst ] dpl.real_fam_lst = [ fam.get_basis_change( B ) for fam in self.real_fam_lst ] return dpl def get_SG( self ): ''' The simple family graph associated to the Neron-Severi lattice of a weak del Pezzo surface is defined as the incidence diagram of self.real_fam_lst, with the edges labeled <=1 removed. All vertices are labeled with the index of the element in self.real_fam_lst. In the mathematical version (see arxiv paper) the vertices are labeled with the dimension of the linear series, which is always 1 with one exception: If len(self.real_fam_lst)==0 and rank==3, then the simple family graph consists of a single vertex labeled 2. Example ------- # The following graph is related to the E8 root system: # dpl = DPLattice.get_cls( 9 )[0] assert set(dpl.get_SG().num_verts()) == {2160} assert set(dpl.get_SG().get_degree()) == {2095} assert set(dpl.get_SG().edge_labels()) == {2,3,4,5,6,7,8} Returns ------- sage_GRAPH, [int, int, list<int>, list<int>, bool, bool, bool, bool ] The simple family graph self.SG and a list self.SG_data associated to the current DPLattice object. Here self.SG_data consists of data that describes self.SG. This method also initializes self.SG and self.SG_data. ''' if self.SG != None: return self.SG, self.SG_data if self.get_rank() == 9 and self.get_numbers()[-1] > 800: NSTools.p( 'Initializing simple family graph of current DPLattice object...', self.get_rank(), self.get_marked_Mtype(), self.get_real_type() ) f = self.real_fam_lst f_range = range( len( f ) ) self.SG = sage_Graph() self.SG.add_vertices( f_range ) for i in f_range: for j in f_range: if f[i] * f[j] > 1: self.SG.add_edge( i, j, f[i] * f[j] ) self.SG_data = [ self.SG.num_verts(), # number of vertices self.SG.num_edges(), # number of edges sorted( list( set( self.SG.degree() ) ) ), # possible numbers of outgoing edges sorted( list( set( self.SG.edge_labels() ) ) ), # possible edge labels self.SG.is_clique(), # True iff the graph is complete. self.SG.is_connected(), self.SG.is_vertex_transitive(), self.SG.is_edge_transitive()] return self.SG, self.SG_data @staticmethod def get_bas_lst( rank = 9 ): ''' See [Algorithm 5, http://arxiv.org/abs/1302.6678] for more info. Parameters ---------- rank : int An integer in [3,...,9]. Returns ------- list<DPLattice> A list of "DPLattice" objects dpl such that dpl.d_lst is the bases of a root subsystem and dpl.Mtype == A0. The list contains exactly one representative for all root subsystems up to equivalence. The list represents a classification of root subsystems of the root system with Dynkin type either: A1, A1+A2, A4, D5, E6, E7 or E8, corresponding to ranks 3, 4, 5, 6, 7, 8 and 9 respectively (eg. A1+A2 if rank equals 4, and E8 if rank equals 9). Note that the root systems live in a subspace of the vector space associated to the Neron-Severi lattice of a weak Del Pezzo surface. ''' # check whether classification of root bases is in cache key = 'get_bas_lst__' + str( rank ) if key in NSTools.get_tool_dct(): return NSTools.get_tool_dct()[key] NSTools.p( 'start' ) A = [ 12, 23, 34, 45, 56, 67, 78] B = [ 1123, 1145, 1456, 1567, 1678, 278 ] C = [ 1127, 1347, 1567, 234, 278, 308 ] D = [ 1123, 1345, 1156, 1258, 1367, 1247, 1468, 1178 ] dpl_lst = [] for ( lst1, lst2 ) in [ ( A, [] ), ( A, B ), ( A, C ), ( [], D ) ]: # restrict to divisors in list, that are of rank at most "max_rank" lst1 = [ Div.new( str( e ), rank ) for e in lst1 if rank >= Div.get_min_rank( str( e ) ) ] lst2 = [ Div.new( str( e ), rank ) for e in lst2 if rank >= Div.get_min_rank( str( e ) ) ] # the involution is trivial Md_lst = [] M = sage_identity_matrix( sage_QQ, rank ) # loop through the lists sub1 = sage_Subsets( range( len( lst1 ) ) ) sub2 = sage_Subsets( range( len( lst2 ) ) ) eta = ETA( len( sub1 ) * len( sub2 ), 20 ) for idx2_lst in sub2: for idx1_lst in sub1: eta.update( 'get_bas_lst rank =', rank ) d_lst = [ lst1[idx1] for idx1 in idx1_lst ] d_lst += [ lst2[idx2] for idx2 in idx2_lst ] if not is_root_basis( d_lst ): continue dpl = DPLattice( d_lst, Md_lst, M ) if dpl not in dpl_lst: dpl.set_attributes() dpl_lst += [dpl] # cache output dpl_lst.sort() NSTools.get_tool_dct()[key] = dpl_lst NSTools.save_tool_dct() return dpl_lst @staticmethod def get_inv_lst( rank = 9 ): ''' Outputs a list representing a classification of root subsystems that define unimodular involutions on the Neron-Severi lattice of a weak del Pezzo surface. We consider root subsystems of the root system with Dynkin type either: A1, A1+A2, A4, D5, E6, E7 or E8, corresponding to ranks 3, 4, 5, 6, 7, 8 and 9 respectively (eg. A1+A2 if rank equals 4, and E8 if rank equals 9). Note that root systems live in a subspace of the vector space associated to the Neron-Severi lattice of a weak Del Pezzo surface. Parameters ---------- max_rank : int An integer in [3,...,9]. Returns ------- list<DPLattice> A list of "DPLattice" objects dpl such that dpl.Md_lst is the bases of a root subsystem and dpl.type == A0. The list contains exactly one representative for root subsystems up to equivalence, so that the root subsystem defines a unimodular involution. ''' # check cache key = 'get_inv_lst__' + str( rank ) if False and key in NSTools.get_tool_dct(): return NSTools.get_tool_dct()[key] bas_lst = DPLattice.get_bas_lst( rank ) NSTools.p( 'rank =', rank ) amb_lst = [] inv_lst = [] eta = ETA( len( bas_lst ), 1 ) for bas in bas_lst: eta.update( bas.type ) M = basis_to_involution( bas.d_lst, rank ) if not is_integral_involution( M ): continue inv = DPLattice( [], bas.d_lst, M ) inv.set_attributes() NSTools.p( 'Found type of involution: ', bas.type ) # real structures with different Dynkin types may be equivalent if inv not in inv_lst: inv_lst += [ inv ] else: inv_prv = [inv2 for inv2 in inv_lst if inv == inv2][0] inv_lst = [inv2 for inv2 in inv_lst if not inv2 == inv] amb_lst += [inv, inv_prv] if inv > inv_prv: inv_lst += [inv] else: inv_lst += [inv_prv] NSTools.p( '\tAmbitious type:', inv.Mtype, '==', inv_prv.Mtype, ' inv>inv_prv: ', inv > inv_prv, ' ambitious types =', [ amb.Mtype for amb in amb_lst if amb == inv ] ) # store in cache inv_lst.sort() NSTools.get_tool_dct()[key] = inv_lst NSTools.save_tool_dct() return inv_lst @staticmethod def get_cls_slow( rank = 7 ): ''' Use get_cls_real_dp() for a faster method. This method does not terminate within reasonable time if rank>7. We still keep the method in order to compare the outcomes in case rank<=9. Parameters ---------- max_rank : int An integer in [3,...,9]. Returns ------- list<DPLattice> A list of DPLattice objects corresponding to Neron-Severi lattices of weak Del Pezzo surfaces of degree (10-rank). The list contains exactly one representative for each equivalence class. All the Div objects referenced in the DPLattice objects of the output have the default intersection matrix: diagonal matrix with diagonal: (1,-1,...,-1). ''' # check cache key = 'get_cls_slow__' + str( rank ) if key in NSTools.get_tool_dct(): return NSTools.get_tool_dct()[key] inv_lst = DPLattice.get_inv_lst( rank ) bas_lst = DPLattice.get_bas_lst( rank ) # we fix an involution up to equivalence and go through # all possible root bases for singularities. dpl_lst = [] eta = ETA( len( bas_lst ) * len( inv_lst ), 20 ) for inv in inv_lst: for bas in bas_lst: orbit_lst = get_root_bases_orbit( bas.d_lst ) eta.update( 'len( orbit_lst ) =', len( orbit_lst ) ) for d_lst in orbit_lst: # check whether involution inv.M preserves d_lst dm_lst = [ d.mat_mul( inv.M ) for d in d_lst ] dm_lst.sort() if dm_lst != d_lst: continue # add to classification if not equivalent to objects # in list, see "DPLattice.__eq__()". dpl = DPLattice( d_lst, inv.Md_lst, inv.M ) if dpl not in dpl_lst: dpl.set_attributes() dpl_lst += [dpl] # store in cache dpl_lst.sort() NSTools.get_tool_dct()[key] = dpl_lst NSTools.save_tool_dct() return dpl_lst @staticmethod def get_num_types( inv, bas, bas_lst ): ''' Returns the number of root bases in the eigenspace of eigenvalue 1 of the involution defined by M.inv. If this number is unknown, then -1 is returned. This method is used by get_cls() before calling seek_bases(). Parameters ---------- inv : DPLattice bas : DPLattice bas_lst : list<DPLattice> We expect this to be the output of get_bas_lst() Thus a list of inequivalent DPLattice objects Returns ------- int If there does not exists a DPLattice in bas_lst whose type is inv.Mtype and bas.type combined, then return 0. Otherwise return either -1 or the number of root bases in the eigenspace of eigenvalue 1 of the involution defined by M.inv. We expect this number to be at most 3. ''' # check whether the combined type exists in bas_lst t1_lst = convert_type( inv.Mtype ) t2_lst = convert_type( bas.type ) type_exists = False for bas2 in bas_lst: if sorted( t1_lst + t2_lst ) == convert_type( bas2.type ): type_exists = True break if not type_exists: return 0 # computes the roots in the eigenspace of eigenvalue 1 # of the involution defined by inv r_lst = get_divs( get_ak( inv.get_rank() ), 0, -2, True ) s_lst = [ r for r in r_lst if r.mat_mul( inv.M ) == r ] if len( s_lst ) == 30: # D6 since #roots=60=2*30 if bas.type in ['2A1', 'A3', '4A1', '2A1+A3', 'A5']: return 2 if bas.type in ['3A1', 'A1+A3']: return 3 return 1 if len( s_lst ) == 63: # E7 since #roots=126=2*63 if bas.type in ['3A1', '4A1', 'A5', 'A1+A3', '2A1+A3', 'A1+A5']: return 2 return 1 return -1 @staticmethod def get_part_roots( inv ): ''' Return two subsets of roots using the input involution. This method is used by get_cls(). Parameters ---------- inv : DPLattice We expect inv.type=='A0'. We will use inv.Mtype and inv.M. Returns ------- list<Div>, list<Div> Let R be defined by the list get_divs( get_ak( inv.get_rank() ), 0, -2, True ) whose elements are Div objects. If r is a Div object, then M(r) is shorthand notation for r.mat_mul(inv.M). The two returned lists correspond respectively to S := { r in R | M(r)=r } and Q union Q' := { r in R | M(r) not in {r,-r} and r*M(r)>0 } where Q = M(Q'). ''' r_lst = get_divs( get_ak( inv.get_rank() ), 0, -2, True ) s_lst = [ r for r in r_lst if r.mat_mul( inv.M ) == r ] tq1_lst = [ r for r in r_lst if r.mat_mul( inv.M ) not in [r, r.int_mul( -1 )] ] tq_lst = [ q for q in tq1_lst if q * q.mat_mul( inv.M ) >= 0 ] q_lst = [] for q in sorted( tq_lst ): if q not in q_lst and q.mat_mul( inv.M ) not in q_lst: q_lst += [q] # q_lst += [ q.int_mul( -1 ) for q in q_lst ] NSTools.p( 'r_lst =', len( r_lst ), r_lst ) NSTools.p( 's_lst =', len( s_lst ), s_lst ) NSTools.p( 'tq1_lst =', len( tq1_lst ), tq1_lst ) NSTools.p( 'tq_lst =', len( tq_lst ), tq_lst ) NSTools.p( 'q_lst =', len( q_lst ), q_lst ) NSTools.p( ' M -->', len( q_lst ), [q.mat_mul( inv.M ) for q in q_lst] ) NSTools.p( 'inv.Md_lst =', inv.Mtype, inv.Md_lst, ', rank =', inv.get_rank() ) return s_lst, q_lst @staticmethod def seek_bases( inv, d_lst, r_lst, eq = False, num = -1, b_lst = [], bas_lst = [] ): ''' Look for root bases in a given set of roots whose Dynkin type is the same as a given root bases. This method is used by get_cls(). Parameters ---------- inv : DPLattice We use inv.Md_lst and inv.M for when creating a new DPLattice object. d_lst : list<Div> We use the intersection matrix associated to d_lst. r_lst : list<Div> A list of roots in which to look for root bases. eq : boolean If True, then the returned bases are pairwise non-equivalent. By default False, in which case only bases that differ by a permutation of elements are considered equivalent. num : int If num>0, then the method will terminate if the number of bases found is equal to num. If num==-1, then the method continues until all possible bases have been reached. b_lst : list<Div> Used for recursive calling this method and represents (a subset of) a candidate root bases. bas_lst : list<DPLattice> Used for recursive calling this method and is the list of DPLattice objects that is returned by this method. Returns ------- list<DPLattice> A list of DPLattice objects "bas" such that bas.type is equal to the Dynkin type of d_lst and bas.Mtype==inv.Mtype and bas.M==inv.M. If eq==True, then the lattice objects are pairwise non-equivalent. If num>0, then the method terminates if the number of bases that are found is equal to num. ''' # check whether the constructed basis defines a new DPLattice object if len( b_lst ) == len( d_lst ): # check if a permutation of b_lst occurred if not eq: for bas in bas_lst: if set( bas.d_lst ) == set( b_lst ): return bas_lst # create a new lattice object bas = DPLattice( b_lst, inv.Md_lst, inv.M ) # check whether there is an equivalent object in bas_lst if eq and bas in bas_lst: return bas_lst # return bas_lst appended with the new DPLattice object return bas_lst + [bas] else: # construct list with intersection numbers s = d_lst[ len( b_lst ) ] m_lst = [ d * s for d in d_lst[:len( b_lst )] ] # go through all possible roots to build up a basis like d_lst for r in r_lst: # check intersection number properties if [b * r for b in b_lst] == m_lst: # recursive call bas_lst = DPLattice.seek_bases( inv, d_lst, r_lst, eq, num, b_lst + [r], bas_lst ) # break out of loop if num bases are found if num > 0 and len( bas_lst ) == num: break return bas_lst @staticmethod def import_cls( cls_lst, inv ): ''' This method is used by get_cls(). Parameters ---------- cls_lst : list<DPLattice> A list of DPLattice objects of rank "inv.get_rank()-1". These lattices correspond to Neron-Severi lattices of weak Del Pezzo surfaces. inv : DPLattice A DPLattice object representing an involution. We expect inv.Md_lst to be set. Returns ------- list<DPLattice> A list of compatible DPLattice objects in cls_lst that are converted so as to have the same rank and involution matrix as inv.get_rank() and inv.M, respectively. The returned list always contains inv itself. ''' out_lst = [] for cls in cls_lst: # convert divisors to new rank Md_lst = [ Div.new( str( d ), inv.get_rank() ) for d in cls.Md_lst ] d_lst = [ Div.new( str( d ), inv.get_rank() ) for d in cls.d_lst ] # import if the involution is compatible if set( Md_lst ) == set( inv.Md_lst ): NSTools.p( 'importing: ', ( inv.get_rank(), cls.get_marked_Mtype(), cls.get_real_type() ), Md_lst, '==', inv.Md_lst ) out = DPLattice( d_lst, inv.Md_lst, inv.M ) out.set_attributes() out_lst += [ out ] # always ensure that at least inv object is contained if out_lst == []: return [inv] # we expect that inv is contained in the out_lst # for correctness of the get_cls() algorithm. assert inv in out_lst return out_lst @staticmethod def get_cls( rank = 9 ): ''' Parameters ---------- rank : int An integer in [1,...,9]. Returns ------- list<DPLattice> A list of DPLattice objects corresponding to Neron-Severi lattices of weak Del Pezzo surfaces of degree (10-rank). The list contains exactly one representative for each equivalence class. All the Div objects referenced in the DPLattice objects of the output have the default intersection matrix: diagonal matrix with diagonal: (1,-1,...,-1). If rank<3 then the empty list is returned. ''' if rank < 3: return [] # check cache key = 'get_cls_' + str( rank ) if key in NSTools.get_tool_dct(): return NSTools.get_tool_dct()[key] NSTools.p( 'rank =', rank ) # collect all lattices with either d_lst==[] of Md_lst==[] bas_lst = DPLattice.get_bas_lst( rank ) inv_lst = DPLattice.get_inv_lst( rank ) # we loop through all involutions NSTools.p( 'start looping through inv_lst: ', len( inv_lst ), [inv.get_marked_Mtype() for inv in inv_lst] ) dpl_lst = [] for inv in inv_lst: NSTools.p( 'looping through inv_lst:', ( rank, inv.get_marked_Mtype(), inv.Md_lst ) ) # recover the known classification if inv.Mtype == 'A0': NSTools.p( 'Since Mtype equals A0 we recover the classification from bas_lst.' ) dpl_lst += [bas for bas in bas_lst] continue # partition the roots into two sets s_lst, q_lst = DPLattice.get_part_roots( inv ) # import classification for rank-1 bas1_lst = DPLattice.import_cls( DPLattice.get_cls( rank - 1 ), inv ) NSTools.p( 'looping through inv_lst continued after recursive call:', ( rank, inv.get_marked_Mtype(), inv.Md_lst ) ) # correct partition of roots (bas1_lst always contains inv) if len( bas1_lst ) > 1: e = Div.new( 'e' + str( rank - 1 ), inv.get_rank() ) s_lst = [ s for s in s_lst if s * e != 0 ] q_lst = [ q for q in q_lst if q * e != 0 ] NSTools.p( 'bas1_lst =', len( bas1_lst ), [( bas1.Mtype, bas1.type ) for bas1 in bas1_lst] ) NSTools.p( 's_lst =', len( s_lst ), s_lst ) NSTools.p( 'q_lst =', len( q_lst ), q_lst ) # collect all possible root bases in s_lst and q_lst bas2_lst = [] bas3_lst = [] visited_type_lst = [] eta = ETA( len( bas_lst ), 1 ) for bas in bas_lst: # display progress info eta.update( 'get_cls seeking bases in s_lst and q_lst: ', ( rank, inv.get_marked_Mtype(), bas.get_real_type() ) ) # each type in bas_lst is treated only once if bas.type in visited_type_lst: continue visited_type_lst += [bas.type] # collect bases of type bas.type in s_lst if DPLattice.get_num_types( inv, bas, bas_lst ) != 0: bas2_lst += DPLattice.seek_bases( inv, bas.d_lst, s_lst ) # collect bases of type bas.type in q_lst if 2 * len( bas.d_lst ) > rank - 1: continue # the rank of a root subsystem is bounded by rank-1 tmp_lst = DPLattice.seek_bases( inv, bas.d_lst, q_lst ) for tmp in tmp_lst: tmp.d_lst += [d.mat_mul( inv.M ) for d in tmp.d_lst ] if is_root_basis( tmp.d_lst ): # the roots and their involutions might have intersection product 1 tmp.d_lst.sort() bas3_lst += [tmp] # debug info NSTools.p( 'Setting Dynkin types of', len( bas2_lst + bas3_lst ), 'items...please wait...' ) eta = ETA( len( bas2_lst + bas3_lst ), len( bas2_lst + bas3_lst ) / 10 ) for bas in bas2_lst + bas3_lst: bas.type = get_dynkin_type( bas.d_lst ) bas.Mtype = get_dynkin_type( bas.Md_lst ) eta.update( bas.get_rank(), bas.get_marked_Mtype(), bas.type ) bas1_lst.sort() bas2_lst.sort() bas3_lst.sort() t_lst1 = [bas.type for bas in bas1_lst] t_lst2 = [bas.type for bas in bas2_lst] t_lst3 = [bas.type for bas in bas3_lst] lst1 = sorted( list( set( [( t, t_lst1.count( t ) ) for t in t_lst1] ) ) ) lst2 = sorted( list( set( [( t, t_lst2.count( t ) ) for t in t_lst2] ) ) ) lst3 = sorted( list( set( [( t, t_lst3.count( t ) ) for t in t_lst3] ) ) ) NSTools.p( 'inv =', inv.get_marked_Mtype(), ', rank =', rank ) NSTools.p( 'bas1_lst =', len( bas1_lst ), lst1 ) NSTools.p( 'bas2_lst =', len( bas2_lst ), lst2 ) NSTools.p( 'bas3_lst =', len( bas3_lst ), lst3 ) # construct a list of combinations of DPLattice objects in bas1_lst bas2_lst and comb_lst = [] total = len( bas1_lst ) * len( bas2_lst ) * len( bas3_lst ) step = total / 10 if total > 10 else total eta = ETA( total, step ) for bas1 in bas1_lst: for bas2 in bas2_lst: for bas3 in bas3_lst: eta.update( 'last loop in get_cls: ( bas1.type, bas2.type, bas3.type )=', ( bas1.type, bas2.type, bas3.type ) ) d_lst = bas1.d_lst + bas2.d_lst + bas3.d_lst # notice that d_lst can be equal to [] if len( d_lst ) > rank - 1: continue # the rank of a root subsystem is bounded by rank-1 if is_root_basis( d_lst ): dpl = DPLattice( d_lst, inv.Md_lst, inv.M ) if dpl not in dpl_lst: dpl.set_attributes() dpl_lst += [dpl] NSTools.p( '\t appended: ', ( rank, dpl.get_marked_Mtype(), dpl.get_real_type() ), ', ( bas1.type, bas2.type, bas3.type ) =', ( bas1.type, bas2.type, bas3.type ) ) # store in cache # dpl_lst.sort() NSTools.get_tool_dct()[key] = dpl_lst NSTools.save_tool_dct() return dpl_lst # overloading of "==" # returns True if isomorphic as Neron-Severi lattices def __eq__( self, other ): # compared with None? if type( self ) != type( other ): return False # cardinality of classes agree? if len( self.d_lst ) != len( other.d_lst ): return False self.set_attributes( 0 ) other.set_attributes( 0 ) if len( self.m1_lst ) != len( other.m1_lst ): return False self.set_attributes( 1 ) other.set_attributes( 1 ) if len( self.fam_lst ) != len( other.fam_lst ): return False self.set_attributes( 2 ) other.set_attributes( 2 ) if len( self.real_d_lst ) != len( other.real_d_lst ): return False self.set_attributes( 3 ) other.set_attributes( 3 ) if len( self.real_m1_lst ) != len( other.real_m1_lst ): return False self.set_attributes( 4 ) other.set_attributes( 4 ) if len( self.real_fam_lst ) != len( other.real_fam_lst ): return False self.set_attributes( 5 ) other.set_attributes( 5 ) if len( self.or_lst ) != len( other.or_lst ): return False self.set_attributes( 6 ) other.set_attributes( 6 ) if len( self.sr_lst ) != len( other.sr_lst ): return False # Dynkin type effective (-2)-classes agree? self.set_attributes( 7 ) other.set_attributes( 7 ) if self.type != other.type: return False # Mtype may differ for equivalent DPLattice objects # check Cremona invariant self.set_attributes( 9 ) other.set_attributes( 9 ) if not self.G.is_isomorphic( other.G, edge_labels = True ): return False return True # operator overloading for != def __ne__( self, other ): return not self.__eq__( other ) # operator overloading for < # Used for sorting lists of DPLattice objects: # <http://stackoverflow.com/questions/1227121/compare-object-instances-for-equality-by-their-attributes-in-python> def __lt__( self, other ): if self.get_rank() != other.get_rank(): return self.get_rank() < other.get_rank() if len( self.Md_lst ) != len( other.Md_lst ): return len( self.Md_lst ) < len( other.Md_lst ) self.set_attributes( 8 ) other.set_attributes( 8 ) if self.Mtype != other.Mtype: return self.Mtype < other.Mtype if self.get_marked_Mtype() != other.get_marked_Mtype(): return self.get_marked_Mtype() < other.get_marked_Mtype() if len( self.d_lst ) != len( other.d_lst ): return len( self.d_lst ) < len( other.d_lst ) if self.type != other.type: return self.type < other.type # more real lines implies smaller self.type! if len( self.real_m1_lst ) != len( other.real_m1_lst ): return len( self.real_m1_lst ) > len( other.real_m1_lst ) if len( self.m1_lst ) != len( other.m1_lst ): return len( self.m1_lst ) > len( other.m1_lst ) if len( self.real_fam_lst ) != len( other.real_fam_lst ): return len( self.real_fam_lst ) > len( other.real_fam_lst ) if len( self.fam_lst ) != len( other.fam_lst ): return len( self.fam_lst ) > len( other.fam_lst ) # overloading of "str()": human readable string representation of object def __str__( self ): self.set_attributes() s = '\n' s += 50 * '=' + '\n' s += 'Degree = ' + str( self.get_degree() ) + '\n' s += 'Rank = ' + str( self.get_rank() ) + '\n' s += 'Intersection = ' + str( list( self.m1_lst[0].int_mat ) ) + '\n' s += 'Real structure = ' + str( self.get_marked_Mtype() ) + '\n' s += 'Singularities = ' + str( self.type ) + '\n' s += 'Cardinalities = ' + '(' + str( len( self.or_lst ) ) + ', ' + str( len( self.sr_lst ) ) + ')\n' arrow = ' ---> ' s += 'Real involution:\n' b_lst = [Div( row ) for row in sage_identity_matrix( sage_ZZ, self.get_rank() ).rows() ] for b in b_lst: s += '\t' + str( b ) + arrow + str( b.mat_mul( self.M ) ) + '\n' s += 'Indecomposable (-2)-classes:\n' for d in self.d_lst: s += '\t' + str( d ) + arrow + str( d.mat_mul( self.M ) ) + '\n' s += '\t#real = ' + str( len( self.real_d_lst ) ) + '\n' s += 'Indecomposable (-1)-classes:\n' for m1 in self.m1_lst: s += '\t' + str( m1 ) + arrow + str( m1.mat_mul( self.M ) ) + '\n' s += '\t#real = ' + str( len( self.real_m1_lst ) ) + '\n' s += 'Classes of conical families:\n' for fam in self.fam_lst: s += '\t' + str( fam ) + arrow + str( fam.mat_mul( self.M ) ) + '\n' s += '\t#real = ' + str( len( self.real_fam_lst ) ) + '\n' s += 50 * '=' + '\n' return s
47,232
34.674471
195
py
ns_lattice
ns_lattice-master/ns_lattice/src/ns_lattice/div_in_lattice.py
''' Use of this source code is governed by a MIT-style license that can be found in the LICENSE file. Created on Aug 11, 2016 @author: Niels Lubbes Algorithm for computing elements in a unimodular lattice. We use this algorithm in the context of Neron-Severi lattices of weak del Pezzo surfaces. See Arxiv: "Computing curves on real rational surfaces" ''' import time from ns_lattice.sage_interface import sage_Combinations from ns_lattice.sage_interface import sage_Compositions from ns_lattice.sage_interface import sage_Partitions from ns_lattice.sage_interface import sage_ZZ from ns_lattice.sage_interface import sage_Permutations from ns_lattice.class_ns_tools import NSTools from ns_lattice.class_div import Div def get_divs( d, dc, cc, perm = False ): ''' Computes divisors in unimodular lattice with prescribed intersection product. Parameters ---------- d : Div object d0*e0 + d1*e1 +...+ dr*er such that * product signature equals (1,d.rank()-1) * d0>0 * d1,...,dr<=0 dc : int A positive integer. cc : int Self intersection. perm : boolean If True, then generators are permuted. Returns ------- list<Div> Returns a sorted list of "Div" objects * c = c0*e0 + c1*e1 +...+ cr*er such that * d.rank() == r+1 * dc == d*c (signature = (1,rank-1)) * cc == c*c (signature = (1,rank-1)) and each Div object satisfies exactly one of the following conditions: * c == ei - ej for 0>i>j>=r, * c == ei for i>0, or * c0 > 0, c1,...,cr <= 0 If "perm" is False, then then only one representative for each c is returned up to permutation of ei for i>0. For example, e0-e1-e2 and e0-e1-e3 are considered equivalent, and only e0-e1-e2 is returned, since e0-e1-e2>e0-e1-e3 (see "get_div_set()" for the ordering). In particular, c1 >= c2 >= ... >= cr. Note ---- If d=[3]+8*[-1], (dc,cc)==(0,-2) and perm=False then the Div classes are '12', '1123', '212' and '308'. See "Div.get_label()" for the notation. These classes correspond to the (-2)-classes in the Neron-Severi lattice associated to a weak del Pezzo surface. If perm==False then only one representative for each q is returned up to permutation of ei for i>0. For example, e0-e1-e2 and e0-e1-e3 are considered equivalent, and only e0-e1-e2 is returned, since e0-e1-e2>e0-e1-e3 (see "Div.__lt__()" for the ordering). ''' # check if input was already computed # key = 'get_divs_' + str( ( d, dc, cc, perm ) ) if key in NSTools.get_tool_dct(): return NSTools.get_tool_dct()[key] # construct div set # NSTools.p( 'Constructing div set classes for ', ( d, dc, cc, perm ) ) out_lst = [] # compute classes of the form ei or ei-ej for i,j>0 # if ( dc, cc ) == ( 1, -1 ) or ( dc, cc ) == ( 0, -2 ): m2_lst = [] # list of divisors of the form ei-ej for i,j>0 m1_lst = [] # list of divisors of the form ei for i>0 if perm: # Example: # >>> list(Combinations( [1,2,3,4], 2 )) # [[1, 2], [1, 3], [1, 4], [2, 3], [2, 4], [3, 4]] # Notice that r=d.rank()-1 if c = c0*e0 + c1*e1 +...+ cr*er. # for comb in sage_Combinations( range( 1, d.rank() ), 2 ): m2_lst += [ Div.new( str( comb[0] ) + str( comb[1] ), d.rank() ) ] m1_lst += [Div.new( 'e' + str( i ), d.rank() ) for i in range( 1, d.rank() )] else: # up to permutation of the generators # we may assume that i==1 and j==2. # m2_lst += [ Div.new( '12', d.rank() ) ] m1_lst += [ Div.new( 'e1', d.rank() ) ] # add the classes that satisfy return # specification to the output list # for c in m1_lst + m2_lst: if ( dc, cc ) == ( d * c, c * c ): out_lst += [c] # # Note: cc = c0^2 - c1^2 -...- cr^2 # c0 = 0 cur_eq_diff = -1 while True: c0 = c0 + 1 dc_tail = d[0] * c0 - dc # = d1*c1 +...+ dr*cr dd_tail = d[0] ** 2 - d * d # = d1^2 +...+ dr^2 cc_tail = c0 ** 2 - cc # = c1^2 +...+ cr^2 # not possible according to io-specs. # if dc_tail < 0 or dd_tail < 0 or cc_tail < 0: NSTools.p( 'continue... (c0, dc_tail, dd_tail, cc_tail) =', ( c0, dc_tail, dd_tail, cc_tail ) ) if dd_tail < 0: raise Exception( 'dd_tail =', dd_tail ) continue # Cauchy-Schwarz inequality <x,y>^2 <= <x,x>*<y,y> holds? # prv_eq_diff = cur_eq_diff cur_eq_diff = abs( dc_tail * dc_tail - dd_tail * cc_tail ) if prv_eq_diff == -1: prv_eq_diff = cur_eq_diff NSTools.p( 'prv_eq_diff =', prv_eq_diff, ', cur_eq_diff =', cur_eq_diff, ', dc_tail^2 =', dc_tail * dc_tail, ', dd_tail*cc_tail =', dd_tail * cc_tail, ', (c0, dc_tail, dd_tail, cc_tail) =', ( c0, dc_tail, dd_tail, cc_tail ) ) if prv_eq_diff < cur_eq_diff and dc_tail * dc_tail > dd_tail * cc_tail: NSTools.p( 'stop by Cauchy-Schwarz inequality...' ) break # out of while loop # obtain all possible [d1*c1+1,...,dr*cr+1] # r = d.rank() - 1 if perm and len( set( d[1:] ) ) != 1: p_lst_lst = sage_Compositions( dc_tail + r, length = r ) else: p_lst_lst = sage_Partitions( dc_tail + r, length = r ) # data for ETA computation total = len( p_lst_lst ) counter = 0 ival = 5000 # obtain [c1,...,cr] from [d1*c1+1,...,dr*cr+1] # for p_lst in p_lst_lst: # ETA if counter % ival == 0: start = time.time() counter += 1 if counter % ival == 0: passed_time = time.time() - start NSTools.p( 'ETA in minutes =', passed_time * ( total - counter ) / ( ival * 60 ), ' (', counter, '/', total, '), c0 =', c0, ', prv_eq_diff =', prv_eq_diff, ', cur_eq_diff =', cur_eq_diff ) # dc_tail=d1*c1 +...+ dr*cr = p1 +...+ pr with pi>=0 p_lst = [ p - 1 for p in p_lst] # obtain c_tail=[c1,...,cr] from [p1,...,pr] valid_part = True c_tail = [] # =[c1,...,cr] for i in range( 0, len( p_lst ) ): if p_lst[i] == 0 or d[i + 1] == 0: c_tail += [p_lst[i]] else: quo, rem = sage_ZZ( p_lst[i] ).quo_rem( d[i + 1] ) if rem != 0: valid_part = False break # out of i-for-loop else: c_tail += [ quo ] if not valid_part: continue # add to out list if valid # c = Div( [c0] + c_tail ) if c.rank() == d.rank() and ( dc, cc ) == ( d * c, c * c ): if perm and len( set( d[1:] ) ) == 1: # since d1==...==dr we do not have to # check each permutation. for pc_tail in sage_Permutations( c_tail ): out_lst += [Div( [c0] + list( pc_tail ) )] else: out_lst += [c] # sort list of "Div" objects out_lst.sort() # cache output NSTools.get_tool_dct()[key] = out_lst NSTools.save_tool_dct() return out_lst def get_indecomp_divs( c_lst, d_lst ): ''' Parameters ---------- c_lst : list<Div> Typically output of "get_divs(...)" d_lst : list<Div> Typically a list of (-2)-classes. Returns ------- list<Div> Returns a list of "Div" objects c in c_lst, so that c*d >= 0 for all d in "d_lst". Note ---- If the Div object represent effective divisor classes in a the Neron-Severi lattice of a weak del Pezzo surface and if d_lst are the classes of singularities, then the output correspond to "indecomposable" classes. Such classes cannot be written as the sum of effective divisors. ''' # check positivity against "d_lst" out_lst = [] for c in c_lst: indecomp = True for d in d_lst: if d * c < 0: indecomp = False break # out of for loop if indecomp: out_lst += [c] return out_lst def get_ak( rank ): ''' Parameters ---------- rank : int Returns ------- Div A Div object of given rank of the form 3e0 - e1 - ... - er Mathematically this is the anticanonical class of the blowup of the projective plane. ''' return Div( [3] + ( rank - 1 ) * [-1] )
9,393
31.061433
233
py
ns_lattice
ns_lattice-master/ns_lattice/src/ns_lattice/ns_basis.py
''' Use of this source code is governed by a MIT-style license that can be found in the LICENSE file. Created on Feb 9, 2017 @author: Niels Lubbes ''' from ns_lattice.sage_interface import sage_identity_matrix from ns_lattice.sage_interface import sage_matrix from ns_lattice.sage_interface import sage_ZZ from ns_lattice.sage_interface import sage_Permutations from ns_lattice.sage_interface import sage_Subsets from ns_lattice.class_div import Div from ns_lattice.div_in_lattice import get_indecomp_divs from ns_lattice.div_in_lattice import get_ak from ns_lattice.div_in_lattice import get_divs from ns_lattice.class_dp_lattice import DPLattice from ns_lattice.class_eta import ETA from ns_lattice.class_ns_tools import NSTools def get_bases_lst( a_lst, M, d_lst, m1_lst, perm = False ): ''' Returns a list of basis with specified generators. Parameters ---------- a_lst : list<Div> A list of linear independent Div objects of the same rank with 3<=rank<=9. It is required that "set(a_lst)==set([ a.mat_mul(M) for a in a_lst ])". M : sage_matrix<sage_ZZ> A unimodular matrix representing an involution. d_lst : list<Div> A list of Div objects d of the same rank as any element in "a_lst", so that "d*k==0" and "d*d==-2". These represent a root basis for the indecomposable (-2)-classes in the Neron-Severi lattice of a weak del Pezzo surface. m1_lst : list<Div> A list of Div objects d of the same rank as any element in "a_lst", so that "d*k==d*d==-1". These represent (-1)-classes in the Neron-Severi lattice of a weak del Pezzo surface. perm : bool If False, then we consider two bases the same if the generators of the first basis can be obtained from the second basis via a permutation matrix. Returns ------- list<tuple<Div>> A list of tuples of Div objects. Each tuple of Div objects represents a basis for the Neron-Severi lattice determined by d_lst and m1_lst. The bases are of the form < a1,...,as, b1,...,bt > with the following property * a1,...,as are defined by the input "a_lst" * bi is an element in m1_lst such that bi*bj=am*bi=0 for all 1<=i<j<=t and 1<=m<=s If "a_lst==[]" then "[[]]" is returned. ''' key = 'get_bases_lst__' + str( ( a_lst, M, d_lst, m1_lst, perm ) ) + '__' + str( M.rank() ) if key in NSTools.get_tool_dct(): return NSTools.get_tool_dct()[key] if a_lst == []: return [[]] if len( a_lst ) == a_lst[0].rank(): return [tuple( a_lst )] e_lst = [] for m1 in get_indecomp_divs( m1_lst, d_lst ): if set( [ m1 * a for a in a_lst ] ) != {0}: continue if m1 * m1.mat_mul( M ) > 0: continue e_lst += [m1] bas_lst = [] for e in e_lst: Me = e.mat_mul( M ) new_d_lst = [ d for d in d_lst if d * e == d * Me == 0 ] new_m1_lst = [ m1 for m1 in m1_lst if m1 * e == m1 * Me == 0 ] add_lst = [e] if e != Me: add_lst += [Me] bas2_lst = get_bases_lst( a_lst + add_lst, M, new_d_lst, new_m1_lst, perm ) if perm: bas_lst += bas2_lst else: for bas2 in bas2_lst: found = False for bas in bas_lst: # check whether the two bases are the same up to # permutation of generators if set( bas ) == set( bas2 ): found = True break # break out of nearest for loop if not found: NSTools.p( 'found new basis: ', bas2, ', bas2_lst =', bas2_lst ) bas_lst += [bas2] # cache output NSTools.get_tool_dct()[key] = bas_lst NSTools.save_tool_dct() return bas_lst def get_webs( dpl ): ''' Returns lists of families of conics for each possible complex basis change. The n-th family in each list correspond to a fixed family wrt. different bases for each n. Parameters ---------- dpl : DPLattice Represents the Neron-Severi lattice of a weak del Pezzo surface. Returns ------- list<list<Div>> A list of lists of Div objects. Each Div object f has the property that f*(3e0-e1-...-er)=2, f*f==0 and f*d>=0 for all d in dpl.d_lst. Such a Div object corresponds geometrically to a family of conics. For each index i, the i-th entry of each list of Div object corresponds to the same family of conics. ''' key = 'get_webs__' + str( dpl ).replace( '\n', '---' ) if key in NSTools.get_tool_dct(): return NSTools.get_tool_dct()[key] ak = get_ak( dpl.get_rank() ) all_m1_lst = get_divs( ak, 1, -1, True ) akc, cc = ( 3, 1 ) M = sage_identity_matrix( dpl.get_rank() ) fam_lst_lst = [] for e0 in get_divs( ak, akc, cc, True ): NSTools.p( 'e0 =', e0 ) for B_lst in get_bases_lst( [e0], M, dpl.d_lst, all_m1_lst, True ): B = sage_matrix( sage_ZZ, [ d.e_lst for d in B_lst ] ) dplB = dpl.get_basis_change( B ) fam_lst_lst += [ dplB.real_fam_lst ] # reduce fam_lst pat_lst_lst = [] rfam_lst_lst = [] for fam_lst in fam_lst_lst: pat_lst = [ 0 if fam[0] != 1 else 1 for fam in fam_lst ] if pat_lst not in pat_lst_lst: pat_lst_lst += [ pat_lst ] rfam_lst_lst += [ fam_lst ] # cache output NSTools.get_tool_dct()[key] = rfam_lst_lst NSTools.save_tool_dct() return rfam_lst_lst def contains_perm( f_lst_lst, c_lst ): ''' Parameters ---------- f_lst_lst : list<list<Div>> A list of lists containing Div objects. c_lst : list<Div> A list of Div objects Returns: -------- bool Returns True if after a permutation of the generators (e1,...,er) the list c_lst is contained in f_lst_lst. For example if c_lst equals [ e0-e1, 2e0-e2-e3-e4-e5 ] then is contained in [ ..., [e0-e2, 2e0-e1-e3-e4-e5], ... ]. ''' if c_lst == []: return [] in f_lst_lst for perm in sage_Permutations( range( c_lst[0].rank() - 1 ) ): pc_lst = [ Div( [c[0]] + [ c[i + 1] for i in perm ], c.rank() ) for c in c_lst ] for f_lst in f_lst_lst: if set( f_lst ) == set( pc_lst ): return True return False def triples( dpl, mval ): ''' Parameters ---------- dpl : DPLattice mval : integer Returns ------- list<(Div,Div,Div)> List of triples in "dpl.fam_lst": [ (a,b,c),... ] so that (1) There does not exists e in "dpl.m1_lst" with the property that a*e==b*e==c*e==0. (2) 1 <= max( a*b, a*c, b*c ) <= mval. ''' key = 'triples__' + str( dpl ).replace( '\n', '---' ) + '---' + str( mval ) if key in NSTools.get_tool_dct(): return NSTools.get_tool_dct()[key] f_lst = dpl.fam_lst e_lst = dpl.m1_lst # obtain list of triples (a,b,c) in f_lst # that are not orthogonal to any element in e_lst t_lst = [] idx_lst_lst = sage_Subsets( range( len( f_lst ) ), 3 ) eta = ETA( len( idx_lst_lst ), 500000 ) for idx_lst in idx_lst_lst: eta.update( 't_lst' ) t = [ f_lst[idx] for idx in idx_lst ] if t[0] * t[1] > mval: continue if t[0] * t[2] > mval: continue if t[1] * t[2] > mval: continue # elements in f_lst correspond to divisor classes of curves on a # surface and thus t[i]*t[j]>=1 for all i,j \in {0,1,2} so that i!=j. cont = False for e in e_lst: if [f * e for f in t] == [0, 0, 0]: cont = True break if cont: continue if not contains_perm( t_lst, t ): t_lst += [t] NSTools.p( 't_lst =', t_lst ) # cache output NSTools.get_tool_dct()[key] = t_lst NSTools.save_tool_dct() return t_lst
8,302
30.214286
97
py
ns_lattice
ns_lattice-master/ns_lattice/src/ns_lattice/class_ns_tools.py
''' Use of this source code is governed by a MIT-style license that can be found in the LICENSE file. Created on Feb 7, 2017 @author: Niels Lubbes ''' from ns_lattice.sage_interface import sage_load from ns_lattice.sage_interface import sage_save import inspect import time import sys import os class NSTools(): ''' For accessing static variables in python see for example: <http://stackoverflow.com/questions/68645/static-class-variables-in-python> ''' # Private dictionary object for caching result # used by ".get_tool_dct()" and ".save_tool_dct()". # If "enable_tool_dct" is false then caching in # disabled. This is useful for example in test # methods. However, it should be noted that it # could take a long time to compute the data. # __tool_dct = None __enable_tool_dct = True # private variable for timer # __start_time = None __end_time = None # private static variables used by ".p()" # If "__filter_fname_lst" equals [] then output is surpressed. # If "__filter_fname_lst" equals None the no output is surpressed # __filter_fname_lst = [] __prev_filter_fname_lst = None @staticmethod def filter( filter_fname_lst ): ''' It is adviced to access this method as statically as .filter(). See .p() for more details. Parameters ---------- filter_fname_lst : list<str> List of file names for Python modules. If None, then no output is surpressed by method ".p()". ''' NSTools.__filter_fname_lst = filter_fname_lst NSTools.__prev_filter_fname_lst = filter_fname_lst @staticmethod def filter_unset(): ''' Output via ".out" will not be surpressed. ''' NSTools.__filter_fname_lst = None @staticmethod def filter_reset(): ''' Resets filter state to before previous ".filter_unset()" call. ''' NSTools.__filter_fname_lst = NSTools.__prev_filter_fname_lst @staticmethod def p( *arg_lst ): ''' Parameters ---------- *arg_lst Variable length argument list. Returns ------- string If ".filter_on(<fname>)" has been called and the file name of the calling module does not coincide with <fname>, then the output is surpressed and "None" is returned. Otherwise, this method prints arguments to "sys.stdout" together with reflection info from "inspect.stack()". Additional returns the output string. Call ".filter_off()" to turn off filter, such that all output is send to "sys.stdout". ''' # collect relevant info from stack trace sk_lst_lst = inspect.stack() file_name = os.path.basename( str( sk_lst_lst[1][1] ) ) # exclude path from file name line = str( sk_lst_lst[1][2] ) method_name = str( sk_lst_lst[1][3] ) # only output when .p() is called from module whose # file name is in .__filter_fname_lst if NSTools.__filter_fname_lst != None: if not file_name in NSTools.__filter_fname_lst: return # construct output string s = method_name + '(' + line + ')' + ': ' for arg in arg_lst: s += str( arg ) + ' ' # print output print( s ) sys.stdout.flush() return s @staticmethod def set_enable_tool_dct( enable_tool_dct ): NSTools.filter_unset() NSTools.p( 'Caching enabled: ', enable_tool_dct ) NSTools.filter_reset() NSTools.__enable_tool_dct = enable_tool_dct @staticmethod def get_tool_dct( fname = 'ns_tools' ): ''' Parameters ---------- fname : str Name of file without extension. Returns ------- dct Sets static private variable "__tool_dct" in memory from file "<local path>/<fname>.sobj" if called for the first time. Returns ".__tool_dct" if ".__enable_tool_dct==True" and "{}" otherwise. ''' if not NSTools.__enable_tool_dct: return {} path = os.path.dirname( os.path.abspath( __file__ ) ) + '/' file_name = path + fname if NSTools.__tool_dct == None: try: NSTools.p( 'Loading from:', file_name ) NSTools.__tool_dct = sage_load( file_name ) except Exception as e: NSTools.filter_unset() NSTools.p( 'Cannot load ".__tool_dct": ', e ) NSTools.filter_reset() NSTools.__tool_dct = {} return NSTools.__tool_dct @staticmethod def save_tool_dct( fname = 'ns_tools' ): ''' Saves ".__tool_dct" to "fname" if ".enable_tool_dct==True" otherwise do nothing. Parameters ---------- fname : str Name of file without extension. ''' if not NSTools.__enable_tool_dct: return path = os.path.dirname( os.path.abspath( __file__ ) ) + '/' file_name = path + fname NSTools.p( 'Saving to:', file_name ) sage_save( NSTools.__tool_dct, file_name ) @staticmethod def start_timer(): ''' Prints the current wall clock time and starts timer. ''' # get time NSTools.__start_time = time.time() # set static variable. NSTools.filter_unset() NSTools.p( 'start time =', NSTools.__start_time ) NSTools.filter_reset() @staticmethod def end_timer(): ''' Prints wall clock time passed since last call of ".start_timer()". ''' NSTools.__end_time = time.time() passed_time = NSTools.__end_time - NSTools.__start_time NSTools.filter_unset() NSTools.p( 'time passed =', passed_time ) NSTools.filter_reset()
6,196
27.689815
97
py
ns_lattice
ns_lattice-master/ns_lattice/src/ns_lattice/__init__.py
1
0
0
py
ns_lattice
ns_lattice-master/ns_lattice/src/ns_lattice/sage_interface.py
''' Use of this source code is governed by a MIT-style license that can be found in the LICENSE file. Created on Jul 12, 2017 @author: Niels Lubbes Sage has a complicated import structure and it is not possible to simply import each need module. It seems that "from sage.all import *" is the only option. Therefore we introduce an interface to Sage so that in the code, it is clear, which libraries of Sage we use. Moreover, we specify below from which modules in the Sage library we import. We explain the naming scheme with the following two examples. The interface method for "PolynomialRing()" is called "sage_PolynomialRing()". However the interface method for "sage_eval()" is not called "sage_sage_eval()" but instead "sage__eval()". The variable "ZZ" is called "sage_ZZ". For the Parameters section in the documentation of types we will use the following abbrevations: sage_POLY: sage.rings.polynomial.multi_polynomial_element.MPolynomial_polydict The type of an element in sage_PolynomialRing sage_RING: sage.rings.* The type of a ring. For example sage_QQ or sage_ZZ or sage_NumberField. sage_GRAPH: sage.graphs.graph The type of a Graph. ''' from sage.all import * from sage.structure.sage_object import register_unpickle_override ################################################# # sage.structure # ################################################# # from sage.structure.proof.proof import proof sage_proof = proof # from sage.structure.sage_object import save def sage_save( *args, **kwargs ): return save( *args, **kwargs ) # from sage.structure.sage_object import load def sage_load( *args, **kwargs ): return load( *args, **kwargs ) # from sage.structure.sage_object import register_unpickle_override def sage_register_unpickle_override( *args, **kwargs ): register_unpickle_override( *args, **kwargs ) ################################################# # sage.misc # ################################################# # from sage.misc.sage_eval import sage_eval def sage__eval( *args, **kwargs ): return sage_eval( *args, **kwargs ) # from sage.misc.functional import n def sage_n( *args, **kwargs ): return n( *args, **kwargs ) ################################################# # sage.symbolic # ################################################# # from sage.symbolic.ring import SR sage_SR = SR # from sage.symbolic.relation import solve def sage_solve( *args, **kwargs ): return solve( *args, **kwargs ) ################################################# # sage.rings # ################################################# # from sage.rings.integer_ring import ZZ sage_ZZ = ZZ # from sage.rings.rational_field import QQ sage_QQ = QQ # import sage.rings.invariant_theory sage_invariant_theory = invariant_theory # from sage.rings.fraction_field import FractionField def sage_FractionField( *args, **kwargs ): return FractionField( *args, **kwargs ) # from sage.rings.polynomial.polynomial_ring_constructor import PolynomialRing # http://doc.sagemath.org/html/en/reference/polynomial_rings/sage/rings/polynomial/polynomial_ring_constructor.html def sage_PolynomialRing( *args, **kwargs ): return PolynomialRing( *args, **kwargs ) # from sage.rings.number_field.number_field import NumberField def sage_NumberField( *args, **kwargs ): return NumberField( *args, **kwargs ) ################################################# # sage.modules # ################################################# # from sage.modules import VectorSpace def sage_VectorSpace( *args, **kwargs ): return VectorSpace( *args, **kwargs ) ################################################# # sage.matrix # ################################################# # from sage.matrix.constructor import matrix def sage_matrix( *args, **kwargs ): return matrix( *args, **kwargs ) # from sage.matrix.constructor import identity_matrix def sage_identity_matrix( *args, **kwargs ): return identity_matrix( *args, **kwargs ) # from sage.matrix.constructor import diagonal_matrix def sage_diagonal_matrix( *args, **kwargs ): return diagonal_matrix( *args, **kwargs ) # from sage.matrix.constructor import vector def sage_vector( *args, **kwargs ): return vector( *args, **kwargs ) ################################################# # sage.arith # ################################################# # from sage.arith.misc import factor def sage_factor( *args, **kwargs ): return factor( *args, **kwargs ) # from sage.arith.misc import gcd def sage_gcd( *args, **kwargs ): return gcd( *args, **kwargs ) ################################################# # sage.calculus # ################################################# # from sage.calculus.functional import diff def sage_diff( *args, **kwargs ): return diff( *args, **kwargs ) # from sage.calculus.functional import expand def sage_expand( *args, **kwargs ): return expand( *args, **kwargs ) # from sage.calculus.var import var def sage_var( *args, **kwargs ): return var( *args, **kwargs ) ################################################# # sage.combinat # ################################################# # from sage.combinat.composition import Compositions def sage_Compositions( *args, **kwargs ): return Compositions( *args, **kwargs ) # from sage.combinat.combination import Combinations def sage_Combinations( *args, **kwargs ): return Combinations( *args, **kwargs ) # from sage.combinat.partitions import Partitions def sage_Partitions( *args, **kwargs ): return Partitions( *args, **kwargs ) # from sage.combinat.permutations import Partitions def sage_Permutations( *args, **kwargs ): return Permutations( *args, **kwargs ) # from sage.subset import Subsets def sage_Subsets( *args, **kwargs ): return Subsets( *args, **kwargs ) # from sage.combinat.root_system.root_system import RootSystem def sage_RootSystem( *args, **kwargs ): return RootSystem( *args, **kwargs ) ################################################# # sage.graphs # ################################################# # from sage.graphs.graph import Graph def sage_Graph( *args, **kwargs ): return Graph( *args, **kwargs )
6,555
28.936073
115
py
MORSE
MORSE-master/Generate_Params_Par.py
#! /usr/bin/env python # -*- coding: utf-8 -*- import sys import os import multiprocessing as mp import traceback import logging import gc import time from argparse import ArgumentParser import numpy from matplotlib import pyplot #from joblib import Parallel, delayed from scipy.signal import argrelextrema import multiprocessing import itertools from constants import c, G, Msun, rho_0, rho_ns, rho_1, rho_2, rho_3, P_0 def generate_params(n, low_gamma=0.5, high_gamma=6.5, Plow=[33.5, 34.5, 35.], Phigh=[34.8, 36., 37.]): """ Generates an array of combinations of P1, P2 and P3, making sure that P3 > P2 > P1. It also sets a lower and upper limit on the polytropic index. Args: n (int) : The number of points in logspace for each parameter. low_gamma (float): The lower limit on the polytropic index, default is 0.5. high_gamma (float): The upper limit on the polytropic index, default is 6.5. Plow (array) : Array of the three lower limits for P1, P2 and P3 in log10, default is [33.5, 34.5, 35.]. Phigh (array) : Array of the three upper limits for P1, P2 and P3 in log10, default is [34.8, 36., 37.]. Returns: out (ndarray) : Returns an array with all the possible combinations of P1, P2, P3. """ excluded = [] P_1 = numpy.logspace(Plow[0], Phigh[0], n) P_2 = numpy.logspace(Plow[1], Phigh[1], n) P_3 = numpy.logspace(Plow[2], Phigh[2], n) #P_1 = numpy.linspace(10**33.5, 10**34.8, n) #P_2 = numpy.linspace(10**34.5, 10**36., n) #P_3 = numpy.linspace(10**35., 10**37., n) iterables = [P_1, P_2, P_3] permutations = [] for t in itertools.product(*iterables): if t[0] < t[1] < t[2]: permutations.append(t) permutations = numpy.array(permutations) for i in range(len(permutations)): P_1 = permutations[i][0] P_2 = permutations[i][1] P_3 = permutations[i][2] gamma_1 = numpy.log10(P_1/P_0) / numpy.log10(rho_1/rho_0) gamma_2 = numpy.log10(P_2/P_1) / numpy.log10(rho_2/rho_1) gamma_3 = numpy.log10(P_3/P_2) / numpy.log10(rho_3/rho_2) if not low_gamma <= gamma_1 <= high_gamma or not low_gamma <= gamma_2 <= high_gamma or not low_gamma <= gamma_3 <= high_gamma: excluded.append(i) permutations = numpy.delete(permutations, excluded, axis=0) return permutations def calc_causal_limit(rho, P_1, P_2, P_3): """ Calculates if an EOS with parameters P1, P2 and P3 is causal at a given density. Args: rho (float): The density at which the function checks causality. P_1 (float): The first parameter of the EOS. P_2 (float): The second parameter of the EOS. P_3 (float): The third parameter of the EOS. Returns: rho (float): If causality is not violated, the density is returned. """ gamma_1 = numpy.log10(P_1/P_0) / numpy.log10(rho_1/rho_0) gamma_2 = numpy.log10(P_2/P_1) / numpy.log10(rho_2/rho_1) gamma_3 = numpy.log10(P_3/P_2) / numpy.log10(rho_3/rho_2) epsilon_0 = rho_0 + P_0/c**2.0 * 1./1.7 a_1 = epsilon_0/(rho_0) - 1. - P_1/((gamma_1 -1.)*rho_0*c**2.0) * (rho_0/rho_1)**gamma_1 a_2 = a_1 + P_1/((gamma_1 -1.)*rho_1*c**2.0) - P_1/((gamma_2 -1.)*rho_1*c**2.0) a_3 = a_2 + P_1/((gamma_2 -1.)*rho_2*c**2.0) * (rho_2/rho_1)**gamma_2 - P_2/((gamma_3 -1.) * rho_2*c**2.0) causality = 0 if rho_0 < rho <= rho_1: pres = P_1 * (rho/rho_1)**gamma_1 epsilon = (1. + a_1) * rho + P_1/((gamma_1 - 1.)*c**2.) *(rho/rho_1)**gamma_1 cs = gamma_1*pres/(epsilon*c**2. + pres) if gamma_1*pres/(epsilon*c**2. + pres) > 1.12: causality = 1 if rho_1 < rho <= rho_2: pres = P_1 * (rho/rho_1)**gamma_2 epsilon = (1. + a_2) * rho + P_1/((gamma_2 - 1.)*c**2.) *(rho/rho_1)**gamma_2 cs = gamma_2*pres/(epsilon*c**2. + pres) if gamma_2*pres/(epsilon*c**2. + pres) > 1.12: causality = 1 if rho > rho_2: pres = P_2 * (rho/rho_2)**gamma_3 epsilon = (1. + a_3) * rho + P_2/((gamma_3 - 1.)*c**2.) *(rho/rho_2)**gamma_3 cs = gamma_3*pres/(epsilon*c**2. + pres) if gamma_3*pres/(epsilon*c**2. + pres) > 1.12: causality = 1 if causality==0: return rho else: return 0.0 def calc_maxrho(parameters, n=1000, low_lim=14.31, up_lim=16.5): """ Generates an array of len(parameters) containing the maximum central density for each EOSs. Args: parameters (ndarray): An array of EOS parameters, dimensions (q,3) for q EOSs. n (int) : The number of points in logspace for each parameter. Default is 10^3. low_lim (float) : The log10 of the lower limit of central densities to test causality for, must be greater than 14.3, default is 14.31. up_lim (float) : The log10 of the lower limit of central densities to test causality for, default is 15.5. Returns: out (ndarray) : Returns an array with all maximum central densities, length (q). """ rhocents = numpy.logspace(low_lim, up_lim, n) max_rho = numpy.zeros(len(parameters)) for i,e in enumerate(parameters): causal_rho = numpy.zeros(len(rhocents)) P_1 = e[0] P_2 = e[1] P_3 = e[2] for j, k in enumerate(rhocents): causal_rho[j] = calc_causal_limit(k, P_1, P_2, P_3) if causal_rho[0]==0.0: max_rho[i] = 0.0 continue locmax = numpy.where(causal_rho==0.0)[0] if len(locmax) < 1: max_rho[i] = 10**(up_lim) else: max_rho[i] = causal_rho[locmax[0]-1] max_rho = numpy.array(max_rho) return max_rho info = mp.get_logger().info def main(n, low_lim, up_lim): logger = mp.log_to_stderr() logger.setLevel(logging.INFO) parameters = generate_params(n, low_lim, up_lim) print len(parameters) nproc = mp.cpu_count()# - 1 nproc = max(1, nproc) div_par = numpy.array_split(parameters, nproc) ntasks = nproc inputs = [[div_par[t], t] for t in xrange(ntasks)] input_q = mp.Queue() output_q = mp.Queue() procs = [ mp.Process(target=worker, args=(input_q,output_q)) for i in xrange(nproc)] for i in xrange(ntasks): input_q.put(inputs[i]) for i in xrange(nproc): input_q.put('STOP') for p in procs: p.start() result = [] while ntasks > 0: result.append(output_q.get()) ntasks -= 1 for p in procs: p.join() result = numpy.array(sorted(result, key=lambda x: x[1])) max_rho_array = numpy.concatenate(result[:,0]).ravel() numpy.save('max_rho', max_rho_array) numpy.save('input_parameters', parameters) def worker(input_q, output_q): start = time.clock() while True: try: tmp = input_q.get() if 'STOP' == tmp : break parameters, task = tmp max_rho = calc_maxrho(parameters) output_q.put([max_rho, task]) except Exception as exception: trace = str(traceback.format_exc()) info(trace) end = (time.clock() - start) info(end) return if __name__ == '__main__': parser = ArgumentParser() parser.add_argument("-n", dest="number", help="Number of values for each parameter", required=True) args = parser.parse_args() main(args.number, 0.5, 6.5)
8,000
28.307692
134
py
MORSE
MORSE-master/constants.py
### Constants used throughout my project code ### c = 3e10 G = 6.67428e-8 Msun = 1.989e33 rho_0 = 10**14.3 rho_ns = 2.7e14 rho_1 = 1.85 * rho_ns rho_2 = 2. * rho_1 #rho_2 = 1.8 * rho_1 rho_3 = 2. * rho_2 P_0 = 1.5281267425e+33 gamma_0 = 2.68019358431 dyncm2_to_MeVfm3 = 1./(1.6022e33) gcm3_to_MeVfm3 = 1./(1.7827e12) oneoverfm_MeV = 197.33
346
18.277778
49
py
MORSE
MORSE-master/PosteriorProbRho.py
#! /usr/bin/env python # -*- coding: utf-8 -*- import sys import os import multiprocessing as mp import traceback import logging import gc import numpy from matplotlib import pyplot from argparse import ArgumentParser from scipy.interpolate import UnivariateSpline, RegularGridInterpolator from scipy.integrate import simps, dblquad import time from constants import G, Msun, c def calc_determinant(JacPart): indices = numpy.where(numpy.invert(numpy.isnan(JacPart).any(axis=1)))[0] deter = numpy.zeros((len(JacPart[indices]), len(JacPart[indices]), len(JacPart[indices]))) derivs = numpy.zeros((6,6)) if indices.size==0: return numpy.nan, numpy.nan, numpy.nan else: for p, i in enumerate(indices): for r, j in enumerate(indices): for w, k in enumerate(indices): derivs[:,0][0:3] = JacPart[i][3:6] derivs[:,1][0:3] = JacPart[j][3:6] derivs[:,2][0:3] = JacPart[k][3:6] derivs[:,3][0:3] = JacPart[i][6::] derivs[:,4][0:3] = JacPart[j][6::] derivs[:,5][0:3] = JacPart[k][6::] derivs[3][[0,3]] = JacPart[i][1:3] derivs[4][[1,4]] = JacPart[j][1:3] derivs[5][[2,5]] = JacPart[k][1:3] deter[p, r, w] = abs(numpy.linalg.det(derivs)) return min(JacPart[indices][:,0]), max(JacPart[indices][:,0]), deter def calculate_norm(distribution): def Multivariate_notNorm(R, M, distribution): Mobs = distribution[0] Robs = distribution[1] sigmaM = distribution[2] sigmaR = distribution[3] rho = distribution[4] return numpy.exp(-1./(2.*(1.-rho**2.)) * ((R-Robs)**2. / sigmaR**2.0 + (M-Mobs)**2. / sigmaM**2.0 - \ 2.*rho*(R-Robs)*(M-Mobs)/(sigmaM*sigmaR))) norm = numpy.zeros(3) for i in range(3): norm[i] = dblquad(Multivariate_notNorm, 0.5, 3.3, lambda M: 2.94*G*M*Msun/(c**2. * 100000), lambda M: 14.3, args=([distribution[i]]))[0] return norm def Pobs(rho1, rho2, rho3, Jac_func, curveM, curveR, obs, norm): meanM = numpy.array([x[0] for x in obs]) meanR = numpy.array([x[1] for x in obs]) sigmaM = numpy.array([x[2] for x in obs]) sigmaR = numpy.array([x[3] for x in obs]) rho = numpy.array([x[4] for x in obs]) Rho = numpy.array([rho1, rho2, rho3]) obs = numpy.zeros(3, dtype=object) for i in range(3): obs[i] = 1./norm[i] * numpy.exp(-1./(2.*(1.-rho[i]**2.)) * ((((curveM(Rho[i])-meanM[i])**2.0)/sigmaM[i]**2.0) +\ (((curveR(Rho[i])-meanR[i])**2.0)/sigmaR[i]**2.0) - (2.*rho[i]*(curveM(Rho[i])-meanM[i])*\ (curveR(Rho[i])-meanR[i])/(sigmaM[i]*sigmaR[i])))) return obs[0] * obs[1] * obs[2] * abs(Jac_func((rho1, rho2, rho3))) def simps_integration(points, low_lim, up_lim, Jac_func, curveM, curveR, obs, norm): n = points rho1s = numpy.linspace(low_lim, up_lim, n) rho2s = numpy.linspace(low_lim, up_lim, n) rho3s = numpy.linspace(low_lim, up_lim, n) integral1 = numpy.zeros(len(rho1s)) integral2 = numpy.zeros(len(rho2s)) for i in range(len(rho3s)): for j in range(len(rho2s)): integral1[j] = simps(Pobs(rho1s, numpy.full(n, rho2s[j]), numpy.full(n, rho3s[i]), Jac_func, curveM, curveR, obs, norm), rho1s) integral2[i] = simps(integral1, rho2s) return simps(integral2, rho3s) info = mp.get_logger().info def main(MRIcurves, Jacobian, Observables, outputfile): logger = mp.log_to_stderr() logger.setLevel(logging.INFO) nproc = mp.cpu_count()# - 1 nproc = max(1, nproc) #norm = numpy.zeros(len(Observables), dtype=object) #for i, e in enumerate(Observables): norm = calculate_norm(Observables) div_MR = numpy.array_split(MRIcurves, nproc) div_jac = numpy.array_split(Jacobian, nproc) ntasks = nproc inputs = [[div_MR[t], div_jac[t], Observables, norm, t] for t in xrange(ntasks)] input_q = mp.Queue() output_q = mp.Queue() procs = [ mp.Process(target=worker, args=(input_q,output_q)) for i in xrange(nproc)] for i in xrange(ntasks): input_q.put(inputs[i]) for i in xrange(nproc): input_q.put('STOP') for p in procs: p.start() result = [] while ntasks > 0: result.append(output_q.get()) ntasks -= 1 result = numpy.array(sorted(result, key=lambda x: x[1])) result = numpy.delete(result, 1, axis=1) result = numpy.concatenate(result).ravel() result = numpy.concatenate(result).ravel() #Prob = numpy.zeros(len(Observables), dtype=object) #for z in range(len(Observables)): # Prob[z] = numpy.array([e[z] for e in result]) numpy.save(outputfile, result) for p in procs: p.join() def worker(input_q, output_q): start = time.clock() while True: try: tmp = input_q.get() if 'STOP' == tmp : break MRIcurves, Jacobian, Observables, norm, task = tmp info(len(Jacobian)) Prob = numpy.zeros(len(Jacobian)) for h, e in enumerate(Jacobian): Det = calc_determinant(e) M, R, I, rhoc = MRIcurves[h] x = numpy.log10(rhoc) curveM = UnivariateSpline(x, M, k=3, s=0) curveR = UnivariateSpline(x, R, k=3, s=0) if numpy.isnan(Det[0]) or len(Det[2])<3: continue rhos = numpy.linspace(Det[0], Det[1], len(Det[2])) Jac_func = RegularGridInterpolator((rhos, rhos, rhos), Det[2]) Prob[h] = simps_integration(25, min(rhos), max(rhos), Jac_func, curveM, curveR, Observables, norm) # Hack to avoid memory leak. Explicitly delete the instance of Jac_func and collect garbage. del Jac_func, curveM, curveR gc.collect() output_q.put([Prob, task]) except Exception as exception: trace = str(traceback.format_exc()) info(trace) end = (time.clock() - start) info(end) return if __name__ == '__main__': Observables = numpy.array([[1.5, 10.7, 0.05*1.5, 0.05*10.7, 0.0], [1.61, 10.5, 0.05*1.61, 0.05*10.5, 0.0], [1.7, 10.2, 0.05*1.7, 0.05*10.2, 0.0]]) parser = ArgumentParser() parser.add_argument("-f", dest="outputFile", help="write probability to FILE", metavar="FILE", required=True) parser.add_argument("-i1", dest="inputMRIcurves", help="use as input MRIcurves", required=True) parser.add_argument("-i2", dest="inputJacobian", help="use as input Jacobian", required=True) parser.add_argument("-i3", dest="inputObservables", help="use as input observables", required=True) args = parser.parse_args() MRIcurves = numpy.load(args.inputMRIcurves) Jacobian = numpy.load(args.inputJacobian) Observables = numpy.load(args.inputObservables) main(MRIcurves, Jacobian, Observables, args.outputFile)
7,431
31.596491
150
py
MORSE
MORSE-master/Plotting.py
import numpy import pandas from matplotlib import pyplot from scipy.interpolate import griddata from InputPosteriors import Pobs, find_CI_level pyplot.rcParams['xtick.direction'] = 'in' pyplot.rcParams['xtick.minor.visible'] = True pyplot.rcParams['ytick.direction'] = 'in' pyplot.rcParams['ytick.minor.visible'] = True pyplot.rcParams['xtick.major.size'] = 5 pyplot.rcParams['ytick.major.size'] = 5 pyplot.rcParams['ytick.right'] = True pyplot.rcParams['xtick.top'] = True pyplot.rcParams['axes.titlesize'] = 15 pyplot.rcParams['axes.labelsize'] = 24 pyplot.rcParams['xtick.labelsize'] = 20 pyplot.rcParams['ytick.labelsize'] = 20 pyplot.rcParams['text.usetex'] = True def Plot_MRcurves(MRIcurves, num): indices = numpy.linspace(0, len(MRIcurves)-1, num, dtype=int) fig, ax = pyplot.subplots(1,1, figsize=(7,6)) for e, i in enumerate(indices): M, R, I, rhoc = MRIcurves[i] ax.plot(R, M, c='blue', zorder=0) ax.set_xlim(5, 16) ax.set_xlabel('Radius (km)') ax.set_ylabel(r'Mass (M$_{\odot}$)') pyplot.show() def Plot_PosteriorInput(PosteriorInput, M, R): mi = numpy.linspace(0.2, 3.6, 400) ri = numpy.linspace(5, 16, 400) mig, rig = numpy.meshgrid(mi, ri) pig = Pobs(mig, rig, PosteriorInput) fig, ax = pyplot.subplots(1,1, figsize=(7,6)) for i in range(len(pig)): pi = numpy.concatenate(pig[i]).ravel() ax.contour(rig, mig, pig[i], linewidth=2.0, rstride=1, cstride=1, vmin=numpy.amin(pig[i]), vmax=numpy.amax(pig[i]), levels=numpy.array([find_CI_level(pi)[0]]), linestyles='--', colors=['red'], extend='max') ax.plot(R, M, c='black') ax.set_xlim(min(R)-3., 17) ax.set_ylim(0., max(M)+.3) ax.set_xlabel('Radius (km)') ax.set_ylabel(r'Mass (M$_{\odot}$)') ax.xaxis.set_tick_params(labelsize=15) ax.yaxis.set_tick_params(labelsize=15) pyplot.show() def Plot_Prob(Prob, Parameters): df = pandas.DataFrame({'P1':Parameters[:,0],'P2':Parameters[:,1], 'P3':Parameters[:,2], 'Prob':Prob}) fig, ax = pyplot.subplots(1,2, figsize=(12,5)) for i, e in enumerate([['P1', 'P2'], ['P2', 'P3']]): df2 = df.groupby([e[0], e[1]]).Prob.sum().reset_index() df2 = numpy.array(df2) values = df2[:,2] points = df2[:,0:2] values=abs(values) X = numpy.log10(points[:,0]) Y = numpy.log10(points[:,1]) Z = values xi = numpy.linspace(X.min(),X.max(),100) yi = numpy.linspace(Y.min(),Y.max(),100) zi = griddata((X, Y), Z, (xi[None,:], yi[:,None]), method='cubic') xig, yig = numpy.meshgrid(xi, yi) surface = ax[i].contour(xig, yig, zi, linewidths=1.5, rstrid=1, cstride=1, vmin=min(Z), vmax=max(Z), levels = find_CI_level(values), colors=('Grey', 'Steelblue')) fmt = {} strs = [r'2 $\sigma$', r'1 $\sigma$'] for l, s in zip(surface.levels, strs): fmt[l] = s ax[i].clabel(surface, inline=1, fontsize=11, fmt=fmt) if i==0: ### P1 P2 ### ax[0].set_xlabel('$\log($P$_{1}$) (dyn cm$^{-2}$)') ax[0].set_ylabel('$\log($P$_{2}$) (dyn cm$^{-2}$)') ax[0].set_yticks([34.5, 35., 35.5, 36.]) ax[0].set_xticks([33.5, 34., 34.5, 35.]) ax[0].text(34.253, 35.115, 'FPS', fontsize=12) ax[0].set_xlim(33.5, 35.) ax[0].set_ylim(34.2, 36.) if i==1: ### P2 P3 ### ax[1].set_xlabel('$\log($P$_{2}$) (dyn cm$^{-2}$)') ax[1].set_ylabel('$\log($P$_{3}$) (dyn cm$^{-2}$)') ax[1].set_xticks([34.5, 35., 35.5, 36.]) ax[1].set_yticks([35., 35.5, 36., 36.5, 37.]) ax[1].text(35.08, 35.89, 'FPS', fontsize=12) ax[1].set_xlim(34.2, 36.) ax[1].set_ylim(35., 37.) pyplot.tight_layout() pyplot.show()
3,918
32.784483
108
py
MORSE
MORSE-master/Create_MRcurves_Par.py
#! /usr/bin/env python # -*- coding: utf-8 -*- import sys import os import multiprocessing as mp import traceback import logging import gc import time from constants import c, G, Msun, rho_0, rho_ns, rho_1, rho_2, rho_3, P_0, gamma_0, dyncm2_to_MeVfm3, gcm3_to_MeVfm3, oneoverfm_MeV import numpy from argparse import ArgumentParser from tqdm import tqdm, trange import itertools from scipy.constants import pi from scipy.interpolate import interp1d, UnivariateSpline from scipy.signal import argrelextrema from scipy.integrate import odeint from matplotlib import pyplot def print_progressbar(i, N): pbwidth = 42 progress = float(i)/N block = int(round(pbwidth*progress)) text = "\rProgress: [{0}] {1:.1f}%".format( "#"*block + "-"*(pbwidth-block), progress*100) sys.stdout.write(text) sys.stdout.flush() if i == (N-1): print " .. done" def crust_EOS(): """ Interpolates the SLy EOS to use for the crust and calculates the minimum pressure. Returns: EOS (function): A function of density representing the EOS Inverse EOS (function): A function of pressure representing the EOS P_min (float): The minimum pressure that is tabulated. """ Pmin = 1e2 Pmax = SLYfit(14.3) rhotest = numpy.logspace(6, 16, 300) prestest = 10**SLYfit(numpy.log10(rhotest)) ptest = numpy.logspace(numpy.log10(Pmin), Pmax, 500) eos = UnivariateSpline(rhotest, prestest, k=3, s=0) inveos = UnivariateSpline(prestest, rhotest, k=3, s=0) return eos, inveos, Pmin def f0(x): return 1./(numpy.exp(x) + 1.) def SLYfit(rho): a = numpy.array([6.22, 6.121, 0.005925, 0.16326, 6.48, 11.4971, 19.105, 0.8938, 6.54, 11.4950, -22.775, 1.5707, 4.3, 14.08, 27.80, -1.653, 1.50, 14.67]) part1 = (a[0] + a[1]*rho + a[2]*rho**3.)/(1. + a[3]*rho) * f0(a[4]*(rho-a[5])) part2 = (a[6] + a[7]*rho)*f0(a[8]*(a[9]-rho)) part3 = (a[10] + a[11]*rho)*f0(a[12]*(a[13]-rho)) part4 = (a[14] + a[15]*rho)*f0(a[16]*(a[17] - rho)) return part1+part2+part3+part4 def eos(rho, P_1, P_2, P_3, eos_crust): """ The parameterized EOS. Args: rho (float): The density at which to evaluate the EOS in g/cm^3. P_1 (float): The first pressure parameter of the parameterization. P_2 (float): The second pressure parameter of the parameterization. P_3 (float): The third pressure parameter of the parameterization. eos_crust (function): The EOS for the low-density part, which inputs a mass density and returns a pressure. Returns: rho (float): The rest-mass density in g/cm^3. epsilon (float): The energy density in g/cm^3 """ rho_ns = 2.7e14 rho_1 = 1.85 * rho_ns rho_2 = 2. * rho_1 rho_3 = 2. * rho_2 gamma_1 = numpy.log10(P_1/P_0) / numpy.log10(rho_1/rho_0) gamma_2 = numpy.log10(P_2/P_1) / numpy.log10(rho_2/rho_1) gamma_3 = numpy.log10(P_3/P_2) / numpy.log10(rho_3/rho_2) k1 = P_0/(rho_0**gamma_1) pres1 = k1*rho_1**gamma_1 k2 = pres1/(rho_1**gamma_2) pres2 = k2*rho_2**gamma_2 k3 = pres2/(rho_2**gamma_3) #gamma0 = deriv_hpd(rho0) * rho0/P0 gamma0 = 2.7 e0 = rho_0 + P_0/c**2.0 * 1./(gamma0 - 1.) a1 = e0/rho_0 - 1. - k1/((gamma_1-1.)*c**2.0) * rho_0**(gamma_1-1.) e1 = (1. + a1)*rho_1 + pres1/(c**2.0 * (gamma_1 -1.)) a2 = e1/rho_1 - 1. - k2/((gamma_2-1.)*c**2.0) * rho_1**(gamma_2-1.) e2 = (1. + a2)*rho_2 + pres2/(c**2.0 * (gamma_2 -1.)) a3 = e2/rho_2 - 1. - k3/((gamma_3-1.)*c**2.0) * rho_2**(gamma_3-1.) if rho <= rho_0: pres = eos_crust(rho) gamma_05 = 1.7 epsilon = rho + pres/c**2. * 1./(gamma_05 - 1) if rho_0 < rho <= rho_1: pres = k1 * rho**gamma_1 epsilon = (1. + a1)*rho + pres/(c**2.0 * (gamma_1 -1.)) if rho_1 < rho <= rho_2: pres = k2 * rho**gamma_2 epsilon = (1. + a2)*rho + pres/(c**2.0 * (gamma_2 -1.)) if rho > rho_2: pres = k3 * rho**gamma_3 epsilon = (1. + a3)*rho + pres/(c**2.0 * (gamma_3 -1.)) return pres, epsilon def inveos(pres, P_1, P_2, P_3, eos_crust, inveos_crust, P_min): """ The inverse of the parameterized EOS. Args: pres (float) : The pressure at which to evaluate the inverse EOS in dyn/cm^2. P_1 (float) : The first pressure parameter of the parameterization. P_2 (float) : The second pressure parameter of the parameterization. P_3 (float) : The third pressure parameter of the parameterization. eos_crust (function) : The EOS for the low-density part, which inputs a mass density and returns a pressure. inveos_crust (function) : The inverse EOS for the low density part, which inputs a pressure and returns a mass density. P_min (float) : The minimum pressure for which the low-density EOS function is defined. Returns: rho (float) : The rest-mass density in g/cm^3. epsilon (float) : The energy density in g/cm^3 """ gamma_1 = numpy.log10(P_1/P_0) / numpy.log10(rho_1/rho_0) gamma_2 = numpy.log10(P_2/P_1) / numpy.log10(rho_2/rho_1) gamma_3 = numpy.log10(P_3/P_2) / numpy.log10(rho_3/rho_2) k1 = P_0/(rho_0**gamma_1) pres1 = k1*rho_1**gamma_1 k2 = pres1/(rho_1**gamma_2) pres2 = k2*rho_2**gamma_2 k3 = pres2/(rho_2**gamma_3) #gamma0 = deriv_hpd(rho0) * rho0/P0 gamma0 = 2.7 e0 = rho_0 + P_0/c**2.0 * 1./(gamma0 - 1.) a1 = e0/rho_0 - 1. - k1/((gamma_1-1.)*c**2.0) * rho_0**(gamma_1-1.) e1 = (1. + a1)*rho_1 + pres1/(c**2.0 * (gamma_1 -1.)) a2 = e1/rho_1 - 1. - k2/((gamma_2-1.)*c**2.0) * rho_1**(gamma_2-1.) e2 = (1. + a2)*rho_2 + pres2/(c**2.0 * (gamma_2 -1.)) a3 = e2/rho_2 - 1. - k3/((gamma_3-1.)*c**2.0) * rho_2**(gamma_3-1.) if pres <= P_0: rho = inveos_crust(pres) #rho = 10**inveos_crust(numpy.log10(pres)) gamma_05 = 1.7 epsilon = rho + pres/c**2. * 1./(gamma_05 - 1) if P_0 < pres <= P_1: rho = (pres/k1)**(1./gamma_1) epsilon = (1. + a1)*rho + pres/(c**2.0 * (gamma_1 -1.)) if P_1 < pres <= P_2: rho = (pres/k2)**(1./gamma_2) epsilon = (1. + a2)*rho + pres/(c**2.0 * (gamma_2 -1.)) if pres > P_2: rho = (pres/k3)**(1./gamma_3) epsilon = (1. + a3)*rho + pres/(c**2.0 * (gamma_3 -1.)) return rho, epsilon ### Define function to integrate def f(initial, r, P_1, P_2, P_3, eos_crust, inveos_crust, P_min): """ The TOV-equations to pass on to scipy's 'odeint'. Args: initial (array) : Array of two values, the inital pressure and the initial mass. r (float) : The radial coordinate of the neutron star (r = 0 is the center). P_1 (float) : The first pressure parameter of the parameterization. P_2 (float) : The second pressure parameter of the parameterization. P_3 (float) : The third pressure parameter of the parameterization. eos_crust (function) : The EOS for the low-density part, which inputs a mass density and returns a pressure. inveos_crust (function): The inverse EOS for the low density part, which inputs a pressure and returns a mass density. P_min (float) : The minimum pressure for which the low-density EOS function is defined. Returns: dpdr (float) : The derivative of the pressure with respect to the radial coordinate. dmdr (float) : The derivative of the mass with respect to the radial coordinate. """ pres, m = initial if pres < P_min: pres = P_min rho, eps = inveos(pres, P_1, P_2, P_3, eos_crust, inveos_crust, P_min) dmdr = 4.*pi*r**2.0 * eps if r==0.0: dpdr = 0.0 else: dpdr = -G * (eps + pres/c**2.) * (m + 4.*pi*r**3. * pres/c**2.) dpdr = dpdr/(r*(r - 2.*G*m/c**2.)) return dpdr, dmdr ### Function to solve the TOV-equations def tovsolve(rhocent, P_1, P_2, P_3, eos_crust, inveos_crust, P_min): """ Solves the TOV-equations using scipy's 'odeint' package. Args: rhocent (float) : The central density of the neutron star. This is the starting value of the differential integration. P_1 (float) : The first pressure parameter of the parameterization. P_2 (float) : The second pressure parameter of the parameterization. P_3 (float) : The third pressure parameter of the parameterization. eos_crust (function) : The EOS for the low-density part, which inputs a mass density and returns a pressure. inveos_crust (function): The inverse EOS for the low density part, which inputs a pressure and returns a mass density. P_min (float) : The minimum pressure for which the low-density EOS function is defined. Returns: M (float) : The mass of the neutron star. R (float) : The radius of the neutron star. """ dr = 800. r = numpy.arange(0.0, 2500000., dr) pcent = eos(rhocent, P_1, P_2, P_3, eos_crust)[0] m0 = 0.0 P0 = pcent y = P0, m0 psol = odeint(f, y, r, args=(P_1, P_2, P_3, eos_crust, inveos_crust, P_min)) indices = numpy.where(psol[:,0]>P_min) index = indices[-1][-1] M_max = psol[index][1]/Msun R_max = r[index]/100000 I = MomentInertia(psol[:,0][indices[0]], psol[:,1][indices[0]], r[indices[0]], P_min) return M_max, R_max, I ### Calculate the Masses and Radii for different central pressures def calculate_MR(logrhomin, logrhomax, n, P_1, P_2, P_3, eos_crust, inveos_crust, P_min): """ Calculate a mass-radius curve by solving the TOV-equations for different central densities. Args: logrhomin (float) : The lower limit of the central density in log(g/cm^3). logrhomax (float) : The upper limit of the central density in log(g/cm^3), based on causality. n (int) : The number of points used in logspace to create the MR-curve. P_1 (float) : The first pressure parameter of the parameterization. P_2 (float) : The second pressure parameter of the parameterization. P_3 (float) : The third pressure parameter of the parameterization. eos_crust (function) : The EOS for the low-density part, which inputs a mass density and returns a pressure. inveos_crust (function): The inverse EOS for the low density part, which inputs a pressure and returns a mass density. P_min (float) : The minimum pressure for which the low-density EOS function is defined. Returns: Masses (array) : An array of length 'n' with the masses of the neutron stars. Radii (array) : An array of length 'n' with the radii of the neutron stars. Inert (array) : An array of length 'n' with the moments of inertia of the stars. rhocent (array) : An array of length 'n' with the central densities of each star. """ rhocent = numpy.logspace(logrhomin, logrhomax, n) Masses = numpy.zeros(len(rhocent)) Radii = numpy.zeros(len(rhocent)) Inert = numpy.zeros(len(rhocent)) for j, k in enumerate(rhocent): Masses[j], Radii[j], Inert[j] = tovsolve(k, P_1, P_2, P_3, eos_crust, inveos_crust, P_min) Masses = numpy.array(Masses) Radii = numpy.array(Radii) Inert = numpy.array(Inert) #Masses = Masses[Masses>0.] #Radii = Radii[Radii>0.] #Inert = Inert[Inert>0.] return Masses, Radii, Inert, rhocent def MomentInertia(P, M, r, P_min): """ Calculate the moment of inertia of a star given a central density and an EOS. Args: P (array) : The pressure profile throughout the star as a function of the radial coordinate. M (array) : The mass profile throughout the star as a function of the radial coordinate. r (array) : The radial coordinates. P_min (float): The minimum pressure for which the EOS is defined. Returns: I (float) : The moment of inertia of the star in (Msun km^2) """ curveM = UnivariateSpline(r, M, k=3, s=0) curveP = UnivariateSpline(r, P, k=3, s=0) Mns = M[-1] Rns = r[-1] dr = 100 rx = numpy.arange(min(r)+.1, max(r), dr) nustart = numpy.log(1. - 2.*G*Mns/(c**2. * Rns)) curvenu = UnivariateSpline(rx, nu(rx, curveM, curveP), k=3, s=0) nufunc = curvenu.antiderivative(1) js = numpy.exp(-.5*(nufunc(rx)-nufunc(Rns)+nustart))*numpy.sqrt((1. - 2.*G*curveM(rx)/(c**2. *rx))) curvej = UnivariateSpline(rx, js, k=3, s=0) derivj = curvej.derivative(1) sol = odeint(omega, [1., 0.], rx, args=(curvej, derivj)) w = sol[:,0] dw = sol[:,1] curvew = UnivariateSpline(rx, w, k=3, s=0) curvedw = UnivariateSpline(rx, dw, k=3, s=0) J = 1./6. * Rns**4. * curvedw(Rns) W = curvew(Rns) + 2.*J/(Rns**3.) #I = quad(Inertia, 0.1, Rns, args=(derivj, curvew, W))[0] *2.*c**2. /(3.*G) I = (1. - curvew(Rns)/W)*Rns**3. * c**2./(2.*G) return I*10**(-10)/Msun def nu(r, curveM, curveP): dvdr = 2*G/c**2. * (curveM(r) + 4.*numpy.pi*r**3. * curveP(r)/c**2.)/(r**2. * (1. - 2.*G*curveM(r)/(r*c**2.))) return dvdr def omega(initial, r, curvej, derivj): x1, x2 = initial if r==0.0: dx1 = 0. dx2 = 0. else: dx1 = x2 dx2 = - 4./(r*curvej(r)) * derivj(r) * x1 - 4./r * x2 -1./curvej(r) *derivj(r) *x2 return dx1, dx2 def Inertia(r, derivj, curvew, W): return -r**3. * derivj(r)*curvew(r)/W def calculate_MR_all(parameters, eos_crust, inveos_crust, task=0, logrhomin=14.4, logrhomax=16.5, n=100, P_min=0.0): """ Calculate the masses, radii and moments of inertia for all input parameters by solving the TOV-equations for different central densities. Args: parameters : The array of parameters for which to solve the TOV-equations. eos_crust (function) : The EOS for the low-density part, which inputs a pressure and returns a mass density. inveos_crust (function) : The inverse EOS for the low density part, which inputs a mass density and returns a pressure. task : At which line to print the progressbar, default is 0. logrhomin (float) : The lower limit of the central density in log(g/cm^3), default is 14.4. logrhomax (float or array): The upper limits of the central density in log(g/cm^3), based on causality. If float, the same value for all EoSs is used, if array, every entry should correspond to the maximum central density of an EoS. n (int) : The number of points used in logspace to create the MR-curve. eos_crust (function) : The EoS for the low density part of the star. The function should take density as input and outputs pressure. inveos_crust (function) : The inverse of the EoS for the low density part of the star. The function should take pressure as input and output density. P_min (float) : The lowest density for which eos_crust and inveos_crust is defined, default is 0.0. Returns: MRIcurves (ndarray) : An array of length (EoS) with for each EoS an array of masses, radii, moments of inertia and the corresponding central densities. Parameters (ndarray) : An array of length (EoS) with the parameters for which the TOV equations generated stable mass-radius curves. """ MR_curves = [] Error_params = [] with tqdm(total=len(parameters), position=task, desc='Process %d' %(task), leave=False) as pbar: for j in range(len(parameters)): P_1 = parameters[j,0] P_2 = parameters[j,1] P_3 = parameters[j,2] try: if isinstance(logrhomax, float): masses, radii, inert, rhocent = calculate_MR(logrhomin, logrhomax, n, P_1, P_2, P_3, eos_crust, inveos_crust, P_min) else: masses, radii, inert, rhocent = calculate_MR(logrhomin, logrhomax[j], n, P_1, P_2, P_3, eos_crust, inveos_crust, P_min) pbar.update(1) except UnboundLocalError: Error_params.append(j) continue else: locmin = argrelextrema(masses, numpy.less)[0] #Check for EOS with local minima if not len(locmin)==0: Error_params.append(j) continue locmax = argrelextrema(masses, numpy.greater)[0] if not len(locmax)==0: if locmax[0] < len(masses)-1: #check to see if there is a sharp kink in the MR-curve right = locmax[0] + 1 left = locmax[0] - 1 if abs(radii[right]-radii[left]) < 0.12: Error_params.append(j) continue MR_curves.append([masses[0:locmax[0]+1], radii[0:locmax[0]+1], inert[0:locmax[0]+1], rhocent[0:locmax[0]+1]]) else: MR_curves.append([masses, radii, inert, rhocent]) MR_curves = numpy.array(MR_curves) Error_params = numpy.array(Error_params) Parameters = numpy.delete(parameters, Error_params, axis=0) return MR_curves, Parameters #info = mp.get_logger().info def main(parameters, maxrho, outputMR, outputPs): #logger = mp.log_to_stderr() #logger.setLevel(logging.INFO) nproc = mp.cpu_count() - 1 nproc = max(1, nproc) print len(parameters) div_par = numpy.array_split(parameters, nproc) div_maxrho = numpy.array_split(maxrho, nproc) ntasks = nproc inputs = [[div_par[t], div_maxrho[t], t] for t in xrange(ntasks)] input_q = mp.Queue() output_q = mp.Queue() procs = [ mp.Process(target=worker, args=(input_q,output_q)) for i in xrange(nproc)] for i in xrange(ntasks): input_q.put(inputs[i]) for i in xrange(nproc): input_q.put('STOP') for p in procs: p.start() result = [] while ntasks > 0: result.append(output_q.get()) ntasks -= 1 for p in procs: p.join() result = numpy.array(sorted(result, key=lambda x: x[2])) result = numpy.delete(result, 2, axis=1) MRcurves = [] Parameters = [] for i in xrange(nproc): for j in range(len(result[i][0])): Parameters.append(result[i][1][j]) MRcurves.append([numpy.array(result[i][0][j][0]), numpy.array(result[i][0][j][1]), numpy.array(result[i][0][j][2]), numpy.array(result[i][0][j][3])]) Parameters = numpy.array(Parameters) MRcurves = numpy.array(MRcurves) numpy.save(outputPs, Parameters) numpy.save(outputMR, MRcurves) def worker(input_q, output_q): start = time.clock() while True: try: tmp = input_q.get() if 'STOP' == tmp : break parameters, maxrho, task = tmp eos_crust, inveos_crust, P_min = crust_EOS() MR_curves, Parameters = calculate_MR_all(parameters, eos_crust, inveos_crust, task, logrhomax=maxrho, n=30, P_min=P_min) output_q.put([MR_curves, Parameters, task]) except Exception as exception: trace = str(traceback.format_exc()) #info(trace) #end = (time.clock() - start) #info(end) return if __name__ == '__main__': parser = ArgumentParser() parser.add_argument("-f1", dest="OutputFileMR", help="write MRIcurves to FILE", metavar="FILE", required=True) parser.add_argument("-f2", dest="OutputFileParams", help="write parameters to FILE", metavar="FILE", required=True) args = parser.parse_args() parameters = numpy.load('input_parameters.npy') maxrho = numpy.load('max_rho.npy') maxrho = numpy.log10(maxrho) main(parameters, maxrho, args.OutputFileMR, args.OutputFileParams)
21,060
35.755672
139
py
MORSE
MORSE-master/InputPosteriors.py
import numpy def Create_Input_Posterior(percent, masses, radii, rho): Robs = radii Mobs = masses sigmaM = numpy.zeros(3) sigmaR = numpy.zeros(3) sigmaM[0] = percent[0]*Mobs[0] sigmaR[0] = percent[0]*Robs[0] sigmaM[1] = percent[1]*Mobs[1] sigmaR[1] = percent[1]*Robs[1] sigmaM[2] = percent[2]*Mobs[2] sigmaR[2] = percent[2]*Robs[2] distribution = [] for i in range(3): distribution.append([Mobs[i], Robs[i], sigmaM[i], sigmaR[i], rho[i]]) distribution = numpy.array(distribution) return distribution def find_CI_level(array): NaN_index = numpy.isnan(array) array[NaN_index] = 0.0 index_68 = numpy.where(numpy.cumsum(numpy.sort(array)[::-1]) < sum(array)*0.6827)[0] index_68 = numpy.argsort(array)[::-1][index_68] index_95 = numpy.where(numpy.cumsum(numpy.sort(array)[::-1]) < sum(array)*0.9545)[0] index_95 = numpy.argsort(array)[::-1][index_95] return min(array[index_95]), min(array[index_68]) def Pobs(M, R, distribution): obs = numpy.zeros(3, dtype=object) for l in range(3): Mobs = distribution[l][0] Robs = distribution[l][1] sigmaM = distribution[l][2] sigmaR = distribution[l][3] rho = distribution[l][4] obs[l] = numpy.exp(-1./(2.*(1.-rho**2.)) *\ ((R-Robs)**2. / sigmaR**2.0 + (M-Mobs)**2. / sigmaM**2.0 - \ 2.*rho*(R-Robs)*(M-Mobs)/(sigmaM*sigmaR))) return obs
1,586
27.339286
95
py
MORSE
MORSE-master/JacobianRho.py
import sys import os import multiprocessing as mp import traceback import logging import gc import time from argparse import ArgumentParser import numpy import pandas from matplotlib import pyplot from scipy.interpolate import UnivariateSpline def UniqueParams(Parameters): df = pandas.DataFrame({'P1':Parameters[:,0], 'P2':Parameters[:,1], 'P3':Parameters[:,2]}) df1 = df.drop_duplicates(['P1', 'P2']) df2 = df.drop_duplicates(['P1', 'P3']) df3 = df.drop_duplicates(['P2', 'P3']) df4 = pandas.concat([df1, df2, df3]) ParametersOut = numpy.array(df4.drop_duplicates()) return ParametersOut def calculate_deriv(t, MRcurves, changeParam, Params): radii = numpy.zeros(len(changeParam)) masses = numpy.zeros(len(changeParam)) rhocM = numpy.zeros(len(changeParam)) rhocR = numpy.zeros(len(changeParam)) #inert = numpy.zeros(len(changeParam)) wrong = [] right = [] back = numpy.empty((len(changeParam), 4)) back.fill(numpy.nan) for i, e in enumerate(changeParam): M, R, I, rc = MRcurves[e] rc = numpy.log10(rc) curveR = UnivariateSpline(rc, R, k=3, s=1e-4) curveM = UnivariateSpline(rc, M, k=3, s=1e-4) derivR = curveR.derivative(1) derivM = curveM.derivative(1) if min(rc) <= t <= max(rc): radii[i] = curveR(t) masses[i] = curveM(t) rhocM[i] = derivM(t) rhocR[i] = derivR(t) right.append(i) else: wrong.append(i) Params2 = numpy.delete(Params, wrong, axis=0) radii = numpy.delete(radii, wrong, axis=0) masses = numpy.delete(masses, wrong, axis=0) rhocM = numpy.delete(rhocM, wrong, axis=0) rhocR = numpy.delete(rhocR, wrong, axis=0) if Params2.size==0 or len(Params2)<4: return back[:,0:2], back[:,2], back[:,3] else: curvePR = UnivariateSpline(Params2, radii, k=3, s=1e-3) curvePM = UnivariateSpline(Params2, masses, k=3, s=1e-3) derivPR = curvePR.derivative(1) derivPM = curvePM.derivative(1) back[right] = numpy.dstack([rhocM, rhocR, derivPM(Params2), derivPR(Params2)]) return back[:,0:2], back[:,2], back[:,3] def calculate_jac(Parameters, ParamUnique, MRIcurves, rhoc): Jac = numpy.zeros((len(Parameters), 9)) Jac.fill(numpy.nan) for i in range(len(ParamUnique)): P1 = ParamUnique[i][0] P2 = ParamUnique[i][1] P3 = ParamUnique[i][2] Pvalues = [P1, P2, P3] combis = [[1, 2], [0, 2], [0, 1]] for j in range(3): h1, h2 = combis[j] change = numpy.where((Parameters[:,h1]==Pvalues[h1]) & (Parameters[:,h2]==Pvalues[h2]))[0] if change.size==0: continue else: Ps = numpy.log10(Parameters[change][:,j]) Jac[:,1:3][change], Jac[:,j+3][change], Jac[:,j+6][change] = calculate_deriv(rhoc, MRIcurves, change, Ps) return Jac info = mp.get_logger().info def main(Parameters, MRIcurves, rhoc, outputfile): logger = mp.log_to_stderr() logger.setLevel(logging.INFO) nproc = mp.cpu_count() - 1 nproc = max(1, nproc) div_rhoc = numpy.array_split(rhoc, nproc) ParamUnique = UniqueParams(Parameters) ntasks = nproc inputs = [[Parameters, ParamUnique, MRIcurves, div_rhoc[t], t] for t in xrange(ntasks)] input_q = mp.Queue() output_q = mp.Queue() procs = [ mp.Process(target=worker, args=(input_q,output_q)) for i in xrange(nproc)] for i in xrange(ntasks): input_q.put(inputs[i]) for i in xrange(nproc): input_q.put('STOP') for p in procs: p.start() result = [] while ntasks > 0: result.append(output_q.get()) ntasks -= 1 for p in procs: p.join() result = numpy.array(sorted(result, key=lambda x: x[1])) result = numpy.delete(result, 1, axis=1) new_jac = [] for i in xrange(nproc): for j in range(len(result[i][0])): new_jac.append(result[i][0][j]) new_jac = numpy.array(new_jac) Jacobian = numpy.zeros((len(Parameters), len(rhoc), 9)) for i in range(len(Parameters)): for j in range(len(rhoc)): Jacobian[i][j] = new_jac[j][i] indices = numpy.invert(numpy.any(numpy.isnan(Jacobian[i][:,1:]), axis=1)) Jacobian[i][:,0][indices] = rhoc[indices] numpy.save(outputfile, Jacobian) def worker(input_q, output_q): start = time.clock() while True: try: tmp = input_q.get() if 'STOP' == tmp : break Parameters, ParamUnique, MRIcurves, rhoc, task = tmp Jacobian = [] for i, e in enumerate(rhoc): #info(e) jacpart = calculate_jac(Parameters, ParamUnique, MRIcurves, e) Jacobian.append(jacpart) output_q.put([Jacobian, task]) except Exception as exception: trace = str(traceback.format_exc()) info(trace) end = (time.clock() - start) info(end) return if __name__ == '__main__': parser = ArgumentParser() parser.add_argument("-f", dest="outputFile", help="write jacobian to FILE", metavar="FILE", required=True) parser.add_argument("-i1", dest="inputMRIcurves", help="use as input MRIcurves", required=True) parser.add_argument("-i2", dest="inputParams", help="use as input Parameters", required=True) args = parser.parse_args() Parameters = numpy.load(args.inputParams) MRIcurves = numpy.load(args.inputMRIcurves) rhoc = numpy.linspace(14.3, 16., 40) print len(Parameters), len(MRIcurves) main(Parameters, MRIcurves, rhoc, args.outputFile)
6,006
27.334906
123
py
UEDGE
UEDGE-master/setup.py
#!/usr/bin/env python # To use: # python setup.py install # import sys import os import os.path import string import site from Forthon.compilers import FCompiler import getopt import logging version='8.0.4.1' try: os.environ['PATH'] += os.pathsep + site.USER_BASE + '/bin' import setuptools import distutils from distutils.core import setup from distutils.core import Extension from distutils.dist import Distribution from distutils.command.build import build from distutils.command.install import install from subprocess import call import numpy except: raise SystemExit("Distutils problem") optlist, args = getopt.getopt(sys.argv[1:], 'gt:F:', ['parallel', 'petsc', 'omp']) machine = sys.platform debug = 0 fcomp = None parallel = 0 petsc = 0 for o in optlist: if o[0] == '-g': debug = 1 elif o[0] == '-t': machine = o[1] elif o[0] == '-F': fcomp = o[1] elif o[0] == '--parallel': parallel = 1 elif o[0] == '--petsc': petsc = 1 elif o[0] == '--omp': os.putenv("OMP","1") if petsc == 1 and os.getenv('PETSC_DIR') == None: raise SystemExit("PETSc requested but PETSC_DIR not set") if os.getenv('PETSC_DIR') != None: petsc = 1 if petsc == 1 and os.getenv('PETSC_ARCH') == None: raise SystemExit("PETSc requested but PETSC_ARCH not set") sys.argv = ['setup2.py']+args fcompiler = FCompiler(machine=machine, debug=debug, fcompname=fcomp) class uedgeInstall(build): def run(self): install.run(self) logging.basicConfig(stream=sys.stderr,level=logging.INFO) log = logging.getLogger() log.info("test") class uedgeBuild(build): def run(self): # with python2 everything is put into a single uedgeC.so file if sys.hexversion < 0x03000000: raise SystemExit("Python versions < 3 not supported") else: if petsc == 0: call(['make', '-f','Makefile.Forthon']) else: call(['make', '-f', 'Makefile.PETSc']) build.run(self) class uedgeClean(build): def run(self): if sys.hexversion < 0x03000000: raise SystemExit("Python versions < 3 not supported") else: if petsc == 0: call(['make', '-f', 'Makefile.Forthon', 'clean']) else: call(['make', '-f', 'Makefile.PETSc', 'clean']) uedgepkgs = ['aph', 'api', 'bbb', 'com', 'flx', 'grd', 'svr', 'wdf', 'ncl'] def makeobjects(pkg): return [pkg+'_p.o', pkg+'pymodule.o'] uedgeobjects = [] # add here any extra dot o files other than pkg.o, pkg_p.o if sys.hexversion < 0x03000000: raise SystemExit("Python versions < 3 not supported") else: dummydist = Distribution() dummydist.parse_command_line() dummybuild = dummydist.get_command_obj('build') dummybuild.finalize_options() builddir = dummybuild.build_temp uedgeobjects = map(lambda p: os.path.join(builddir, p), uedgeobjects) if os.getenv('PACT_DIR') != None: library_dirs = fcompiler.libdirs + [ os.path.join(os.getenv('PACT_DIR'), 'lib')] libraries = ['pdb', 'pml', 'score', 'blas', 'm'] + fcompiler.libs else: library_dirs = fcompiler.libdirs libraries = fcompiler.libs if petsc: # PETSC_DIR = '/homes/mccomic/petsc-uedge' # PETSC_ARCH = 'linux-uedge' PETSC_DIR = os.getenv('PETSC_DIR') PETSC_ARCH = os.getenv('PETSC_ARCH') library_dirs = fcompiler.libdirs + \ [os.path.join(PETSC_DIR, PETSC_ARCH, 'lib')] libraries = ['petscts', 'petscsnes', 'petscksp', 'petscdm', 'petscmat', 'petscvec', 'petsc', 'HYPRE', 'mpich', 'lapack', 'blas', 'X11', 'pthread', 'rt', 'stdc++', 'm'] + fcompiler.libs libraries = ['petsc'] + fcompiler.libs if parallel: library_dirs = fcompiler.libdirs + ['/usr/lpp/ppe.poe/lib'] libraries = fcompiler.libs + ['mpi'] # uedgeobjects = uedgeobjects + ['/usr/local/mpi/ifc_farg.o'] with open('pyscripts/__version__.py','w') as ff: ff.write("__version__ = '%s'\n"%version) with open('pyscripts/__src__.py','w') as ff: ff.write("__src__ = '%s'\n"%os.getcwd()) define_macros=[("WITH_NUMERIC", "0"), ("FORTHON_PKGNAME", '\"uedgeC\"'), ("FORTHON","1")] # check for readline rlncom = "echo \"int main(){}\" | gcc -x c -lreadline - " rln = os.system(rlncom) if rln == 0: define_macros = define_macros + [("HAS_READLINE","1")] os.environ["READLINE"] = "-l readline" libraries = ['readline'] + libraries setup(name="uedge", version=version, author='Tom Rognlien', author_email="trognlien@llnl.gov", maintainer='Bill Meyer', maintainer_email='meyer8@llnl.gov', description="2D Fluid simulation of plasma and neutrals in magnetic fusion devices", platforms="Unix, Windows (cygwin), Mac OSX", packages=['uedge'], package_dir={'uedge': 'pyscripts'}, # include_package_data=True, scripts=['pyscripts/pdb2hdf5', 'pyscripts/bas2py', 'pyscripts/hdf52pdb'], ext_modules=[Extension('uedge.uedgeC', ['uedgeC_Forthon.c', os.path.join(builddir, 'Forthon.c'), 'com/handlers.c', 'com/vector.c','bbb/exmain.c'], include_dirs=[builddir, numpy.get_include()], library_dirs=library_dirs, libraries=libraries, define_macros=define_macros, extra_objects=uedgeobjects, extra_link_args=['-g','-DFORTHON'] + fcompiler.extra_link_args, extra_compile_args=fcompiler.extra_compile_args )], cmdclass={'build': uedgeBuild, 'clean': uedgeClean}, test_suite="pytests", install_requires=['forthon', 'easygui'], # note that include_dirs may have to be expanded in the line above classifiers=['Programming Language :: Python', 'Programming Language :: Python :: 3'] )
6,228
30.780612
90
py
UEDGE
UEDGE-master/localrules.py
#rules for converting mppl to f90 # This is generic down to the UEDGE section import copy import os def Use2use(s): #Return s if this is a comment if (not s[0].isspace()) and s[0] != "U": return s #Do a substitution if line contains "Use" and it is not #part of a comment if (s.find("Use")+1): if (s.find("!")==-1) or (s.find("Use")<s.find("!")): sout=s.replace("("," ") sout=sout.replace(")"," ") sout=sout.replace("Use"," use") return sout return s def Allot(s): # converts MPPL allot to F90 allocate if (s.find("allot")+1): if (s.find("call allot")+1): s=s.replace("call allot","allocate") s=s.replace('"','') s=s.replace("'","") s=s.replace(",","(") s=s.replace(")","))") return s s=s.replace(", allot","") s=s.replace(",allot","") return s def Nopdb(s): if os.getenv('PACT_DIR') == None: if s.startswith("c!nopdb"): s=s.replace("c!nopdb"," ") return s def Petsc(s): if os.getenv('PETSC_DIR') != None: if os.getenv('PARALLEL') == None: if s.startswith("cunipetsc"): s=s.replace("cunipetsc","") if s.startswith("cpetsc"): s=s.replace("cpetsc","") return s def Omp(s): if os.getenv('OMP') != None: if s.startswith("c!omp"): s=s.replace("c!omp"," ") return s saved_dec=0 in_uses=0 savedlines=[] def MoveDecs(s): global saved_dec,in_uses # Return if this is a comment if (not s[0].isspace()) and (not saved_dec) and (not in_uses): return s # collect lines for declarations # if we find an "implicit none" statement, store it and remove the line sls=s.lstrip().lower() indfunc=sls.find("function") indcom=sls.find("!") functest= (indfunc == -1 or -1<indcom<indfunc) # tests to exclude "real function" but allow "real ! function as # part of declaration block if (sls[0:8]=="implicit") or (sls[0:4]=="real" and functest) \ or (sls[0:7]=="integer" and functest) \ or (sls[0:9]=="character") or (sls[0:9]=="parameter") or \ (sls[0:8]=="external") or (sls[0:9] == "intrinsic") or \ (sls[0:7]=="logical" and functest) or (sls[0:9]=="dimension") or \ (sls[0:4] == "data"): savedlines.append(s) saved_dec=1 in_uses=0 return None # if we are in the midst of declarations, save also comments (except for # "Common block") and continuations and blank lines as part of # what is moved.) if (saved_dec==1) and (sls == "" or s[0].lower() == "c" or sls[0]=="!" \ or s[0]=="*") and (in_uses == 0) \ and (s.find("Common block")==-1): savedlines.append(s) return None # Check for continuation line in midst of declarations if (saved_dec==1) and (len(s)>6): if (s[5].isspace() == 0): savedlines.append(s) return None if (sls[0:3] == "use"): in_uses=1 if (saved_dec==1) and (sls != "") and s[0] != "c" and sls[0] != "!" and \ (sls[0:3] != "use"): #This is our first executable statement. Add it to our saved # declarations lines and return them now templines = copy.copy(savedlines) templines.append(s) #empty out savedlines del savedlines[0:] saved_dec = 0 in_uses=0 return templines return s inelseif = 0 savedelselines="" def Elseifthen(s): # put a "then" at the end of an elseif if it isn't already there # need to check to see if next line is a continue global inelseif,savedelselines # return s if this is a comment if (not s[0].isspace()): return s if s.find("elseif")+1: if s.find("then")+1: return s # set a flag that we are in an "in-else-if" block that needs work inelseif = 1 # If there is no "then" we need to save it to test if the next # line is a continuation savedelselines = s return(None) if (inelseif and len(s)>6 and not s[5].isspace()): # This is a continue line, add it to savedelselines savedelselines += s return(None) if (inelseif and (len(s)<6 or s[5].isspace())): # No longer in a continue, so process lines if savedelselines.find("then")+1: savedelselines += s inelseif = 0 return savedelselines if savedelselines.split("\n")[-2].find("!")+1: # if last line in saved lines has a comment, # find index of last comment sign last = savedelselines.rfind("!") savedelselines=savedelselines[0:last]+ \ " then "+savedelselines[last:] + s inelseif=0 return savedelselines #Otherwise the last line has no comment so insert "then" at end savedelselines = savedelselines[0:-1]+" then\n" + s inelseif=0 return savedelselines return s M2Fsubrules = [("#","!"),Use2use, ("c!ifdef","#ifdef"), ("c!else","#else"), ("c!endif","#endif"), ("(Size4)","(kind=4)::"), (":: function"," function"), (" break "," exit "), (" break\n"," exit\n"), ("while (","do while ("), ("endwhile","end do"), (" call ruthere","c call ruthere"), ("c!include ","#include "), Nopdb, Petsc, Omp, Allot, Elseifthen, MoveDecs ] #------------------------------------- # Special for UEDGE wordsizectr=0 def grdproc(s): # process to eliminate ifelse write construction global wordsizectr if (s.find("ifelse([WORDSIZE]")+1): s="#if WORDSIZE == 64\n 2001 format(1p3e23.15)\n#else\n 2001 format(1p3d23.15)\n#endif\n" # s="#ifndef WORDSIZE\n 2001 format(1p3e23.15)\n#else\n 2001 format(1p3d23.15)\n#endif\n" wordsizectr=4 return s elif wordsizectr > 0: wordsizectr -= 1 return None else: wordsizectr = 0 return s M2Fsubrules.insert(6,grdproc) M2Fsubrules.insert(3,("do i1=","do_i1: do i1=")) M2Fsubrules.insert(4,("break (2) ! exit do_i1","exit do_i1")) M2Fsubrules.insert(5,("enddo ! do_i1","enddo do_i1")) M2Fsubrules.insert(6,("float","real")) M2Fsubrules.insert(6,("dfloat","real"))
6,579
32.917526
97
py
UEDGE
UEDGE-master/pyexamples/d3dHsm/runcase.py
#-import uedge from uedge import * #-import hdf5 routines from uedge.hdf5 import * #-import graphics, math, etc. import matplotlib.pyplot as plt import matplotlib.gridspec as gridspec import numpy as np import os #-import some utilities for using OS ###execfile(os.path.join(os.environ['HOME'], 'utils/python/osfun.py')) #-in .bashrc: "export PYLIB=/home/umansky1/PyUEDGE/uedge/pylib" execfile(os.environ['PYLIB']+"/plotmesh.py") execfile(os.environ['PYLIB']+"/plotcontour.py") execfile(os.environ['PYLIB']+"/plotvar.py") execfile(os.environ['PYLIB']+"/paws.py") execfile(os.environ['PYLIB']+"/osfun.py") plt.ion() #-read UEDGE settings execfile("rd_d3dHsm_in.py") #-do a quick preliminary run to set all internals bbb.restart=0; bbb.ftol=1e10; bbb.dtreal = 1e-6; bbb.exmain() #-show grid plotmesh(iso=1) wait = raw_input("PAUSING, PRESS ENTER TO CONTINUE...") #-run to steady state bbb.restart=1; bbb.ftol=1e-8; bbb.isbcwdt=1 bbb.dtreal = 1e-14; bbb.itermx=30; bbb.exmain() bbb.t_stop=1e0 bbb.rundt() bbb.dtreal=1e20; bbb.isbcwdt=0; bbb.exmain() #-show some results plotvar(bbb.te/bbb.ev) #-export the solution in hdf5 file hdf5_save('mycase.h5') #-can be imported with this command #hdf5_restore('mycase.h5') ###-refine the grid, interpolate to new grid, and restart: #com.nxleg[0,0]=20; bbb.newgeo=1; bbb.icntnunk=0 #bbb.dtreal = 1e-14; bbb.isbcwdt=1; bbb.itermx=30; bbb.exmain() ###-time advance by another second #bbb.t_stop=2e0; bbb.rundt() ###-now to steady state (infinite time) #bbb.dtreal=1e20; bbb.isbcwdt=0; bbb.exmain() ###-show some results #plotvar(bbb.te/bbb.ev)
1,603
21.591549
70
py
UEDGE
UEDGE-master/pyexamples/d3dHsm/rd_d3dHsm_in.py
# # ########################################################################### # DESCRIPTION OF PROBLEM (d3dHsm) from FACETS test suite: # DIII-D single-null geometry with 5 variables (ni,upi,te,ti,ng) and a # (16+2)*(8+2)=18x10 [poloidal*radial] mesh yielding 900 variables. # Solver used is Newton Krylov (svrpkg="nksol") and preconditioner uses a # direct banded solver for the LU decomposition (premeth="banded"). Iterates # to steady-state solution from an initial profile file (HF5). ########################################################################### ###import uedge from uedge import * # Set the geometry bbb.mhdgeo = 1 #=1 use MHD equilibrium files #flx.aeqdskfname = "aeqdskd3d" #name of EFIT 'a' file for flux-surface mesh #flx.geqdskfname = "neqdskd3d" #name of EFIT 'g' or 'n' file for flux-sur mesh flx.psi0min1 = 0.98 #normalized flux on core bndry flx.psi0min2 = 0.98 #normalized flux on pf bndry flx.psi0sep = 1.00001 #normalized flux at separatrix flx.psi0max = 1.07 #normalized flux on outer wall bndry bbb.ngrid = 1 #number of mesh sequenc. (always set to 1) com.nxleg[0,0] = 4 #pol. mesh pts from inner plate to x-point com.nxcore[0,0] = 4 #pol. mesh pts from x-point to top on inside com.nxcore[0,1] = 4 #pol. mesh pts from top to x-point on outside com.nxleg[0,1] = 4 #pol. mesh pts from x-point to outer plate com.nysol[0] = 6 #rad. mesh pts in SOL com.nycore[0] = 2 #rad. mesh pts in core # Finite-difference algorithms (upwind, central diff, etc.) bbb.methn = 33 #ion continuty eqn bbb.methu = 33 #ion parallel momentum eqn bbb.methe = 33 #electron energy eqn bbb.methi = 33 #ion energy eqn bbb.methg = 33 #neutral gas continuity eqn # Boundary conditions bbb.ncore[0] = 2.5e19 #hydrogen ion density on core ## iflcore = 0 #flag; =0, fixed Te,i; =1, fixed power on core bbb.tcoree = 100. #core Te bbb.tcorei = 100. #core Ti bbb.tedge = 2. #fixed wall,pf Te,i if istewcon=1, etc bbb.recycp[0] = 0.8 #hydrogen recycling coeff at plates # Transport coefficients (m**2/s) bbb.difni[0] = 1. #D for radial hydrogen diffusion bbb.kye = 1. #chi_e for radial elec energy diffusion bbb.kyi = 1. #chi_i for radial ion energy diffusion bbb.travis[0] = 1. #eta_a for radial ion momentum diffusion # Flux limits bbb.flalfe = 0.21 #electron parallel thermal conduct. coeff bbb.flalfi = 0.21 #ion parallel thermal conduct. coeff bbb.flalfv = 1. #ion parallel viscosity coeff bbb.flalfgx = 1.e20 #neut. gas in poloidal direction bbb.flalfgy = 1.e20 #neut. gas in radial direction # Solver package bbb.svrpkg = "nksol" #Newton solver using Krylov method bbb.premeth = "banded" #Solution method for precond. Jacobian matrix # Restart from a HDF5 or PDB savefile bbb.restart = 1 #Begin from savefile, not estimated profiles bbb.allocate() #allocates storage for arrays #from uedge.hdf5 import * #hdf5_restore("d3dHsm.h5") if (0): ue.restore("d3dHsm.h5") bbb.dtreal = 1e20; bbb.exmain() else: #-set up some initial state ###ev=1.6022e-19 bbb.ngs=1e14; bbb.ng=1e14 bbb.nis=1e20; bbb.ni=1e20 bbb.ups=0.0; bbb.up=0.0 bbb.tes=bbb.ev; bbb.te=bbb.ev bbb.tis=bbb.ev; bbb.ti=bbb.ev # Atomic data switches com.istabon = 0 #-analytic rates ###com.istabon = 10 #=10 specifics hydrogen data file ehr2.dat
3,347
36.617978
80
py
UEDGE
UEDGE-master/pyexamples/box2/plotcontour.py
import matplotlib import matplotlib.gridspec as gridspec gs=gridspec.GridSpec(2, 2) plt.figure(10) plt.subplot(gs[0,0]) CS = plt.contour(com.zm[:,:,0], com.rm[:,:,0], bbb.te/ev) plt.clabel(CS, inline=1, fontsize=10) params = {'mathtext.default': 'regular' } plt.rcParams.update(params) plt.title('T$\mathregular{_e}$ [ev]') plt.ylabel('R [m]') plt.grid(True) plt.subplot(gs[0,1]) CS = plt.contour(com.zm[:,:,0], com.rm[:,:,0], bbb.ti/ev) plt.clabel(CS, inline=1, fontsize=10) plt.title('T$\mathregular{_i}$ [ev]') plt.grid(True) plt.subplot(gs[1,0]) CS = plt.contour(com.zm[:,:,0], com.rm[:,:,0], bbb.ni[:,:,0]/1e20) plt.clabel(CS, inline=1, fontsize=10) plt.title('N$\mathregular{_i}$/1e20 [m-3]') plt.xlabel('Z [m]') plt.ylabel('R [m]') plt.grid(True) plt.subplot(gs[1,1]) CS = plt.contour(com.zm[:,:,0], com.rm[:,:,0], bbb.up[:,:,0]/1e3) plt.clabel(CS, inline=1, fontsize=10) plt.title('U$\mathregular{_p}$/1e3 [m/s]') plt.xlabel('Z [m]') plt.grid(True) plt.show()
987
23.097561
66
py
UEDGE
UEDGE-master/pyexamples/box2/runcase.py
#import uedge from uedge import * #-import hdf5 routines from uedge.hdf5 import * #-import graphics, math, etc. import matplotlib.pyplot as plt import matplotlib.gridspec as gridspec import numpy as np import os #-import some utilities for using OS ###execfile(os.path.join(os.environ['HOME'], 'utils/python/osfun.py')) ##-how to do this better? #-in .bashrc: "export PYLIB=/home/umansky1/PyUEDGE/uedge/pylib" execfile(os.environ['PYLIB']+"/plotmesh.py") execfile(os.environ['PYLIB']+"/plotcontour.py") execfile(os.environ['PYLIB']+"/plotvar.py") execfile(os.environ['PYLIB']+"/paws.py") execfile(os.environ['PYLIB']+"/osfun.py") #execfile("../../plotmesh.py") #execfile("../../pylib/plotvar.py") #execfile("../../pylib/plotr.py") #execfile("../../pylib/showrange.py") #execfile("../../pylib/paws.py") plt.ion() #-read UEDGE settings execfile("box2_in.py") #-do a quick preliminary run to set all internals bbb.restart=0; bbb.ftol=1e10; bbb.dtreal = 1e-6; bbb.exmain() #-show grid plotmesh() wait = raw_input("PAUSING, PRESS ENTER TO CONTINUE...") #-this should be done in uefacets #ev=1.6022e-19 if (0): hdf5_restore('mycase.h5') bbb.dtreal = 1e20; bbb.exmain() else: #-set up some initial state bbb.ngs=1e14; bbb.ng=1e14 bbb.nis=1e20; bbb.ni=1e20 bbb.ups=0.0; bbb.up=0.0 bbb.tes=bbb.ev; bbb.te=bbb.ev bbb.tis=bbb.ev; bbb.ti=bbb.ev # #-Note: if you make a gap here then it will change the logic of if-else! # #-run to steady state bbb.restart=1; bbb.ftol=1e-8; bbb.isbcwdt=1 bbb.dtreal = 1e-14; bbb.itermx=30; bbb.exmain() bbb.t_stop=1e0 bbb.rundt() bbb.dtreal=1e20; bbb.isbcwdt=0; bbb.exmain() hdf5_save('mycase.h5') ###execfile('plotcontour.py') plotcontour() paws() ##-now refine the solution on a larger grid #com.nycore[0]=2 #com.nysol[0]=6 #com.nxleg[0,1]=8 #com.nxcore[0,1]=8 #bbb.restart=1; bbb.newgeo=1; bbb.icntnunk=0 #bbb.dtreal = 1e-14; bbb.ftol=1e-10; #bbb.isbcwdt=1; bbb.itermx=30; bbb.exmain() #plotmesh() #paws() #bbb.t_stop=2e0; bbb.ftol=1e-8; bbb.rundt() #bbb.dtreal=1e20; bbb.isbcwdt=0; bbb.exmain() #execfile('plotcontour.py') #paws() #==========================================================================#
2,306
23.284211
76
py
UEDGE
UEDGE-master/pyexamples/box2/box2_in.py
########################################################################### # DESCRIPTION OF PROBLEM (box2): # # This is a Python version of the box case from Andreas Holm ########################################################################### #-Geometry bbb.mhdgeo=-1 #-set cartesian geometry bbb.isfixlb=2 #left boundary as sym. plane; no flux at cut grd.radx= 4.e-2 #-outer "radial" wall grd.rad0=0.0 #-location of 'radial' separ'x for cylinder or slab grd.radm=-1.e-2 #-minimum "radial" position grd.za0 = 0. #-poloidal symmetry plane location grd.zax=3.0 #-poloidal location of divertor plate grd.zaxpt=2.25 #-poloidal location of x-point grd.alfyt=-2.0 #radial nonuniformity factor; < 0 => expanding grd.alfxt=2.76 #poliodal nonuniformity factor; to make smooth #transition to exp. grid, alfxt should satisfy #the eqn dzun = (zax-zaxpt+dzun) # (1-exp(-alfxt/(nx-ixpt2+1))) / # (1-exp(-alfxt)) #where dzun = (zaxpt-za0)/ixpt2 and #ixpt2 = ncore(1,2). grd.btfix = 2. #constant total B-field grd.bpolfix = .2 #constant poloidal B-field #-Grid bbb.gengrid=1; #-Note: for slab the grid is not saved in gridue bbb.ngrid=1 com.nycore[0]=2 com.nysol[0]=4 com.nxleg[0,1]=3 com.nxcore[0,1]=3 #-Boundary conditions bbb.isnicore[0]=1 #-same density at all core points bbb.ncore=1.1e19 #-density on core boundary bbb.iflcore=1 #if=1, specify core power bbb.tcoree=25.0 #-used if iflcore=0 bbb.tcorei=25.0 #-used if iflcore=0 bbb.pcoree = 2.5e4 #-used if iflcore=1 bbb.pcorei = 2.5e4 #-used if iflcore=1 bbb.recycp=0.98 #-recycling coef at plates if ndatlb,rb=0 bbb.albdsi=0.99 #-albedos at inner gas source locations bbb.albdso=0.99 #-albedos at inner gas source locations bbb.istepfc=0; bbb.istipfc=0 #-priv. flux has zero temp. deriv. bbb.istewc=0; bbb.istiwc=0 #-wall has zero temp. deriv. bbb.bcee = 4.; bbb.bcei = 2.5 #-energy transmission coeffs. bbb.bcen = 0. #-energy transmission coefficint for neutrals bbb.isupss = 0 #-parallel vel sonic bbb.isupcore = 0 #-parallel vel =0 on core bndry #-Transport coefficients bbb.difni=0.5 bbb.kye=0.7 bbb.kyi=0.7 bbb.travis=1.0 bbb.parvis=1.0 #-Flux limits bbb.flalfe=0.2 bbb.flalfi=0.2 bbb.flalfgx=1.e0 bbb.flalfgy=1.e0 bbb.flalfgxy=1.e0 bbb.flalfv=0.5 # Finite difference algorithms bbb.methe=33;bbb.methu=33;bbb.methg=33 bbb.methn=33;bbb.methi=33 #-Solver package bbb.svrpkg = "nksol" #Newton solver using Krylov method bbb.mfnksol=-3 bbb.epscon1=0.005 bbb.ftol=1e-10 bbb.premeth = "ilut" #Solution method for precond. Jacobian matrix bbb.runtim=1e-07 bbb.rlx=0.9 ###bbb.del=1.e-8 #-this one causes syntax error! #-Neutral gas propeties bbb.tfcx=5.;bbb.tfcy=5. #Franck-Condon temperatures bbb.cngfx=1.;bbb.cngfy=1. #turn-on grad(T_g) flux if =1 bbb.cngflox=1.;bbb.cngfloy=0. #turn-on drift with ions if =1 bbb.cngmom = 1. #ion-gas momentum transfer bbb.eion = 5. #birth energy of ions bbb.ediss = 10. #dissoc. energy lost from elecs (eion=2*ediss) bbb.isrecmon = 1 #=1 turns on recombination bbb.cfupcx=1.0 # factor multipling momentum cx bbb.cfticx=1.0 # factor multipling cx terms in ion energy eqn #-Parallel neutral momentum equation bbb.isupgon[0]=1 if (bbb.isupgon[0] == 1): bbb.isngon=0 com.ngsp=1 com.nhsp=2 ###bbb.ziin[com.nhsp-1]=1 bbb.ziin[0]=1 bbb.ziin[1]=0 #-the following are probably default, set them anyway to be sure bbb.cngmom=0 bbb.cmwall=0 bbb.cngtgx=0 bbb.cngtgy=0 bbb.kxn=0 bbb.kyn=0 #-Currents and potential parameters bbb.isphion=0 bbb.rsigpl=1.e-8 #anomalous cross-field conductivity bbb.cfjhf=0. #turn-on heat flow from current (fqp) bbb.jhswitch=0 #Joule Heating switch # Atomic physics packages #com.istabon=10 #DEGAS rates com.istabon=0 #-analytic rates #-Misc bbb.restart=0
4,169
27.758621
75
py
UEDGE
UEDGE-master/pyscripts/paws.py
def paws(): programPause = raw_input("Press the <ENTER> key to continue...")
81
26.333333
68
py
UEDGE
UEDGE-master/pyscripts/uedge_lists.py
""" This module uses some of the Forthon methods to provide routines for listing and searching the Uedge compiled packages. """ import uedge import re packages = [uedge.com,uedge.aph,uedge.api,uedge.bbb,uedge.flx,uedge.grd,uedge.svr,uedge.wdf,uedge.ncl] def packagename2object(package): for p in packages: if p.name() == package: return p return None def list_packages(objects=None): """ Return list of package string names or objects if object argument set. """ if objects != None: return packages pnames = [] for p in packages: pnames.append(p.name()) return pnames def list_package_variables(package,attribute='',vars=None): """ Return list of variable string names from package. package - string name of Uedge package. attribute='search string' can be either either the group name or an attribute. Search is case sensitive and must be exact. vars=[varlist] selection limited to varlist """ ret = [] if type(package) == type(''): p = packagename2object(package) if p != None: ret.extend(p.varlist(attribute)) else: ret.extend(package.varlist(attribute)) if vars == None: return ret else: return list(set(ret) & set(vars)) def list_variable(var): """ Print variable information of name passed as a string Do not include the package in the variable name. """ for p in packages: if var in p.varlist(): print(p.listvar(var)) def list_variables_glob(s,verbose=False,veryverbose=False,vars=None): """ Print variables where variable name contains string Case insensitive verbose=True will cause variable comment to print veryverbose=True will cause all variable info to print vars=[varlist] search limited to varlist """ ret = [] for p in packages: for var in p.varlist(): if s.upper() in var.upper(): if verbose: print(var+' : '+p.getvardoc(var)) if veryverbose: print(p.listvar(var)) ret.append(var) if vars == None: return ret else: return list(set(ret) & set(vars)) def list_variables_apropos(s,verbose=False,veryverbose=False,vars=None): """ Print variables where comment contains string Case insensitive verbose=True will cause variable comment to print veryverbose=True will cause all variable info to print vars=[varlist] search limited to varlist """ ret = [] for p in packages: for var in p.varlist(): if s.upper() in p.getvardoc(var).upper(): if verbose: print(var+' : '+p.getvardoc(var)) if veryverbose: print(p.listvar(var)) ret.append(var) if vars == None: return ret else: return list(set(ret) & set(vars)) def list_variables_regex(r,verbose=False,veryverbose=False,vars=None): """ Print variables where comment matches regular expression. verbose=True will cause variable comment to print veryverbose=True will cause all variable info to print vars=[varlist] search limited to varlist """ ret = [] for p in packages: for var in p.varlist(): if re.search(r,p.getvardoc(var)): if verbose: print(var+' : '+p.getvardoc(var)) if veryverbose: print(p.listvar(var)) ret.append(var) if vars == None: return ret else: return list(set(ret) & set(vars)) def varlistattr(a): """ Return list of variables with the given attribute. Includes the package prefix for use in file save functions. """ ret = [] for p in packages: for var in p.varlist(): if a in p.getvarattr(var).split(): ret.append(p.name()+'.'+var) return ret
4,004
30.289063
103
py
UEDGE
UEDGE-master/pyscripts/osfun.py
def date(): os.system("date") def ls(opts=""): os.system("ls " + opts) def more(fname): os.system("more " + fname) def pwd(): os.system("pwd") def cp(opts=""): os.system("cp " + opts) def mv(opts=""): os.system("mv " + opts)
254
13.166667
30
py
UEDGE
UEDGE-master/pyscripts/convert1.py
#!/usr/bin/env python # # $Id: convert1.py,v 7.1 2019/11/01 22:38:19 meyer8 Exp $ # # To try solving linear critical gradient import sys import os import getopt import string from . import convert from .convert import * # define the mppl to f90 class class M2F(generic): suffixin = "m" suffixout = "F" subrules = globalsubrules + M2Fsubrules def usage(): print("Usage: convert1.py -i <indir> -o <outdir> <infile>") r""" main(argv: array of strings) """ def main(argv): try: opts, args = getopt.getopt(sys.argv[1:], "hi:o:", ["help", "indir=", "outdir="]) except getopt.GetoptError: # print help information and exit: usage() sys.exit(2) indir = "." outdir = "." # Go through args for o, a in opts: if o in ("-h", "--help"): usage() sys.exit() if o in ("-i", "--indir"): indir = a elif o in ("-o", "--outdir"): outdir = a # print "args =", args # Extract the file fn = args[0] # print "Converting " + fn # Do the conversion convert.M2F = M2F m2f = M2F(indir, outdir) # m2f.outdir = outdir m2f.processfile(fn) return 0 if __name__ == '__main__': sys.exit(main(sys.argv))
1,164
17.492063
62
py
UEDGE
UEDGE-master/pyscripts/bas2py_rules.py
from uedge import * import uedge.uedge_lists as ul subrules = [ ['\(','['], ['\)',']'], [';','\n'], ['^!','#!'], ['^ *',''], ['^\t*',''], [r'\ballocate\b','bbb.allocate()'], [r'\bexmain\b','bbb.exmain()'], [r'\bexponseed\b','grd.exponseed()'], ] warnrules = [] def raw_string(s): s = s.encode('unicode-escape').decode() return s for p in ul.list_packages(): subrules.append([r'\b'+raw_string('package '+ p)+r'\b','from uedge import '+p]) po = ul.packagename2object(p) for v in ul.list_package_variables(p): subrules.append([r'\b'+raw_string(v)+r'\b',p+'.'+v]) if "Dimension:" in po.listvar(v): d = po.listvar(v).split("Dimension:")[1].split("\n") if "0:" in d[0]: warnrules.append([r'\b'+raw_string(v)+r'\b','base 0, '+d[0]]) subrules.append([r'\bbbb.del\b','bbb.delpy'])
878
25.636364
82
py
UEDGE
UEDGE-master/pyscripts/rdinitdt.py
# Setup file to run time-dependently using dtreal # Change dtreal for starting dt and savefname to change pfb file name # Once variables are set, read rdrundt to execute a time-dependent run from uedge import * i_stor = 0 nfe_tot = 0 savefn = "savedt.hdf5" # name of hdf5 savefile written every timestep bbb.rdtphidtr = 1e20 # ratio dtphi/dtreal bbb.ismfnkauto = 1 # if =1, mfnksol=3 for dtreal<dtmfnk3, otherwise=-3 bbb.dtmfnk3 = 5.e-4 # dtreal for mfnksol sign change if ismfnkauto=1 bbb.mult_dt = 3.4 # factor expanding dtreal after ii2max steps bbb.ii1max = 500 # number of changes to dtreal bbb.ii2max = 5 # number of timesteps at current dtreal bbb.itermxrdc = 7 # value of itermx used by rdcontdt bbb.incpset = 7 # iterations until Jacobian is recomputed bbb.ftol_dt = 1.e-5 # fnrm tolerance for the time-dependent steps bbb.ftol_min = 1e-9 # value of fnrm where time advance will stop bbb.dt_tot = 0. # tot time accumulated for run (output, not input) bbb.t_stop = 100. # value of dt_tot (sec) where calculation will stop bbb.dt_max = 100. # maximum time step for dtreal bbb.dt_kill = 1e-14 # min allowed time step; rdcontdt stops if reached bbb.deldt_min = 0.04 # minimum relative change allowed for model_dt > 0 bbb.initjac = 0 # if=1, calc initial Jac upon reading rdcontdt bbb.numrevjmax = 2 # number of dt reductions before Jac recalculated bbb.numfwdjmax = 1 # number of dt increases before Jac recalculated ###bbb.ismmaxuc = 1 # =1 for intern calc mmaxu; =0,set mmaxu & dont chng bbb.irev = -1 # flag to allow reduced dt advance after cutback bbb.rlx = 0.9 # max. change in variable at each linear iteration bbb.itermx = 7 # max. number of linear iterations allowed bbb.tstor_s = 1e-5 # beginning time for storing solution bbb.tstor_e = 1e-3 # ending time for storing solution bbb.n_stor = 0 # number of linearly spaced storage points bbb.ipt = 1 # index of variable; value printed at step # if ipt not reset from unity, ipt=idxte(nx,iysptrx+1)
2,050
52.973684
76
py
UEDGE
UEDGE-master/pyscripts/ruthere.py
from numpy import * import time import signal import sys ############################################################################# # From Dave Grote: # --- Setup signal handler to capture Control-C # --- To use, first call arminterrupt(). Then at the place where the interrupt # --- is allowed, call ruthere(). This will raise a KeyboardInterrupt if # --- Control-C had been pressed. # --- When a interrupt request is received, all this handler does is set a # --- flag to that effect. Then, a subsequent call to ruthere will check # --- that flag, and if set, raise an exception. This allows a graceful # --- stop with the current time step completed. # --- Set the following two in case ruthere is called before arminterrupt. _defaultcontrolC = signal.getsignal(signal.SIGINT) _controlCrecieved = False savetracebacklimit = 0 def _handlecontrolC(signum, frame): global _controlCrecieved _controlCrecieved = True def ruthere(reset=True): """ Checks if an interrupt was requested (usually control-C). If so, then raise an exception. If reset is True, restore the original interrupt handler so that the calling code does not have to, and so that, if there is an exception, it gets restored (since the calling code is not returned to). """ global _controlCrecieved global _defaultcontrolC global savetracebacklimit if _controlCrecieved: if reset: signal.signal(signal.SIGINT, _defaultcontrolC) _controlCrecieved = False raise KeyboardInterrupt("Interrupt requested") def arminterrupt(): global _controlCrecieved global _defaultcontrolC global savetracebacklimit _controlCrecieved = False _defaultcontrolC = signal.getsignal(signal.SIGINT) try: savetracebacklimit = sys.tracebacklimit except: savetracebacklimit = None signal.signal(signal.SIGINT, _handlecontrolC) sys.tracebacklimit = 0 def disarminterrupt(): global _defaultcontrolC global savetracebacklimit signal.signal(signal.SIGINT, _defaultcontrolC) sys.tracebacklimit = savetracebacklimit #========================================================================= arminterrupt()
2,199
31.835821
82
py
UEDGE
UEDGE-master/pyscripts/__version__.py
__version__ = '8.0.0'
22
10.5
21
py
UEDGE
UEDGE-master/pyscripts/checkver.py
import json pkg = 'uedge' try: import importlib.metadata thisver = importlib.metadata.version(pkg) except: import pkg_resources thisver = pkg_resources.get_distribution(pkg).version try: import urllib.request contents = urllib.request.urlopen('https://pypi.org/pypi/'+pkg+'/json').read() data = json.loads(contents.decode()) thatver = data['info']['version'] except: import urllib contents = urllib.urlopen('https://pypi.org/pypi/'+pkg+'/json').read() data = json.loads(contents) thatver = str(data['info']['version']) print() if thisver > thatver: #print('Uedge version '+thisver+' is newer than available with pip ('+thatver+')') pass elif thisver == thatver: #print('Uedge version '+thisver+' is up-to-date') pass elif thisver < thatver: print('Uedge version '+thisver+', an update is available to '+thatver) else: print('Some error checking pypi version') print()
928
24.108108
85
py
UEDGE
UEDGE-master/pyscripts/uedge.py
try: import IPython from IPython.terminal.prompts import Prompts,Token from IPython.terminal.embed import InteractiveShellEmbed except: pass try: from traitlets.config.loader import Config except: pass import sys,os,__main__ import numpy as np from numpy import array,tanh,exp,arange ArrayType = np.ndarray if sys.hexversion >= 0x03000000: # --- With Python3, the so files of each Fortran package are imported # --- separately. The dlopen flag needs to be set so that cross references # --- among the packages can be satisfied. sys.setdlopenflags(os.RTLD_LAZY | os.RTLD_GLOBAL) from . import uedgeC from .uedgeC import * #from Forthon import * if sys.hexversion >= 0x03000000: from .compy import com from .grdpy import grd from .flxpy import flx from .bbbpy import bbb from .svrpy import svr from .wdfpy import wdf from .apipy import api from .aphpy import aph from .nclpy import ncl else: from wdfpy import wdf from grdpy import grd from flxpy import flx from bbbpy import bbb from svrpy import svr from apipy import api from aphpy import aph from compy import com from nclpy import ncl import time import os.path import __main__ # import all of the neccesary packages def gettypecode(x): return x.dtype.char def oldnonzero(a): return a.nonzero()[0] # Import the uedgeC shared object which contains all of UEDGE try: import PyPDB from PyPDB import PW, PR from PyPDB.pypdb import * except: # print "Unable to import PyPDB or * from PyPDB.pypdb." # print "Will proceed to try to import pypdb in case of old installation." try: from pypdb import * except: # print "pypdb not found." pass # --- The UEDGE modules must be imported in the order below because of # --- linking dependencies. # --- Set default runid to first filename in the command line, stripping off # --- the .py suffix. if sys.argv[0]: if sys.argv[0][-3:] == '.py': h, t = os.path.split(sys.argv[0][:-3]) runid = t del h, t else: h, t = os.path.split(sys.argv[0]) runid = t del h, t # --- Check if the compiler was ifort - if so, set the stacksize unlimited # --- The fcompname is not yet be available yet if Forthon is not up to date try: if fcompname == 'ifort': import resource resource.setrlimit(resource.RLIMIT_STACK, (-1, -1)) except: pass try: class MyPrompt(Prompts): def in_prompt_tokens(self, cli=None): return [(Token.Prompt, 'UEDGE>>> ')] def out_prompt_tokens(self, cli=None): return [(Token.Prompt, 'UEDGE>>> ')] get_ipython except: sys.ps1='UEDGE>>> ' else: ip = get_ipython() ip.prompts = MyPrompt(ip) ############################################################################## ###### Don't put anything below this line!!! ################################ ##############################################################################
3,031
23.650407
78
py
UEDGE
UEDGE-master/pyscripts/uexec.py
import sys try: from importlib import reload,import_module except: from importlib import import_module import builtins def uexec(mname,returns=globals()): if mname in sys.modules: _m = reload(sys.modules[mname]) else: _m = import_module(mname) # is there an __all__? if so respect it if "__all__" in _m.__dict__: names = _m.__dict__["__all__"] else: # otherwise we import all names that don't begin with _ names = [x for x in _m.__dict__ if not x.startswith("_")] # now drag them in for k in names: #print k,getattr(_m,k) returns[k] = getattr(_m,k)
652
20.064516
65
py
UEDGE
UEDGE-master/pyscripts/uedgeplots.py
import numpy as np import matplotlib.pyplot as plt import matplotlib.cm as cm import matplotlib.axes as ax import sys from matplotlib.collections import PolyCollection from matplotlib.colors import LinearSegmentedColormap from matplotlib import rcdefaults from matplotlib import interactive from uedge import * from numpy import sin, cos from scipy import spatial from skimage.util import img_as_ubyte as bytescale # This file defines a function to plot the UEDGE mesh, and # then calls the function to plot the entire mesh. # To use this file in a simple way, give the following ands: # read plotmesh # nf # The function could then be used in a more sophisticated way # to plot portions of the mesh, possibly with customized plot limits # (by resetting acom.ny of r_min, r_max, z_min, and z_max): # call plotmesh(ixmn,ixmx,iymn,iymx) # nf # where ixmn, ixmx, iymn, and iymx are integer variables or # expressions. Always give an "nf" and after reading the file # plotmesh or calling the function plotmesh. # DEFINE THE PLOT FUNCTION -- def plotmesh(ixmin=None, ixmax=None, iymin=None, iymax=None, r_min=None, r_max=None, z_min=None, z_max=None, title=None, block=False, figsize=(4.0, 8.0),xlabel=None,ylabel=None): """ plotmesh(ixmin=<int>,ixmax=<int>,iymin=<int>,iymax=<int> title=<string>,r_min=<val>,r_max=<val>,z_min=<val>,z_max=<val>, block=<True|False>,xlabel=None,ylabel=None,zlabel=None) Plot the uedge grid. where ixmin, ixmax, iymin, and iymax are integer variables or expressions used to plot a portion of the grid. title is used as both the title and the figure name. Block default is True. The plot axis limits may be specified with r_rmin,r_max,z_min,z_max. """ zrefl = com.zm zlim = com.ylim zreflbdry = com.zbdry if str(com.geometry) == str([b'uppersn ']): zrefl = 2.0 * com.zmid - com.zm zlim = 2.0 * com.zmid - com.ylim zreflbdry = 2.0 * com.zmid - com.zbdry if ixmin == None: ixmin = com.nxomit if ixmax == None: ixmax = (com.nxm-1) if iymin == None: iymin = 0 if iymax == None: iymax = (com.ny-1) if r_min == None: r_min = com.rm.min() if r_max == None: r_max = com.rm.max() if z_min == None: z_min = zrefl.min() if z_max == None: z_max = zrefl.max() rcdefaults() if title == None: title = 'Uedge Grid' fig,ax = plt.subplots(figsize=figsize) ax.set_title(title) try: ax.plot(com.xlim, zlim, 'k-', label='Limiter', linewidth=3) ax.plot(com.xlim, zlim, 'y-', label='Limiter', linewidth=1) ax.plot(com.rbdry, zreflbdry, 'b-', label='Last Closed') except: pass for ix in range(ixmax-ixmin+1): for iy in range(iymax-iymin+1): r0 = [com.rm[ix, iy, 1], com.rm[ix, iy, 2], com.rm[ix, iy, 4], com.rm[ix, iy, 3], com.rm[ix, iy, 1]] z0 = [zrefl[ix, iy, 1], zrefl[ix, iy, 2], zrefl[ix, iy, 4], zrefl[ix, iy, 3], zrefl[ix, iy, 1]] ax.plot(r0, z0, 'k-', label='Grid', linewidth=1) if ylabel == None: ax.set_ylabel('Z (m)') else: ax.set_ylabel(ylabel) if xlabel == None: ax.set_xlabel('R (m)') else: ax.set_xlabel(xlabel) ax.set_ylim(z_min, z_max) ax.set_xlim(r_min, r_max) ax.set_aspect('equal') plt.ion() plt.show(block=block) plt.pause(0.001) def plotanymesh(verts, r_min=None, r_max=None, z_min=None, z_max=None, title=None, block=False, figsize=(4.0, 8.0),xlabel=None,ylabel=None): """ plotanymesh(verts, title=<string>,r_min=<val>,r_max=<val>,z_min=<val>,z_max=<val>, block=<True|False>,xlabel=None,ylabel=None) Plot any polynomial NxM grid. verts dimensions are [0:N,0:M,0:nverts,0:2]. Last dim is [:,:,:,0] is R array, [:,:,:,1] is Z array title is used as both the title and the figure name. Block default is True. The plot axis limits may be specified with r_rmin,r_max,z_min,z_max. """ zrefl = com.zm zlim = com.ylim zreflbdry = com.zbdry if str(com.geometry) == str([b'uppersn ']): zrefl = 2.0 * com.zmid - com.zm zlim = 2.0 * com.zmid - com.ylim zreflbdry = 2.0 * com.zmid - com.zbdry if r_min == None: r_min = np.min(verts[:,:,:,0]) if r_max == None: r_max = np.max(verts[:,:,:,0]) if z_min == None: z_min = np.min(verts[:,:,:,1]) if z_max == None: z_max = np.max(verts[:,:,:,1]) rcdefaults() if title == None: title = 'Grid' fig,ax = plt.subplots(figsize=figsize) ax.set_title(title) try: ax.plot(com.xlim, zlim, 'k-', label='Limiter', linewidth=3) ax.plot(com.xlim, zlim, 'y-', label='Limiter', linewidth=1) ax.plot(com.rbdry, zreflbdry, 'b-', label='Last Closed') except: pass s = verts.shape xlen = s[0] ylen = s[1] for ix in range(xlen): for iy in range(ylen): r0 = [verts[ix, iy, 0, 0], verts[ix, iy, 1, 0], verts[ix, iy, 2, 0], verts[ix, iy, 3, 0], verts[ix, iy, 0, 0]] z0 = [verts[ix, iy, 0, 1], verts[ix, iy, 1, 1], verts[ix, iy, 2, 1], verts[ix, iy, 3, 1], verts[ix, iy, 0, 1]] ax.plot(r0, z0, 'k-', label='Grid', linewidth=1) if ylabel == None: ax.set_ylabel('Z (m)') else: ax.set_ylabel(ylabel) if xlabel == None: ax.set_xlabel('R (m)') else: ax.set_xlabel(xlabel) ax.set_ylim(z_min, z_max) ax.set_xlim(r_min, r_max) ax.set_aspect('equal') plt.ion() plt.show(block=block) plt.pause(0.001) def plotmeshval(val, ixmin=None, ixmax=None, iymin=None, iymax=None, r_min=None, r_max=None, z_min=None, z_max=None, title=None, units=None, block=False,xlabel=None,ylabel=None,zlabel=None,figsize=(5.0,8.0)): """ plotmeshval(val,ixmin=<int>,ixmax=<int>,iymin=<int>,iymax=<int> title=<string>,units=<string>,block=<True|False> xlabel=None,ylabel=None,zlabel=None) Display Uedge 2-D quantity using polyfill. where ixmin, ixmax, iymin, and iymax are integer variables or expressions used to plot a portion of the grid. title is used as both the title and the figure name. Units are displayed in the side colorbar. Block default is True. The plot axis limits may be specified with r_rmin,r_max,z_min,z_max. """ zrefl = com.zm zlim = com.ylim zreflbdry = com.zbdry if str(com.geometry) == str([b'uppersn ']): zrefl = 2.0 * com.zmid - com.zm zlim = 2.0 * com.zmid - com.ylim zreflbdry = 2.0 * com.zmid - com.zbdry if ixmin == None: ixmin = com.nxomit if ixmax == None: ixmax = (com.nxm-1) if iymin == None: iymin = 0 if iymax == None: iymax = (com.ny-1) if r_min == None: r_min = com.rm.min() if r_max == None: r_max = com.rm.max() if z_min == None: z_min = zrefl.min() if z_max == None: z_max = zrefl.max() rcdefaults() if title == None: title = 'Uedge' fig, ax = plt.subplots(figsize=figsize) verts = np.array([]) z = np.array([]) for ix in range(ixmax-ixmin+1): for iy in range(iymax-iymin+1): v = [] v.append([com.rm[ix, iy, 1], zrefl[ix, iy, 1]]) v.append([com.rm[ix, iy, 2], zrefl[ix, iy, 2]]) v.append([com.rm[ix, iy, 4], zrefl[ix, iy, 4]]) v.append([com.rm[ix, iy, 3], zrefl[ix, iy, 3]]) verts = np.append(verts, v) z = np.append(z, val[ix, iy]) verts = verts.reshape(len(z), 4, 2) ax.set_title(title) if ylabel == None: ax.set_ylabel('Z (m)') else: ax.set_ylabel(ylabel) if xlabel == None: ax.set_xlabel('R (m)') else: ax.set_xlabel(xlabel) try: ax.plot(com.xlim, zlim, 'k-', label='Limiter', linewidth=3) ax.plot(com.xlim, zlim, 'y-', label='Limiter', linewidth=1) ax.plot(com.rbdry, zreflbdry, 'b-', label='Last Closed') except: pass coll = PolyCollection(verts, array=z, cmap=cm.jet, edgecolors='face') ax.add_collection(coll) ax.autoscale_view() cbar = fig.colorbar(coll, ax=ax,label=zlabel) # if units != None: cbar.ax.set_ylabel(units,rotation=-90,va='bottom') if units != None: cbar.ax.set_ylabel(units, va='bottom') ax.set_ylim(z_min, z_max) ax.set_xlim(r_min, r_max) ax.set_aspect('equal') plt.ion() plt.show(block=block) plt.pause(0.001) def plotanymeshval(verts,z, r_min=None, r_max=None, z_min=None, z_max=None, title=None, units=None, block=False,xlabel=None,ylabel=None,zlabel=None): """ plotanymeshval(verts, val, title=<string>,units=<string>,block=<True|False>, xlabel=None,ylabel=None,zlabel=None) Display 2-D (NxM) quantity, val, using polyfill of NxM polynomial grid verts verts dimensions are [0:N,0:M,0:nverts,0:2]. Last dim is [:,:,:,0] is R array, [:,:,:,1] is Z array title is used as both the title and the figure name. Units are displayed in the side colorbar. Block default is True. The plot axis limits may be specified with r_rmin,r_max,z_min,z_max. """ zrefl = com.zm zlim = com.ylim zreflbdry = com.zbdry if str(com.geometry) == str([b'uppersn ']): zrefl = 2.0 * com.zmid - com.zm zlim = 2.0 * com.zmid - com.ylim zreflbdry = 2.0 * com.zmid - com.zbdry if r_min == None: r_min = com.rm.min() if r_max == None: r_max = com.rm.max() if z_min == None: z_min = zrefl.min() if z_max == None: z_max = zrefl.max() rcdefaults() if title == None: title = 'Uedge' fig, ax = plt.subplots() ax.set_title(title) if ylabel == None: ax.set_ylabel('Z (m)') else: ax.set_ylabel(ylabel) if xlabel == None: ax.set_xlabel('R (m)') else: ax.set_xlabel(xlabel) try: ax.plot(com.xlim, zlim, 'k-', label='Limiter', linewidth=3) ax.plot(com.xlim, zlim, 'y-', label='Limiter', linewidth=1) ax.plot(com.rbdry, zreflbdry, 'b-', label='Last Closed') except: pass coll = PolyCollection(verts, array=z, cmap=cm.jet, edgecolors='face') ax.add_collection(coll) ax.autoscale_view() cbar = fig.colorbar(coll, ax=ax,label=zlabel) # if units != None: cbar.ax.set_ylabel(units,rotation=-90,va='bottom') if units != None: cbar.ax.set_ylabel(units, va='bottom') ax.set_ylim(z_min, z_max) ax.set_xlim(r_min, r_max) ax.set_aspect('equal') plt.ion() plt.show(block=block) plt.pause(0.001) def mkdensityfile(filename, ival, renmin=None, renmax=None, samples=[500, 500, 500], xrange=[-2.4, 2.4], yrange=[-2.4, 2.4], zrange=[0, 3.2], tree=None): """ mkdensityfile(filename, ival,renmin=<float>,renmax=<float>, samples=[<xsamps>,<ysamps>,<zsamps>], xrange=[xmin,xmax],yrange=[ymin,ymax],zrange=[zmin,zmax], tree=<cKDTree object> ) Output Povray include and density field file (df3) for rendering. where: renmin,renmax are the values scaled to 0,255 in the final bytescaling samples is an array of three values giving the volume sampling for the density file (def [500,500,500]) xrange, yrange, zrange are the vessel dimensions of the sampled volume (m) (def xrange[-2.4,2.4], yrange[-2.4,2.4], zrange[0,3.2]) tree is returned and may be reused for another call for efficiency The defaults are set for DIII-D and will sample the full torus at about 1cm r resolution and .6cm in z. """ zrefl = com.zm zlim = com.ylim zreflbdry = com.zbdry if str(com.geometry) == str([b'uppersn ']): zrefl = 2.0 * com.zmid - com.zm zlim = 2.0 * com.zmid - com.ylim zreflbdry = 2.0 * com.zmid - com.zbdry if renmin == None: renmin = np.min(ival) if renmax == None: renmax = np.max(ival) nx, ny, nz = samples dims = np.array([nx, ny, nz], dtype=np.int16) file = open(filename, 'wb') if sys.byteorder == 'little': file.write(dims.byteswap(True)) else: file.write(dims) rrm = com.rm[:, :, 0].ravel() rzm = zrefl[:, :, 0].ravel() if tree == None: tree = spatial.cKDTree(list(zip(rrm, rzm))) treelen = tree.data.shape[0] rpts = np.array([]) zpts = np.array([]) z, x, y = np.mgrid[ zrange[0]:zrange[1]:complex(0, nz), xrange[0]:xrange[1]:complex(0, nx), yrange[0]:yrange[1]:complex(0, ny) ] r = (x*x + y*y)**0.5 pts = list(zip(r.ravel(), z.ravel())) #d,i = tree.query(pts,k=1,distance_upper_bound=0.1) d, i = tree.query(pts, k=1) val = ival val[0, :] = 0 val[-1, :] = 0 val[:, 0] = 0 val[:, -1] = 0 vf = np.append(val.ravel(), [renmin]) #dens = bytescale(np.average(vf[i],axis=1,weights=1./d),cmin=renmin,cmax=renmax) dens = bytescale((vf[i] - renmin)/(renmax - renmin)) file.write(dens) file.close() return tree def profile(rval, zval, title=None, style=None, linewidth=None, xlabel=None, ylabel=None, figsize=(4.0, 8.0), block=False,marker=None): """ profile(xval,yval,title=<None>,style=<None>,linewidth=<None>,xlabel=<None>,ylabel=<None>,block=<True|False>,marker=<none>) title is used as both the title and the figure name. Interactive is turned on so subsequent calls go to the same plot Style encoded color, line, and marker. See matplotlib documention. examples: black solid line - style='k-' red circle marks - style='ro' green x marks and dotted line - style='gx--' """ rcdefaults() interactive(True) if title == None: title = 'Uedge Profile' if style == None: style = 'k-' if linewidth == None: lw = 1 fig,ax = plt.subplots(figsize=figsize) ax.set_title(title) try: ax.plot(rval, zval, style, linewidth=lw,marker=marker) except: pass if ylabel != None: ax.set_ylabel(ylabel) if xlabel != None: ax.set_xlabel(xlabel) plt.ion() plt.show(block=block) plt.pause(0.001)
14,603
33.524823
135
py
UEDGE
UEDGE-master/pyscripts/rundt.py
# Holm10 Nov 5 2019, based on rdcontdt.py # 191121 - Created hdf5-routines to read and save time-dependent data # Writes and reads dictionary with multi-dimensional arrays # containing all restore-parameters. # 230210 - Updated old rundt function to RunData class, modularizing # all functionalities. Added a comprehensive diagnostics suite # plotting fnrm evolution as function of exmains(), plasma # time, and wall-clock time. Still need to test time-slicing # ` procedures # 230327 - Removed old routine, created wrapper function rundt for # object. Renamed Object to UeRun. # 230522 - Fixed bug associated with itroub, improved itroub visualization from matplotlib.pyplot import ion ion() class UeRun(): ''' Class containing information on run ''' def __init__(self, n_stor = False): from time import time from numpy import array from uedge import bbb, com # TODO: Add restore/recover from timeslice # TODO: Add plot timeslice directly # NOTE: No -> Utilize direct I/O from file instead self.tstart = time() self.numvar = bbb.numvar self.nx = com.nx self.ny = com.ny self.ixpt1 = com.ixpt1[0] self.ixpt2 = com.ixpt2[0] self.iysptrx = com.iysptrx self.equationkey = array([b'te', b'ti', b'phi', b'up', b'ni', b'ng', b'tg']) self.classvars = ['slice_ni', 'slice_ng', 'slice_up', 'slice_te', 'slice_ti', 'slice_tg', 'slice_phi', 'slice_dttot', 'time', 'fnorm', 'nfe', 'dt_tot', 'dtreal', 'ii1', 'ii2', 'ii1fail', 'ii2fail', 'dtrealfail', 'itrouble', 'troubleeq', 'troubleindex', 'ylfail', 'isteon', 'istion', 'isupon', 'isphion', 'isupgon', 'isngon', 'istgon', 'ishymol', 'nisp', 'ngsp', 'nhsp', 'nhgsp', 'nzsp', 'b0', 'ncore', 'pcoree', 'pcorei', 'internaleq', 'internalspecies', 'yldotsfscalfail'] # Intiialize all variables to empty lists in class for var in self.classvars: self.__setattr__(var, []) def itroub(self): ''' Function that displays information on the problematic equation ''' from numpy import mod, argmax, where, array, argmin from uedge import bbb from copy import deepcopy self.equations = [bbb.idxte, bbb.idxti, bbb.idxphi, bbb.idxu, bbb.idxn, bbb.idxg, bbb.idxtg] equationsdescription = [ 'Electron energy', 'Ion energy', 'Potential', 'Ion momentum', 'Ion density', 'Gas density', 'Gas temperature'] # Find the fortran index of the troublemaking equation self.neq = bbb.neq self.itrouble.append(deepcopy(argmax(abs(bbb.yldot*\ bbb.sfscal)[:bbb.neq])+1)) print("** Fortran index of trouble making equation is:\n{}".format(\ self.itrouble[-1])) # Print equation information print("** Number of equations solved per cell:\n numvar = {}\n"\ .format(self.numvar)) self.troubleeq.append(mod(self.itrouble[-1]-1, bbb.numvar)+1) species = '' self.internaleq.append([abs(x - self.itrouble[-1]).min() for x in \ self.equations].index(0)) if self.equations[self.internaleq[-1]].ndim == 3: self.internalspecies.append( where(\ self.equations[self.internaleq[-1]] == self.itrouble[-1])\ [-1][0] + 1) species = ' of species {}'.format(self.internalspecies[-1]) else: self.internalspecies.append(0) print('** Troublemaker equation is:\n{} equation{}: iv_t={}\n'\ .format(equationsdescription[self.internaleq[-1]], species, self.troubleeq[-1])) # Display additional information about troublemaker cell self.troubleindex.append(deepcopy(bbb.igyl[self.itrouble[-1]-1,])) self.dtrealfail.append(deepcopy(bbb.dtreal)) self.ylfail.append(deepcopy(bbb.yl[self.itrouble[-1]-1])) self.yldotsfscalfail.append(deepcopy((bbb.yldot*bbb.sfscal)\ [self.itrouble[-1]-1])) print('** Troublemaker cell (ix,iy) is:\n' + \ '{}\n'.format(self.troubleindex[-1])) print('** Timestep for troublemaker equation:\n' + \ '{:.4e}\n'.format(self.dtrealfail[-1])) print('** yl for troublemaker equation:\n' + \ '{:.4e}\n'.format(self.ylfail[-1])) print('** yl*sfscal for troublemaker equation:\n' + \ '{:.4e}\n'.format(self.yldotsfscalfail[-1])) def savesuccess(self, ii1, ii2, savedir, savename, fnrm=None): from time import time from uedge import bbb from copy import deepcopy self.time.append(time()) if fnrm is None: bbb.pandf1 (-1, -1, 0, bbb.neq, 1., bbb.yl, bbb.yldot) self.fnorm.append(deepcopy((sum((bbb.yldot[:bbb.neq]*\ bbb.sfscal[:bbb.neq])**2))**0.5)) else: self.fnorm.append(fnrm) self.nfe.append(deepcopy(bbb.nfe)) self.dt_tot.append(deepcopy(bbb.dt_tot)) self.dtreal.append(deepcopy(bbb.dtreal)) self.ii1.append(ii1) self.ii2.append(ii2) self.neq = bbb.neq try: self.save('{}_UeCase.hdf5'.format(savefname.split('.')[0])) except: pass self.save_intermediate(savedir, savename) def store_timeslice(self): from copy import deepcopy from uedge import bbb self.slice_ni.append(deepcopy(bbb.ni)) self.slice_ng.append(deepcopy(bbb.ng)) self.slice_up.append(deepcopy(bbb.up)) self.slice_te.append(deepcopy(bbb.te)) self.slice_ti.append(deepcopy(bbb.ti)) self.slice_tg.append(deepcopy(bbb.tg)) self.slice_phi.append(deepcopy(bbb.phi)) self.slice_dttot.append(deepcopy(bbb.dt_tot)) def save_intermediate(self, savedir, savename): from uedge.hdf5 import hdf5_save from uedge import bbb, com from h5py import File for var in [ 'isteon', 'istion', 'isupon', 'isphion', 'isupgon', 'isngon', 'istgon', 'ishymol']: self.__setattr__(var, bbb.__getattribute__(var)) for var in [ 'nisp', 'ngsp', 'nhsp', 'nhgsp', 'nzsp']: self.__setattr__(var, com.__getattribute__(var)) try: hdf5_save('{}/{}_last_ii2.hdf5'.format(savedir,savename)) except: print('Folder {} not found, saving output to cwd...'\ .format(savedir)) hdf5_save('{}_last_ii2.hdf5'.format(savename)) # Try to store ready-made Case-file, if possible try: self.save('{}/{}_last_ii2_Case.hdf5'.format(savedir,savename)) except: pass try: file = File('{}/{}_last_ii2.hdf5'.format(savedir, savename), 'r+') except: file = File('{}_last_ii2.hdf5'.format(savename), 'r+') file.require_group('convergence') group = file['convergence'] group.create_dataset('t_start', data=self.tstart) group.create_dataset('numvar', data=self.numvar) group.create_dataset('neq', data=self.neq) group.create_dataset('nx', data=self.nx) group.create_dataset('ny', data=self.ny) group.create_dataset('ixpt1', data=self.ixpt1) group.create_dataset('ixpt2', data=self.ixpt2) group.create_dataset('iysptrx', data=self.iysptrx) group.create_dataset('equationkey', data=self.equationkey) group.create_dataset('itermx', data=self.itermx) group.create_dataset('incpset', data=self.incpset) group.create_dataset('ii1max', data=self.ii1max) group.create_dataset('ii2max', data=self.ii1max) group.create_dataset('numrevjmax', data=self.numrevjmax) group.create_dataset('numfwdjmax', data=self.numfwdjmax) group.create_dataset('numtotjmax', data=self.numtotjmax) group.create_dataset('rdtphidtr', data=self.rdtphidtr) group.create_dataset('deldt_min', data=self.deldt_min) group.create_dataset('rlx', data=self.rlx) for var in self.classvars: group.create_dataset(var, data=self.__getattribute__(var)) file.close() def convergenceanalysis(savefname, savedir='../solutions', fig=None, xaxis = 'exmain', logx = False, color='k', label=None, ylim = (None, None)): from h5py import File from matplotlib.pyplot import subplots from datetime import timedelta from matplotlib.ticker import FuncFormatter from numpy import cumsum, ones if fig is None: f, ax = subplots(1, 3, figsize=(15, 5)) else: ax = fig.get_axes() if len(ax) < 3: print('Three subplots required for plots! Aborting...') return f = fig try: file = File('{}/{}'.format(savedir, savefname), 'r') except: print('File {}/{} not found. Aborting!'.format(savedir, savefname)) return data = file['convergence'] try: data = file['convergence'] except: print('Convergence data not found in {}/{}. Aborting!'.format(\ savedir, savefname)) return if xaxis == 'exmain': xlabel = 'Exmain calls' xones = ones(data['ii2'][()].shape) x = cumsum(xones) elif xaxis == 'nfe': xlabel = 'nfe internal iterations' x = cumsum(data['nfe'][()][:, 0, 0]) elif xaxis == 'time': xlabel = 'Total wall-clock time [HH:MM]' x = [timedelta(t - data['t_start'][()]) for t in data['time'][()]] x = data['time'][()] - data['t_start'][()] if logx is True: ax[0].loglog(x, data['fnorm'][()], '-', color=color, label=label) ax[1].loglog(data['dt_tot'][()], data['fnorm'][()], '-', color=color, label=label) ax[2].loglog(x, data['dtreal'][()], '-', color=color, label=label) else: ax[0].semilogy(x, data['fnorm'][()], '-', color=color, label=label) ax[1].semilogy(data['dt_tot'][()], data['fnorm'][()], '-', color=color, label=label) ax[2].semilogy(x, data['dtreal'][()], '-', color=color, label=label) ax[0].set_xlabel(xlabel) ax[1].set_xlabel('Accumulated plasma simualtion time [s]') ax[2].set_xlabel(xlabel) ax[1].set_title('Total exmain evaluations: {}'.format\ (len(data['dtreal'][()]))) ax[0].set_ylabel('Initial fnorm') ax[1].set_ylabel('Initial fnorm') ax[2].set_ylabel('Time-step (dtreal) [s]') ax[0].set_ylim(ylim) ax[1].set_ylim(ylim) if xaxis == 'time': ax[0].xaxis.set_major_formatter(FuncFormatter(lambda t, pos :\ str(timedelta(seconds=t))[:-3])) ax[2].xaxis.set_major_formatter(FuncFormatter(lambda t, pos :\ str(timedelta(seconds=t))[:-3])) if label is not None: ax[0].legend() return f def failureanalysis(savefname, savedir='../solutions', equation=None): from h5py import File from matplotlib.pyplot import subplots from numpy import histogram, zeros from matplotlib.collections import PolyCollection f, ax = subplots(2,1, figsize=(10, 7)) try: file = File('{}/{}'.format(savedir, savefname), 'r') except: print('File {}/{} not found. Aborting!'.format(savedir, savefname)) try: data = file['convergence'] except: print('Convergence data not found in {}/{}. Aborting!'.format(\ savedir, savefname)) return if equation is not None: iequation = [x.decode('UTF-8') for x in data['equationkey']]\ .index(equation) # Bin the equation errors nspecies = 1/(data['nisp'][()] + 1) nbins = 7*data['nisp'][()] counts, bins = histogram((data['internaleq'][()]+\ data['internalspecies']*nspecies)-0.5, bins=nbins, range=(-0.5,6.5)) h, e = histogram(data['internaleq'][()] - 0.5, bins=7, range=(-0.5,6.5)) ax[0].bar([x for x in range(7)], h, width=1, edgecolor='k', color=(0, 87/255, 183/255)) ax[0].hist(bins[3*data['nisp'][()]:-1], bins[3*data['nisp'][()]:], weights=counts[3*data['nisp'][()]:], edgecolor='k', color=(255/255, 215/255, 0)) ax[0].set_xticks(range(7)) ax[0].set_xticklabels([x.decode('UTF-8') for x in \ data['equationkey'][()]]) ax[0].grid(linestyle=':', linewidth=0.5, axis='y') ax[0].set_xlim((-0.5,6.5)) ax[0].set_ylabel('Counts') for i in range(7): ax[0].axvline(i-0.5, linewidth=1, color='k') # Visualize error locations nx = data['nx'][()] ny = data['ny'][()] ixpt1 = data['ixpt1'][()] ixpt2 = data['ixpt2'][()] iysptrx = data['iysptrx'][()] frequency = zeros((nx+2, ny+2)) cells = [] for i in range(nx+2): for j in range(ny+2): cells.append([[i-.5, j-.5], [i+.5, j-.5], [i+.5, j+.5], [i-.5, j+.5]]) polys = PolyCollection(cells, edgecolors='k', linewidth=0.5, linestyle=':') for i in range(len(data['itrouble'])): coord = data['troubleindex'][()][i] if equation is None: frequency[coord[0], coord[1]] += 1 elif iequation == data['internaleq'][()][i]: frequency[coord[0], coord[1]] += 1 polys.set_cmap('binary') polys.set_array(frequency.reshape(((nx+2)*(ny+2),))) cbar = f.colorbar(polys, ax=ax[1]) cbar.ax.set_ylabel('N trouble'+' for {}'.format(equation)*\ (equation is not None), va='bottom', labelpad=20) ax[1].plot([.5, nx+.5, nx+.5, .5, .5], [.5, .5, ny+.5, ny+.5, .5], 'k-', linewidth=1) ax[1].set_xlabel('Poloidal index') ax[1].set_ylabel('Radial index') ax[1].add_collection(polys) ax[1].plot([.5, nx+.5],[iysptrx+.5, iysptrx+.5], 'k-', linewidth=1) ax[1].plot([ixpt1+.5, ixpt1+.5], [.5, iysptrx+.5], 'k-', linewidth=1) ax[1].plot([ixpt2+.5, ixpt2+.5], [.5, iysptrx+.5], 'k-', linewidth=1) file.close() return f def converge(self, dtreal=2e-9, ii1max=5000, ii2max=5, itermx=7, ftol=1e-5, dt_kill=1e-14, t_stop=100, dt_max=100, ftol_min = 1e-9, incpset=7, n_stor=0, storedist='lin', numrevjmax=2, numfwdjmax=1, numtotjmax=0, tstor=(1e-3, 4e-2), ismfnkauto=True, dtmfnk3=5e-4, mult_dt=3.4, reset=True, initjac=False, rdtphidtr=1e20, deldt_min=0.04, rlx=0.9, tsnapshot=None, savedir='../solutions', ii2increase=0): ''' Converges the case by increasing dt dtreal : float [1e-9] Original time-step size ii1max : int [500] Outer loop iterations, i.e. time-step changes ii2max : int [5] Inner loop iterations, i.e. time-steps per time-step change dt_kill : float [1e-14] Time-step limit for aborting simulation itermx : int [7] Maximum iterations per time-step used internally in routine ftol : float [1e-5] Internal fnrm tolerance for time-steps incpset : int [7] savedir : str ['../solutions'] numtotjmax : int [None] ii2increase : float [1.5] ftol_min : float [1e-9] Value of fnrm where time-advance will stop t_stop : float [100.] Maximum total accumulated plasma-time before stopping if fnorm has not decreased below ftol_min dt_max : float [100.] Maximum allowable time-step size numrevjmax : int [2] Number of time-step reducitons before Jacobian is recomputed numfwdjmax : int [2] Number of time-step increases before Jacobian is recomputed n_stor : int [0] Number of time-slices to be stored in interval tstor tstor : tuple of floats [(1e-3, 4e-2)] Time-interval in which time-slices are stored (lower, upper) storedist : str ['lin'] Distribution of time-slices in tstor. Options are 'lin' and 'log' for linear and logarithmic distributions, respectively reset : bool [True] Switch whether to reset the total time etc of the case initjac : bool [False] Switch to re-evaluate Jacobian on first convegnec time-step or not ismfnkauto : bool [True] If True, sets mfnksol=3 for time-steps smaller that dtmfnk3, mfnksol=-3 for larger time-step sizes dtmfnk3 : float [5e-4] Time-step size for which ismfnkauto controls mfnksol if ismfnkauto is True mult_dt : float [3.4] Time-step size increase factor after successful inner loop rdtphidtr : float [1e20] Ratio of potential-equation time-step to plasma equaiton time-step size: dtphi/dtreal deldt_min : float [0.04] Minimum relative change allowed for model_dt>0 rlx : float [0.9] Maximum change in variable at each internal linear iteration tsnapshot : list [None] If None, uses linear/logarithmic interpolation according to storedist in the interval tstor. Snapshot times can be defined in a list and supplied. Then, the tnsaphsot list defines the time-slices ''' from numpy import linspace, logspace, log10, append from copy import deepcopy from uedge import bbb from os.path import exists # Check if requested save-directory exists: if not, write to cwd if not exists(savedir): print('Requested save-path {} not found, writing to cwd!'.format(\ savedir)) savedir = '.' self.orig = {} self.orig['itermx'] = deepcopy(bbb.itermx) self.orig['dtreal'] = deepcopy(bbb.dtreal) self.orig['icntnunk'] = deepcopy(bbb.icntnunk) self.orig['ftol'] = deepcopy(bbb.ftol) self.orig['mfnksol'] = deepcopy(bbb.mfnksol) self.orig['rlx'] = deepcopy(bbb.rlx) self.orig['deldt'] = deepcopy(bbb.deldt) self.orig['isdtsfscal'] = deepcopy(bbb.isdtsfscal) self.orig['incpset'] = deepcopy(bbb.incpset) if numtotjmax == 0: numtotjmax = numrevjmax + numfwdjmax self.itermx = itermx self.incpset = incpset self.ii1max = ii1max self.ii2max = ii2max self.numrevjmax = numrevjmax self.numfwdjmax = numfwdjmax self.numtotjmax = numtotjmax self.rdtphidtr = rdtphidtr self.deldt_min = deldt_min self.rlx = rlx # TODO: Add variable to control reduciton factor? # TODO: Should dtreal = min(x, t_stop) actually be t_stop or dt_max? def restorevalues(self): ''' Restores the original UEDGE values ''' for key, value in self.orig.items(): bbb.__setattr__(key, value) def message(string, separator='-', pad='', seppad = '', nseparator=1): ''' Prints formatted message to stdout ''' # TODO: add formatting for len>75 strings message = pad.strip() + ' ' + string.strip() + ' ' + pad.strip() for i in range(nseparator): print(seppad + separator*(len(message)-2*len(seppad)) + seppad) print(message) print(seppad + separator*(len(message)-2*len(seppad)) + seppad) def scale_timestep(scaling): ''' Increases/decreases time-step ''' bbb.dtreal *= scaling def exmain_isaborted(self): ''' Checks if abort is requested ''' from uedge import bbb bbb.exmain() # Abort flag set, abort case if bbb.exmain_aborted == 1: # Reset flag bbb.exmain_aborted == 0 # Restore parameters modified by script restorevalues(self) return True def issuccess(self, t_stop, ftol_min): ''' Checks if case is converged ''' from datetime import timedelta from time import time if (bbb.iterm == 1): bbb.ylodt = bbb.yl bbb.dt_tot += bbb.dtreal bbb.pandf1 (-1, -1, 0, bbb.neq, 1., bbb.yl, bbb.yldot) self.fnrm_old = sum((bbb.yldot[:bbb.neq-1]*\ bbb.sfscal[:bbb.neq-1])**2)**0.5 self.savesuccess(ii1, ii2, savedir, bbb.label[0].strip(\ ).decode('UTF-8'), self.fnrm_old) if (bbb.dt_tot>=t_stop or self.fnrm_old<ftol_min): print('') message('SUCCESS: ' + 'fnrm < bbb.ftol'\ *(self.fnrm_old<ftol_min) + \ 'dt_tot >= t_stop'*(bbb.dt_tot >= t_stop), pad='**', separator='*') print('Total runtime: {}'.format(timedelta( seconds=round(time()-self.tstart)))) restorevalues(self) return True def isfail(dt_kill): ''' Checks whether to abandon case ''' if (bbb.dtreal < dt_kill): message('FAILURE: time-step < dt_kill', pad='**', separator='*') restorevalues(self) return True def setmfnksol(ismfnkauto, dtmfnk3): ''' Sets mfnksol according to setup ''' if ismfnkauto is True: bbb.mfnksol = 3*(-1)**(bbb.dtreal > dtmfnk3) def calc_fnrm(): ''' Calculates the initial fnrm ''' from uedge import bbb bbb.pandf1 (-1, -1, 0, bbb.neq, 1., bbb.yl, bbb.yldot) return sum((bbb.yldot[:bbb.neq-1]*bbb.sfscal[:bbb.neq-1])**2)**0.5 ''' TIME-SLICING SETUP ''' if tsnapshot is None: if storedist == 'lin': # Linearly spaced time slices for writing dt_stor = linspace(tstor[0], tstor[1], n_stor) elif storedist == 'log': # Logarithmically spaced time-slices dt_stor = logspace(log10(tstor[0]), log10(tstor[1]), n_stor) else: dt_stor = tsnapshot # Add end-point to avoid tripping on empty arrays dt_stor = append(dt_stor, 1e20) if reset is True: bbb.dt_tot = 0 ''' TIME-STEP INITIALIZATION ''' bbb.rlx = rlx bbb.dtreal = dtreal bbb.ftol = ftol if (bbb.iterm == 1) and (bbb.ijactot > 0): message('Initial successful time-step exists', separator='') else: message('Need to take initial step with Jacobian; ' + \ 'trying to do here', seppad='*') # Ensure time-step is taken bbb.icntnunk = 0 # Take timestep and see if abort requested if exmain_isaborted(self): return # Increase time # Verify time-step was successful if (bbb.iterm != 1): restorevalues(self) message('Error: converge an initial time-step first; then ' + \ 're-execute command', seppad='*') return bbb.incpset = incpset bbb.itermx = itermx deldt_0 = deepcopy(bbb.deldt) isdtsf_sav = deepcopy(bbb.isdtsfscal) # TODO: Replace with some more useful information? # if (bbb.ipt==1 and bbb.isteon==1): # set ipt to te(nx,iysptrx+1) # #if no user value # ipt = bbb.idxte[nx-1,com.iysptrx] #note: ipt is local, # # bbb.ipt global bbb.dtphi = rdtphidtr*bbb.dtreal svrpkg=bbb.svrpkg.tostring().strip() bbb.ylodt = bbb.yl self.fnrm_old = calc_fnrm() if initjac is True: self.fnrm_old = 1e20 else: bbb.newgeo=0 # Intialize counters irev = -1 numfwd = 0 numrev = 0 numrfcum = 0 # Compensate for first time-step before entering loop scale_timestep(1/(3*(irev == 0) + mult_dt*(irev != 0))) ''' OUTER LOOP - MODIFY TIME-STEP SIZE''' # TODO: Add logic to always go back to last successful ii2 to # precondition the Jacobian, to avoid downwards cascades? # NOTE: experomental functionality successivesuccesses = 0 for ii1 in range(ii1max): setmfnksol(ismfnkauto, dtmfnk3) # adjust the time-step # dtmult=3 only used after a dt reduc. success. completes loop ii2 # for fixed dt either increase or decrease dtreal; depends # on mult_dt scale_timestep(3*(irev == 0) + mult_dt*(irev != 0)) bbb.dtreal = min([bbb.dtreal, dt_max]) bbb.dtphi = rdtphidtr*bbb.dtreal bbb.deldt = min([bbb.deldt, deldt_0, deldt_min]) message('Number of time-step changes = ''{} New time-step: {:.2E}\n'\ .format((ii1+1), bbb.dtreal), pad='***', nseparator=1) # Enter for every loop except first, unless intijac == True if ii1 > -int(initjac): # Main time-stepping switch: controls increase/decrease in # dtreal and Jacobian preconditioning if (irev == 1): # decrease in bbb.dtreal if (numrev < numrevjmax and \ numrfcum < numtotjmax): #dont recom bbb.jac bbb.icntnunk = 1 numrfcum += 1 else: # force bbb.jac calc, reset numrev bbb.icntnunk = 0 numrev = -1 # yields api.zero in next statement numrfcum = 0 numrev += 1 numfwd = 0 else: # increase in bbb.dtreal if (numfwd < numfwdjmax and \ numrfcum < numtotjmax): #dont recomp bbb.jac bbb.icntnunk = 1 numrfcum += 1 else: bbb.icntnunk = 0 # recompute jacobian for increase dt numfwd = -1 numrfcum = 0 numfwd += 1 numrev = 0 #bbb.restart counter for dt reversals bbb.isdtsfscal = isdtsf_sav # Dynamically decrease ftol as the initial ftol decreases bbb.ftol = max(min(ftol, 0.01*self.fnrm_old),ftol_min) # Take timestep and see if abort requested if exmain_isaborted(self): return if issuccess(self, t_stop, ftol_min): return bbb.icntnunk = 2 bbb.isdtsfscal = 0 # NOTE: experomental functionality bbb.ii2max = ii2max + round(ii2increase*successivesuccesses) # Take ii2max time-steps at current time-step size while # time-steps converge: if not, drop through for ii2 in range(bbb.ii2max): if (bbb.iterm == 1): bbb.ftol = max(min(ftol, 0.01*self.fnrm_old),ftol_min) # Take timestep and see if abort requested message("Inner iteration #{}".format(ii2+1), nseparator=0, separator='') if exmain_isaborted(self): return if issuccess(self, t_stop, ftol_min): return message("Total time = {:.4E}; Timestep = {:.4E}".format(\ bbb.dt_tot-bbb.dtreal,bbb.dtreal), nseparator=0, separator='') # TODO: replace with more useful information # print("variable index ipt = ",ipt, " bbb.yl[ipt] = ", # bbb.yl[ipt]) # Store variable if threshold has been passed if (bbb.dt_tot >= dt_stor[0]): # Remove storing time-points smaller than current # simulation time while bbb.dt_tot >= dt_stor[0]: dt_stor = dt_stor[1:] self.store_timeslice() irev -= 1 # Output and store troublemaker info # NOTE: experomental functionality successivesuccesses += 1 if (bbb.iterm != 1): # NOTE: experomental functionality successivesuccesses = 0 self.itroub() ''' ISFAIL ''' if isfail(dt_kill): self.save_intermediate(savedir, bbb.label[0].strip()\ .decode('UTF-8')) break irev = 1 message('Converg. fails for bbb.dtreal; reduce time-step by '+\ '3, try again', pad = '***', nseparator=0) scale_timestep(1/(3*mult_dt)) bbb.dtphi = rdtphidtr*bbb.dtreal bbb.deldt *= 1/(3*mult_dt) setmfnksol(ismfnkauto, dtmfnk3) # bbb.iterm = 1 # bbb.iterm = -1 # Ensure subsequent repetitions work as intended def rundt(**kwargs): runcase=UeRun() runcase.converge(**kwargs)
30,206
40.83795
81
py
UEDGE
UEDGE-master/pyscripts/hdf5.py
import numpy as np import h5py import uedge try: import __version__ as pyv pyver = pyv.__version__ except: pyver = uedge.__version__ from .uedge import bbb from .uedge import com from .uedge_lists import * import time from Forthon import packageobject def hdf5_restore(file): """ Read a hdf5 file previously written from pyUedge. This reads the file recursively and will attempt to restore all datasets. This will restore a file saved by either hdf5_save or hdf5_dump. """ try: hf = h5py.File(file, 'r') except: print("Couldn't open hdf5 file ", file) raise try: dummy = hf['bbb'] # force an exception if the group not there hfb = hf.get('bbb') try: for var in ['tes', 'tis', 'ups', 'nis', 'phis', 'ngs', 'tgs']: packageobject('bbb').__setattr__(var, hf['bbb'][var][()]) except: raise except: print("Old style hdf5 file") try: bbb.ngs[...] = np.array(hf.get('ngs@bbb')) except ValueError as error: print("Couldn't read ngs from ", file) print(error) except: print("Couldn't read ngs from ", file) try: bbb.nis[...] = np.array(hf.get('nis@bbb')) except ValueError as error: print("Couldn't read nis from ", file) print(error) except: print("Couldn't read nis from ", file) try: bbb.phis[...] = np.array(hf.get('phis@bbb')) except ValueError as error: print("Couldn't read phis from ", file) print(error) except: print("Couldn't read phis from ", file) try: bbb.tes[...] = np.array(hf.get('tes@bbb')) except ValueError as error: print("Couldn't read tes from ", file) print(error) except: print("Couldn't read tes from ", file) try: bbb.tis[...] = np.array(hf.get('tis@bbb')) except ValueError as error: print("Couldn't read tis from ", file) print(error) except: print("Couldn't read tis from ", file) try: bbb.ups[...] = np.array(hf.get('ups@bbb')) except ValueError as error: print("Couldn't read ups from ", file) print(error) except: print("Couldn't read ups from ", file) try: bbb.tgs[...] = np.array(hf.get('tgs@bbb')) except ValueError as error: print("Couldn't read tgs from ", file) print(error) except: print("Couldn't read tgs from ", file) try: bbb.tipers[...] = np.array(hf.get('tipers@bbb')) except ValueError as error: print("Couldn't read tipers from ", file) print(error) except: print("Couldn't read tipers from ", file) hf.close() return True def hdf5_save(file, varlist=['bbb.ngs', 'bbb.ng', 'bbb.ni', 'bbb.nis', 'bbb.phi', 'bbb.phis', 'bbb.te', 'bbb.tes', 'bbb.ti', 'bbb.tis', 'bbb.up', 'bbb.ups', 'bbb.tg', 'bbb.tgs', 'bbb.ev', 'bbb.prad', 'bbb.pradhyd', 'bbb.tipers','com.nx', 'com.ny', 'com.rm', 'com.zm'], addvarlist=[]): """ Save HDF5 output for restarting and plotting. varlist=[] a list of variables to save specified as strings. package prefix required. Default list is usual variable list. Example use: varlist=['bbb.ni','bbb.te'] addvarlist=[] a list of variables to save in addition to the ones in varlist. Syntax is the same as varlist parameter. Envisioned use is to add output in addition to the default list in varlist. """ grps = {} vars = {} try: hf = h5py.File(file, 'w') hfb = hf.create_group('bbb') grps['bbb'] = {'h5': hfb} grps['bbb']['vars'] = ['uedge_ver'] grps['bbb'] hfb.attrs['time'] = time.time() hfb.attrs['ctime'] = time.ctime() hfb.attrs['code'] = 'UEDGE' hfb.attrs['ver'] = bbb.uedge_ver try: hfb.attrs['pyver'] = pyver grps['bbb']['vars'].append('pyver') except: print("couldn\'t write pyver to header") except ValueError as error: print("HDF5 file open failed to ", file) print(error) raise except: print("HDF5 file open failed to ", file) raise for lvt in varlist: try: vt = lvt.split('.') if not vt[0] in grps.keys(): hfb = hf.create_group(vt[0]) grps[vt[0]] = {'h5': hfb} grps[vt[0]]['vars'] = [] else: hfb = grps[vt[0]]['h5'] pck = packagename2object(vt[0]) po = pck.getpyobject(vt[1]) if vt[1] in grps[vt[0]]['vars']: print(vt[1], " already written, skipping") else: grps[vt[0]]['vars'].append(vt[1]) d = hfb.create_dataset(vt[1], data=po) d.attrs['units'] = pck.getvarunit(vt[1]) d.attrs['comment'] = pck.getvardoc(vt[1]) except ValueError as error: print("HDF5 write failed to ", file, ' var ', vt[1]) print(error) except: print("HDF5 write failed to ", file, ' var ', vt[1]) for lvt in addvarlist: try: vt = lvt.split('.') if not vt[0] in grps.keys(): hfb = hf.create_group(vt[0]) grps[vt[0]] = {'h5': hfb} grps[vt[0]]['vars'] = [] else: hfb = grps[vt[0]]['h5'] pck = packagename2object(vt[0]) po = pck.getpyobject(vt[1]) if vt[1] in grps[vt[0]]['vars']: print(vt[1], " already written, skipping") else: grps[vt[0]]['vars'].append(vt[1]) d = hfb.create_dataset(vt[1], data=po) d.attrs['units'] = pck.getvarunit(vt[1]) d.attrs['comment'] = pck.getvardoc(vt[1]) except ValueError as error: print("HDF5 write failed to ", file, ' var ', vt[1]) print(error) except: print("HDF5 write failed to ", file, ' var ', vt[1]) hf.close() return True def hdf5_dump(file, packages=list_packages(objects=1), vars=None, globals=None): """ Dump all variables from a list of package objects into a file. Default packages are output of uedge.uedge_lists.list_packages() vars=[varlist] dump limited to intersection of varlist and packages """ try: hf = h5py.File(file, 'w') except ValueError as error: print("Couldn't open hdf5 file ", file) print(error) raise except: print("Couldn't open hdf5 file ", file) raise for p in packages: hfg = hf.create_group(p.name()) hfg.attrs['time'] = time.time() hfg.attrs['ctime'] = time.ctime() hfg.attrs['code'] = 'UEDGE' hfg.attrs['ver'] = bbb.uedge_ver try: hfg.attrs['pyver'] = pyver except: pass for v in list_package_variables(p, vars=vars): if p.allocated(v): try: d = hfg.create_dataset(v, data=p.getpyobject(v)) d.attrs['units'] = p.getvarunit(v) d.attrs['comment'] = p.getvardoc(v) except ValueError as error: print("Couldn't write out: "+p.name()+'.'+v) print(error) except: print("Couldn't write out: "+p.name()+'.'+v) else: print(p.name()+'.'+v+" is not allocated") if globals != None: hfg = hf.create_group('globals') hfg.attrs['time'] = time.time() hfg.attrs['ctime'] = time.ctime() hfg.attrs['code'] = 'UEDGE' hfg.attrs['ver'] = bbb.uedge_ver try: hfg.attrs['pyver'] = pyver except: pass for v in list(set(globals.keys()) & set(vars)): try: d = hfg.create_dataset(v, data=globals[v]) d.attrs['units'] = 'none' d.attrs['comment'] = 'Global Variable' except ValueError as error: print("Couldn't write out: "+p.name()+'.'+v) print(error) except: print("Couldn't write out: "+p.name()+'.'+v) hf.close() return True def h5py_dataset_iterator(g, prefix=''): for key in g.keys(): item = g[key] path = '{}/{}'.format(prefix, key) if isinstance(item, h5py.Dataset): # test for dataset yield (path, item) elif isinstance(item, h5py.Group): # test for group (go down) # following yield is not python 2.7 compatible # yield from h5py_dataset_iterator(item, path) for (path, item) in h5py_dataset_iterator(item, path): yield (path, item) def hdf5_restore_dump(file, scope=None, hdffile=None,quiet=False): """ Restore all variables from a previously saved HDF5 file. This is called by hdf5_restore and the recommended way to restore. """ if hdffile == None: try: hf = h5py.File(file, 'r') except: print("Couldn't open hdf5 file ", file) raise else: hf = hdffile try: try: g = hf['bbb'] if not quiet: prfileattrs = False print('File attributes:') print(' written on: ', g.attrs['ctime']) print(' by code: ', g.attrs['code']) print(' version: ', np.char.strip(g.attrs['ver'])) print(' physics tag: ', np.char.strip(g.attrs['ver'])) print(' python version: ', np.char.strip(g.attrs['pyver'])) except: if not quiet: print('No file attributes, trying to restore') for (path, dset) in h5py_dataset_iterator(hf): vt = path.split('/') if vt[1] == 'globals': if scope != None: if dset.size > 1: scope[vt[2]] = np.array(dset[()]) else: scope[vt[2]] = dset[()] else: try: pck = packagename2object(vt[1]) po = pck.getpyobject(vt[2]) if dset.size > 1: po[...] = np.array(dset[()]) else: setattr(pck, vt[2], dset[()]) except ValueError as error: print("Couldn't read tes from ", file) print(error) except: print('Couldn\'t read dataset ', path) raise except: print("Couldn't read hdf5 file ", file) raise if hdffile == None: hf.close() return True
11,545
33.363095
90
py
UEDGE
UEDGE-master/pyscripts/filelists.py
# Methods to create a list of files in a directory, and a sub-list containing sspecified suffixes. import os import string def filelist(path): return os.listdir(path) def filesublist(path,suffix): sublist = [] biglist = filelist(path) for filename in biglist: filesplit = filename.split(".") if filesplit[-1] == suffix: sublist.append(filename) return sublist
404
21.5
98
py
UEDGE
UEDGE-master/pyscripts/convert.py
# A Python translator that will convert files following a set # of file-suffix-specific substitution rules # Written by Ron Cohen, February 2007. Last revision March 15, 2007. # import utilities to create filtered list of files from filelists import * import os import filecmp """ create a list of classes to enstantiate which will have conversion rules for different language files. Usage: Create a file "localrules.py" that defines globalsubrules to apply to all file types and any of cppsubrules, pythonsubrules,fortransubrules, mpplsubrules,F90subrules that you want to apply to specific file types. These rules are of the form: globalsubrules = [("in1","out1"),("in2","out2"), ...] where string in1 will be converted to string out1, etc. It is an ordered list which will be executed first to last. Another option is that an entry in globalsubrules can be the name of a function of one argument (the string being processed). Note currently, if the ordered list returns "None" (the python None, not the string "None")for a line, no further proceessing of that line is done and no line is written. This can be used to re-order lines, since a function entry used in globalsubrules can be used to return a list of lines, or "None". But no further processing is done of the returned list of lines, so any such re-ordering must be done as the last of the rules.. Note in additon to duples, one can also include in the subrules methods which take as its sole argument the string the method operates on; use this for more complicated operations than simple replacements Also create a local script in the directory where you want to convert files. Begin that script with "from convert import *", making sure that convert.py and localrules.py are in your path. Warning: don't end the local script with the suffix .py or the convertor will appy your rules to `your local script (assuming you use the default suffix as the suffxin for class Py). Further notes 1.If desired, the default suffixes searched to designate files of a specific type can be edited in your local script. to do so import convert, edit the suffix for the sub-class (for example, convert.py.suffix = "newsuffix"), then type "from convert import *". 2.Create sub-classes that inherit from class generic, for any additional file suffix that will be converted. In each subclass, set suffixin to a string that gives the suffix of files to be converted, e.g. ".cc" or ".py". If the processed files are to have a different suffix, set suffixout to the desired suffix. 3.Append processall.classlist, a list of sub-class names that will be processed to do the conversion, to add the names of any classes added in step 2. Then, to proces files: 1.Execute processall(indir,outdir) to process all the specified file types. indir and outdir are optional arguments to specify paths for the directory to be processed and the directory where processed files will be written. By default indir = ".", the current directory, and outdir = "./converted". 2.You can also create instances of specific classes created in step 2, and just process those files. For example a = Py(indir,outdir). See documentation for class generic to see available methods. WARNING: If you've run the script before and already created the outdir, the script will overwrite files in outdir with the same name 3.The script will only process files if the source file is newer than the target (or the target doesn't exist). 4.You can also use processdirs(dirlist) to run processall in a list of directory names specified in dirlist (names in quotes). """ # Global substitution dictionary, globalsubdict = {"instring1:outstring1",...} globalsubdict = {} # subrules of form [("in1","out1"),("in2","out2"), ...] globalsubrules = [] # substitution rules to be appended to globalsubrules for each language cppsubrules = [] # e.g. cppsubrules = [ ("hier::Box<NDIM>", "Box"), # ("hier::IntVector<NDIM>", "IntVect"), # ("hier::Index<NDIM>", "IntVect") ] pythonsubrules = [] fortransubrules = [] F90subrules = [] F90_90subrules = [] TupleType = type((0,0)) ListType = type([0,0]) StringType = type("fubar") MPPLsubrules = [] try: from localrules import * except: print("No file or problem with 'localrules.py'; proceeding with default rules") def fnconvert(name,suffixout): # Converts a file name "name" by substituting suffixout for the # existing suffix, or if there is no suffix, appending suffixout suffixin = name.split(".")[-1] if (suffixin != name): nameout = name.rstrip(suffixin)+suffixout # this coding makes sure only last occurence of suffix is repalced else: nameout = name+"."+suffixout return nameout from stat import ST_MTIME from numpy import greater def newer(file0,file1): # returns 1 if file0 is newer than file1, 0 otherwise # if file0 does not exist, raise a standard python error # if file1 does not exist, return 2 time0 = os.stat(file0)[ST_MTIME] try: time1 = os.stat(file1)[ST_MTIME] except: return 2 return greater(time0,time1) class processdirs: # Process a list of directories to do conversions with processall # The default list of directories is ".", the current directory # Establishes a set of subdirectories with processed files # in a root directory whose default is called "converted" and # is parallel to ".". def __init__(self,indirs = ["."],outroot = "../converted",clean=None): # create the output root directory if it needs to be curpath = os.getenv("PWD") try: os.mkdir(outroot) print("Creating directory "+outroot) except: try: os.chdir(outroot) print("Output directory already exists; proceeding") os.chdir(curpath) except: raise "Can't create output directory" for direc in indirs: print("Entering directory "+direc) if direc == ".": curdir = curpath.split("/")[-1] outdir = outroot+"/"+curdir else: outdir = outroot + "/"+direc processall(direc,outdir,clean) class processall: # Process all files in a directory with rules according to file type # as designated by suffix. # List of classes of distict file types classlist = ["Py","Cpp","Cpp_h","Cpp_hh","Fortran","F90","F90_90","MPPL"] def __init__(self,indir=".",outdir="./converted",clean=None): for entry in self.classlist: a=eval(entry+"("+repr(indir)+","+repr(outdir)+",clean="+repr(clean)+")") # Only process file types with non-empty substition rules if (a.subrules != []): print("processing for file type "+entry) a.process() class generic: """ generic class for replacing strings in a series of files. Methods: process(): processes the list of files in the directory indir with suffix suffixin, and for each one creates a new file in directory outdir with the same root name appended by the string suffixout (which by default is suffixin). processfile(filename): processes a specific file Notes: indir is string with path. By default it is ".", the current directory outdir is by default "./converted". If the output directory does not exist it will be created. Can be overwritten after instance created """ suffixin = "" suffixout = "" subrules = globalsubrules def __init__(self,indir=".",outdir="./converted",clean=None): self.indir = indir self.outdir = outdir self.clean=clean self.doclean = 0 # default is no removal of duplicate files if (self.suffixout == ""): self.suffixout = self.suffixin # create the output directory if it hasn't been created try: os.mkdir(self.outdir) except: pass def process(self): # get the list of files to process self.filelist = filesublist(self.indir,self.suffixin) print( "processing directory" + self.indir+ "to directory" + self.outdir) # exclude convert.py, localrules.py and filelists.py from filelist if (self.suffixin == "py"): try: self.filelist.remove("convert.py") except: pass try: self.filelist.remove("localrules.py") except: pass try: self.filelist.remove("filelists.py") except: pass # Set a flag to remove unchanged files if input and output directories are different if self.clean and (self.outdir != self.indir): # if the input and output directories are different, eliminate files # from the output directory that duplicate those in the input directory. self.doclean = 1 else: self.doclean = 0 # Now process files remaining for file in self.filelist: self.processfile(file) def processfile(self,filename): # Convert an individual file self.infile=self.indir+"/"+filename if (self.suffixout == self.suffixin): outfilename=filename else: outfilename=fnconvert(filename,self.suffixout) self.outfile = self.outdir+"/"+outfilename if (newer(self.infile,self.outfile)): # Only process infile if it is newer than outfile print("converting file "+self.infile+" to "+self.outfile) f=open(self.infile,"r") g=open(self.outfile,"w") lines=f.readlines() iline = 0 # just a diagnostic, counts lines for line in lines: iline = iline + 1 for rule in self.subrules: if (line != None): if type(rule)==TupleType or type(rule)==ListType: line = line.replace(rule[0],rule[1]) else: # if it's not a tuple or list, assume it is a method # that executes something # more complicated than a simple replace line = rule(line) if (line != None): if (type(line) == StringType): g.write(line) if (type(line) == ListType): g.writelines(line) g.close() f.close() if self.doclean == 1: if filecmp.cmp(self.outfile,self.infile) and self.outfile != self.infile: print(self.outfile + " is the same as " + self.infile + ", removing") os.remove(self.outfile) class Py(generic): suffixin="py" subrules = globalsubrules + pythonsubrules class Cpp(generic): suffixin = "C" subrules = globalsubrules + cppsubrules class Cpp_h(Cpp): suffixin = "h" class Cpp_hh(Cpp): suffixin = "hh" class Fortran(generic): suffixin = "f" subrules = globalsubrules + fortransubrules class F90(generic): suffixin = "F" subrules = globalsubrules + F90subrules class F90_90(F90): suffixin = "F90" class MPPL(generic): suffixin = "m" subrules = globalsubrules + MPPLsubrules
11,542
37.734899
92
py
UEDGE
UEDGE-master/pyscripts/double.py
from uedge import * def uedouble(): com.nxleg=2*com.nxleg com.nxcore=2*com.nxcore com.nycore=2*com.nycore com.nysol=2*com.nysol if com.nxomit > 0: if com.geometry=="dnbot": com.nxomit = com.nxleg[0,0]+com.nxcore[0,0] + 2*com.nxxptxx + 1 else: com.nxomit=2*(com.nxomit-2*com.nxxptxx) + 2*com.nxxpt # assumes com.nxomit removes 1/2 SOL if com.nyomitmx == 1: com.nysol = 1 if grd.kxmesh == 4: dxgasold=grd.dxgas alfxold=grd.alfx grd.alfx=alfxold/2. grd.dxgas=dxgasold*(exp(grd.alfx)-1)/(exp(alfxold)-1) grd.nxgas=2*grd.nxgas bbb.restart=1 bbb.newgeo=1 bbb.gengrid=1 bbb.isnintp=1 grd.ixdstar = com.nxcore[0,1]+1
739
28.6
103
py
UEDGE
UEDGE-master/pyscripts/__init__.py
from .uedge import * from os import path from pathlib import Path try: from uedge.__version__ import __version__ from uedge.__src__ import __src__ import uedge.checkver except: try: from __version__ import __version__ from __src__ import __src__ import checkver except: __version__ = 'unknown' __src__ = 'unknown' # # Load the startup file .uedgerc.py from cwd or home. # _homepath = path.expanduser('~') _homefile = Path('{}/.uedgerc.py'.format(_homepath)) _localpath = path.expanduser('.') _localfile = Path('{}/.uedgerc.py'.format(_localpath)) if path.exists(_localfile): with open(_localfile) as f: exec(open(_localfile).read()) elif path.exists(_homefile): with open(_homefile) as f: exec(open(_homefile).read())
805
24.1875
54
py
UEDGE
UEDGE-master/pyscripts/__src__.py
__src__ = '/Users/meyer8/gitstuff/UEDGE'
41
20
40
py
UEDGE
UEDGE-master/pyscripts/cdf4.py
import numpy as np import netCDF4 as nc from .uedge import bbb def cdf4_restore(file): """ Read a cdf4 file previously written from Uedge. This reads the file and puts the 6 standard variables into the correct format. """ try: cf = nc.Dataset(file) except: print("Couldn't open cdf4 file ",file) return try: bbb.ngs[...] = np.array(cf.variables['ngs']) except: print("Couldn't read ngs from ",file) try: bbb.nis[...] = np.array(cf.variables['nis']) except: print("Couldn't read nis from ",file) try: bbb.phis[...] = np.array(cf.variables['phis']) except: print("Couldn't read phis from ",file) try: bbb.tes[...] = np.array(cf.variables['tes']) except: print("Couldn't read tes from ",file) try: bbb.tis[...] = np.array(cf.variables['tis']) except: print("Couldn't read tis from ",file) try: bbb.ups[...] = np.array(cf.variables['ups']) except: print("Couldn't read ups from ",file) try: bbb.tgs[...] = np.array(cf.variables['tgs']) except: print("Couldn't read tgs from ",file) cf.close() def cdf4_save(file): """ Write the 6 standard variables into an cdf4 file. """ try: hf = h5py.File(file,'w') hfg = hf.create_group('bbb') except: print("Couldn't open cdf4 file ",file) try: hfg.create_dataset('ngs',data=bbb.ngs) except: print("Couldn't write ngs to ",file) try: hfg.create_dataset('nis',data=bbb.nis) except: print("Couldn't write nis to ",file) try: hfg.create_dataset('phis',data=bbb.phis) except: print("Couldn't write phis to ",file) try: hfg.create_dataset('tes',data=bbb.tes) except: print("Couldn't write tes to ",file) try: hfg.create_dataset('tis',data=bbb.tis) except: print("Couldn't write tis to ",file) try: hfg.create_dataset('ups',data=bbb.ups) except: print("Couldn't write ups to ",file) try: hfg.create_dataset('tgs',data=bbb.tgs) except: print("Couldn't write ups to ",file) hf.close()
2,293
22.649485
84
py
UEDGE
UEDGE-master/pyscripts/sources.py
# # Bill Meyer - 7/24/2019 # meyer8@llnl.gov # # from __future__ import print_function import sys def sources(): """ This routine simply dumps all the modules as a sorted list along with the "__file__" attribute. Prints "unknown" for those that don't have this attribute. This is just for debug. When a user reports a problem it may be useful to have them run this to make sure they are getting the modules from expected places. """ for m in sorted(sys.modules.keys()): t = sys.modules[m] try: f = t.__file__ except: f = 'unknown' finally: print(m,'\t--\t',f)
670
20.645161
75
py
UEDGE
UEDGE-master/pyscripts/pdb2h5.py
#!/usr/bin/env python # @description This script will convert a pdb data file into an hdf5 # data file # # @keywords Py-UEDGE # # @contact John Cary # # @version $id: $ # import sys for system commands import sys, os # Import uefacets for interface convenience from . import uefacets # import getopts for cmd line parsing import getopt # import uedge and all its structures from .uedge import * def usage(code): print("pdb2h5 [-h] -i <infile> -o <outfile> [-p <paramfile>]") print(" -h: Print this help") sys.exit(code) try: olst, _ = getopt.getopt(sys.argv[1:], "hi:o:p:") except: usage(1) for o in olst: if o[0] == "-h": usage(0) if o[0] == "-i": infile = o[1] if o[0] == "-o": outfile = o[1] if o[0] == "-p": prmfile = o[1] # initialize the uedge object uefacets.init() ue = uefacets.Uedge() try: ue.readParams(prmfile) except NameError as _: print("No parameter file specified, continuing...") ue.buildData() ue.restore(infile) ue.dump(outfile) uefacets.final() sys.exit(0)
1,094
16.380952
68
py
UEDGE
UEDGE-master/pyscripts/rdcontdt.py
# This file runs a time-dependent case using dtreal. First, obtain a converged # solution for a (usually small) dtreal; xuedge must report iterm=1 at the end. # Then adjust control parameters in rdinitdt; read this file, which reads rdinitdt. # If a mistake is made, to restart this file without a Jacobian evaluation, # be sure to reset iterm=1 (=> last step was successful) # IMPORT UEDGE (assuming starting from ipython before any imports) from .uedge import * from .ruthere import * from .uexec import * from numpy import zeros # IMPORT HDF5 routines for saving solutions below from .hdf5 import * # INITIALIZE PARAMS -- SHOULD BE DONE IN MASTER SCRIPT OR TERMINAL SESSION # BEFORE INVOKING THIS SCRIPT uexec("uedge.rdinitdt",returns=globals()) no = 0;yes = 1 echo = no # Set precisions of floating point output ###import print_options ###print_options.set_float_precision(4) # Check if successful time-step exists (bbb.iterm=1) if (bbb.iterm == 1): print("Initial successful time-step exists") bbb.dtreal = bbb.dtreal*bbb.mult_dt #compensates dtreal divided by mult_dt below else: print("*---------------------------------------------------------*") print("Need to take initial step with Jacobian; trying to do here") print("*---------------------------------------------------------*") bbb.icntnunk = 0 bbb.exmain() ruthere() bbb.dtreal = bbb.dtreal*bbb.mult_dt #compensates dtreal divided by mult_dt below if (bbb.iterm != 1): print("*--------------------------------------------------------------*") print("Error: converge an initial time-step first; then retry rdcontdt") print("*--------------------------------------------------------------*") exit() nx=com.nx;ny=com.ny;nisp=com.nisp;ngsp=com.ngsp;numvar=bbb.numvar isteon=bbb.isteon if (i_stor==0): ni_stor = zeros((bbb.n_stor,nx+1+1,ny+1+1,nisp),"d") # set time storage arrays up_stor = zeros((bbb.n_stor,nx+1+1,ny+1+1,nisp),"d") te_stor = zeros((bbb.n_stor,nx+1+1,ny+1+1),"d") ti_stor = zeros((bbb.n_stor,nx+1+1,ny+1+1),"d") ng_stor = zeros((bbb.n_stor,nx+1+1,ny+1+1,ngsp),"d") tg_stor = zeros((bbb.n_stor,nx+1+1,ny+1+1,ngsp),"d") phi_stor = zeros((bbb.n_stor,nx+1+1,ny+1+1),"d") tim_stor = zeros((bbb.n_stor),"d") dtreal_stor = zeros((bbb.n_stor),"d") nfe_stor = zeros((bbb.n_stor),"l") dt_stor = (bbb.tstor_e - bbb.tstor_s)/(bbb.n_stor - 1) i_stor = max(i_stor,1) # set counter for storage arrays bbb.dt_tot = max(bbb.dt_tot,0.) nfe_tot = max(nfe_tot,0) deldt_0 = bbb.deldt isdtsf_sav = bbb.isdtsfscal if (bbb.ipt==1 and bbb.isteon==1): # set ipt to te(nx,iysptrx+1) if no user value ipt = bbb.idxte[nx-1,com.iysptrx] #note: ipt is local, bbb.ipt global bbb.irev = -1 # forces second branch of irev in ii1 loop below if (bbb.iterm == 1): # successful initial run with dtreal bbb.dtreal = bbb.dtreal/bbb.mult_dt # gives same dtreal after irev loop else: # unsuccessful initial run; reduce dtreal bbb.dtreal = bbb.dtreal/(3*bbb.mult_dt) # causes dt=dt/mult_dt after irev loop if (bbb.initjac == 0): bbb.newgeo=0 dtreal_sav = bbb.dtreal bbb.itermx = bbb.itermxrdc bbb.dtreal = bbb.dtreal/bbb.mult_dt #adjust for mult. to follow; mult_dt in rdinitdt bbb.dtphi = bbb.rdtphidtr*bbb.dtreal neq=bbb.neq svrpkg=bbb.svrpkg.tostring().strip() # bbb.ylodt = bbb.yl bbb.pandf1 (-1, -1, 0, bbb.neq, 1., bbb.yl, bbb.yldot) fnrm_old = sqrt(sum((bbb.yldot[0:neq]*bbb.sfscal[0:neq])**2)) if (bbb.initjac == 1): fnrm_old=1.e20 print("initial fnrm =",fnrm_old) for ii1 in range( 1, bbb.ii1max+1): if (bbb.ismfnkauto==1): bbb.mfnksol = 3 # adjust the time-step if (bbb.irev == 0): # Only used after a dt reduc. success. completes loop ii2 for fixed dt bbb.dtreal = min(3*bbb.dtreal,bbb.t_stop) #first move forward after reduction bbb.dtphi = bbb.rdtphidtr*bbb.dtreal if (bbb.ismfnkauto==1 and bbb.dtreal > bbb.dtmfnk3): bbb.mfnksol = -3 bbb.deldt = 3*bbb.deldt else: # either increase or decrease dtreal; depends on mult_dt bbb.dtreal = min(bbb.mult_dt*bbb.dtreal,bbb.t_stop) bbb.dtphi = bbb.rdtphidtr*bbb.dtreal if (bbb.ismfnkauto==1 and bbb.dtreal > bbb.dtmfnk3): bbb.mfnksol = -3 bbb.deldt = bbb.mult_dt*bbb.deldt bbb.dtreal = min(bbb.dtreal,bbb.dt_max) bbb.dtphi = bbb.rdtphidtr*bbb.dtreal if (bbb.ismfnkauto==1 and bbb.dtreal > bbb.dtmfnk3): bbb.mfnksol = -3 bbb.deldt = min(bbb.deldt,deldt_0) bbb.deldt = max(bbb.deldt,bbb.deldt_min) nsteps_nk=1 print('--------------------------------------------------------------------') print('--------------------------------------------------------------------') print(' ') print('*** Number time-step changes = ',ii1,' New time-step = ', bbb.dtreal) print('--------------------------------------------------------------------') bbb.itermx = bbb.itermxrdc if (ii1>1 or bbb.initjac==1): # first time calc Jac if initjac=1 if (bbb.irev == 1): # decrease in bbb.dtreal if (bbb.numrev < bbb.numrevjmax and \ bbb.numrfcum < bbb.numrevjmax+bbb.numfwdjmax): #dont recom bbb.jac bbb.icntnunk = 1 bbb.numrfcum = bbb.numrfcum + 1 else: # force bbb.jac calc, reset numrev bbb.icntnunk = 0 bbb.numrev = -1 # yields api.zero in next statement bbb.numrfcum = 0 bbb.numrev = bbb.numrev + 1 bbb.numfwd = 0 else: # increase in bbb.dtreal if (bbb.numfwd < bbb.numfwdjmax and \ bbb.numrfcum < bbb.numrevjmax+bbb.numfwdjmax): #dont recomp bbb.jac bbb.icntnunk = 1 bbb.numrfcum = bbb.numrfcum + 1 else: bbb.icntnunk = 0 #recompute jacobian for increase dt bbb.numfwd = -1 bbb.numrfcum = 0 bbb.numfwd = bbb.numfwd + 1 bbb.numrev = 0 #bbb.restart counter for dt reversals bbb.isdtsfscal = isdtsf_sav bbb.ftol = min(bbb.ftol_dt, 0.01*fnrm_old) bbb.ftol = max(bbb.ftol, bbb.ftol_min) exmain() # take a single step at the present bbb.dtreal ruthere() if (bbb.iterm == 1): bbb.dt_tot = bbb.dt_tot + bbb.dtreal nfe_tot = nfe_tot + bbb.nfe[0,0] bbb.ylodt = bbb.yl bbb.pandf1 (-1, -1, 0, bbb.neq, 1., bbb.yl, bbb.yldot) fnrm_old = sqrt(sum((bbb.yldot[0:neq-1]*bbb.sfscal[0:neq-1])**2)) if (bbb.dt_tot>=0.9999999*bbb.t_stop or fnrm_old<bbb.ftol_min): print(' ') print('*****************************************************') print('** SUCCESS: frnm < bbb.ftol; or dt_tot >= t_stop **') print('*****************************************************') break bbb.icntnunk = 1 bbb.isdtsfscal = 0 for ii2 in range( 1, bbb.ii2max+1): #take ii2max steps at the present time-step if (bbb.iterm == 1): bbb.itermx = bbb.itermxrdc bbb.ftol = min(bbb.ftol_dt, 0.01*fnrm_old) bbb.ftol = max(bbb.ftol, bbb.ftol_min) bbb.exmain() ruthere() if (bbb.iterm == 1): bbb.ylodt = bbb.yl bbb.pandf1 (-1, -1, 0, bbb.neq, 1., bbb.yl, bbb.yldot) fnrm_old = sqrt(sum((bbb.yldot[0:neq-1]*bbb.sfscal[0:neq-1])**2)) print("Total time = ",bbb.dt_tot,"; Timestep = ",bbb.dtreal) print("variable index ipt = ",ipt, " bbb.yl[ipt] = ",bbb.yl[ipt]) dtreal_sav = bbb.dtreal bbb.dt_tot = bbb.dt_tot + bbb.dtreal nfe_tot = nfe_tot + bbb.nfe[0,0] hdf5_save(savefn) if (bbb.dt_tot>=0.999999999999*bbb.t_stop or fnrm_old<bbb.ftol_min): print(' ') print('*****************************************************') print('** SUCCESS: frnm < bbb.ftol; or dt_tot >= t_stop **') print('*****************************************************') break print(" ") ## Store variables if a storage time has been crossed if (bbb.dt_tot >= dt_stor*i_stor and i_stor<=bbb.n_stor): i_stor1 = i_stor-1 ni_stor[i_stor1,:,:,:] = ni up_stor[i_stor1,:,:,:] = up te_stor[i_stor1,:,:] = te ti_stor1[i_stor1,:,:] = ti ng_stor[i_stor1,:,:,:] = ng phi_stor1[i_stor1,:,:] = phi tim_stor[i_stor1] = bbb.dt_tot nfe_stor[i_stor1] = nfe_tot dtreal_stor[i_stor1] = bbb.dtreal i_stor = i_stor + 1 ## End of storage section if (bbb.dt_tot>=bbb.t_stop or fnrm_old<bbb.ftol_min): break # need for both loops bbb.irev = bbb.irev-1 if (bbb.iterm != 1): #print bad eqn, cut dtreal by 3, set irev flag ####### a copy of idtroub script ######################## oldecho=echo echo=no # integer ii # real8 ydmax scalfac = bbb.sfscal if (svrpkg != "nksol"): scalfac = 1/(bbb.yl + 1.e-30) # for time-dep calc. ydmax = 0.999999999*max(abs(bbb.yldot*scalfac)) itrouble = 0 for ii in range(neq): if (abs(bbb.yldot[ii]*scalfac[ii]) > ydmax): itrouble=ii print("** Fortran index of trouble making equation is:") print(itrouble+1) break print("** Number of variables is:") print("numvar = ", numvar) print(" ") iv_t = (itrouble).__mod__(numvar) + 1 print("** Troublemaker equation is:") print("iv_t = ",iv_t) print(" ") print("** Troublemaker cell (ix,iy) is:") print(bbb.igyl[itrouble,]) print(" ") print("** Timestep for troublemaker equation:") print(bbb.dtuse[itrouble]) print(" ") print("** yl for troublemaker equation:") print(bbb.yl[itrouble]) print(" ") echo=oldecho ######## end of idtroub script ############################## if (bbb.dtreal < bbb.dt_kill): print(' ') print('*************************************') print('** FAILURE: time-step < dt_kill **') print('*************************************') break bbb.irev = 1 print('*** Converg. fails for bbb.dtreal; reduce time-step by 3, try again') print('----------------------------------------------------------------- ') bbb.dtreal = bbb.dtreal/(3*bbb.mult_dt) bbb.dtphi = bbb.rdtphidtr*bbb.dtreal if (bbb.ismfnkauto==1 and bbb.dtreal > bbb.dtmfnk3): bbb.mfnksol = -3 bbb.deldt = bbb.deldt/(3*bbb.mult_dt) bbb.iterm = 1 echo = yes
10,702
41.472222
88
py
UEDGE
UEDGE-master/test/Forthon_cases/Forthon_case3/rd_forthon_case3.py
#Typical mesh (nx=64, ny=32) for DIII-D MHD equilibrium #Uses inertial neutrals, so six variables (ni,upi,Te,Ti,ng,upg) # ##package flx;package grd;package bbb # Initialize pyuedge from uedge import * # Set the geometry bbb.mhdgeo = 1 #use MHD equilibrium os.system('rm -f aeqdsk neqdsk') #change names of MHD eqil. files os.system('cp aeqdskd3d aeqdsk') # (Cannot tab or indent these 3 lines) os.system('cp neqdskd3d neqdsk') flx.psi0min1 = 0.98 #normalized flux on core bndry flx.psi0min2 = 0.98 #normalized flux on pf bndry flx.psi0sep = 1.00001 #normalized flux at separatrix flx.psi0max = 1.07 #normalized flux on outer wall bndry flx.alfcy = 3. bbb.ngrid = 1 #number of meshes (always set to 1) com.nxleg[0,0] = 16 #pol. mesh pts from inner plate to x-point com.nxcore[0,0] = 16 #pol. mesh pts from x-point to top on inside com.nxcore[0,1] = 16 #pol. mesh pts from top to x-point on outside com.nxleg[0,1] = 16 #pol. mesh pts from x-point to outer plate com.nysol[0] = 24 #rad. mesh pts in SOL com.nycore[0] =8 #rad. mesh pts in core # Finite-difference algorithms (upwind, central diff, etc.) bbb.methn = 33 #ion continuty eqn bbb.methu = 33 #ion parallel momentum eqn bbb.methe = 33 #electron energy eqn bbb.methi = 33 #ion energy eqn bbb.methg = 33 #neutral gas continuity eqn # Boundary conditions bbb.ncore[0] = 4.0e19 #hydrogen ion density on core bbb.isnwcono = 3 #=3; use lyne bbb.lyni = 0.05 ## iflcore = 0 #flag; =0, fixed Te,i; =1, fixed power on core bbb.tcoree = 300. #core Te bbb.tcorei = 300. #core Ti bbb.istewc = 3 #=3 for gradient-length = lyte bbb.istiwc = 3 #=3 for gradient-length = lyti bbb.lyte = 0.05 bbb.lyti = 0.05 bbb.recycp[0] = 1.0 #hydrogen recycling coeff at plates bbb.recycw[0] = 1.0 bbb.nwsor = 1 bbb.matwso[0] = 1 # Transport coefficients bbb.difni[0] = 0.3 #D for radial hydrogen diffusion bbb.kye = 0.5 #chi_e for radial elec energy diffusion bbb.kyi = 0.5 #chi_i for radial ion energy diffusion bbb.travis[0] = 1. #eta_a for radial ion momentum diffusion # Flux limits bbb.flalfe = 0.21 #electron parallel thermal conduct. coeff bbb.flalfi = 0.21 #ion parallel thermal conduct. coeff bbb.flalfv = 0.5 #ion parallel viscosity coeff bbb.flalfgx = 1. #neut. gas in poloidal direction bbb.flalfgy = 1. #neut. gas in radial direction bbb.lgmax = 0.1 bbb.lgtmax = 0.1 bbb.lgvmax = 0.1 # Solver package bbb.svrpkg = "nksol" #Newton solver using Krylov method bbb.premeth = "ilut" #Solution method for precond. Jacobian matrix # Neutral parallel momentum eqn bbb.isupgon[0] = 1 bbb.isngon[0] = 0 com.ngsp = 1 com.nhsp = 2 bbb.ziin[com.nhsp-1] = 0 bbb.cfnidh = 0. #coeff. heating from charge-exchange # Restart from a pfb savefile bbb.restart = 1 #Begin from savefile, not estimated profiles bbb.allocate() #allocate space for savevariables restore('pfd3d_ex_upg.64x32m') #read in the solution from pfb file ###os.system('ln -s ~/Uedge/uedge/in/aph aph6') aph.isaphdir = 0 #=0 if atomic data file in run directory com.istabon = 10 # Execute uedge bbb.exmain() # Print out a few variables across outer midplane print'' print'*** Printing variables versus radial index at outer midplane' print'' print '************\nradial position relative to separatrix [m]' print(com.yyc) print '************\n ion density, ni [m**-3] = ' print(bbb.ni[bbb.ixmp,]) print '************\n parallel ion velocity, up [m/s] = ' print(bbb.up[bbb.ixmp,]) print '************\n electron temp, te [eV] = ' print(bbb.te[bbb.ixmp,]/bbb.ev) print '************\n ion temp, ti [eV] = ' print(bbb.ti[bbb.ixmp,]/bbb.ev) print '************\n gas density, ng [m**-3] = ' print(bbb.ng[bbb.ixmp,])
3,833
34.831776
74
py
UEDGE
UEDGE-master/test/Forthon_cases/Forthon_case2/rd_forthon_case2.py
#Coarse mesh (nx=16, ny=8) for DIII-D MHD equilibrium #Uses diffusive neutrals, so five variables (ni,upi,Te,Ti,ng) # ##package flx;package grd;package bbb # Initialize pyuedge from uedge import * #from uefacets import * from uedge.pdb_restore import * from uedge.hdf5 import * #bbb.uedge_petscInit() # Set the geometry bbb.mhdgeo = 1 #use MHD equilibrium os.system('rm -f aeqdsk neqdsk') #change names of MHD eqil. files os.system('cp aeqdskd3d aeqdsk') # (Cannot tab or indent these 3 lines) os.system('cp neqdskd3d neqdsk') flx.psi0min1 = 0.98 #normalized flux on core bndry flx.psi0min2 = 0.98 #normalized flux on pf bndry flx.psi0sep = 1.00001 #normalized flux at separatrix flx.psi0max = 1.07 #normalized flux on outer wall bndry bbb.ngrid = 1 #number of meshes (always set to 1) com.nxleg[0,0] = 4 #pol. mesh pts from inner plate to x-point com.nxcore[0,0] = 4 #pol. mesh pts from x-point to top on inside com.nxcore[0,1] = 4 #pol. mesh pts from top to x-point on outside com.nxleg[0,1] = 4 #pol. mesh pts from x-point to outer plate com.nysol[0] = 6 #rad. mesh pts in SOL com.nycore[0] =2 #rad. mesh pts in core # Finite-difference algorithms (upwind, central diff, etc.) bbb.methn = 33 #ion continuty eqn bbb.methu = 33 #ion parallel momentum eqn bbb.methe = 33 #electron energy eqn bbb.methi = 33 #ion energy eqn bbb.methg = 33 #neutral gas continuity eqn # Boundary conditions bbb.ncore[0] = 2.5e19 #hydrogen ion density on core ## iflcore = 0 #flag; =0, fixed Te,i; =1, fixed power on core bbb.tcoree = 100. #core Te bbb.tcorei = 100. #core Ti bbb.tedge = 2. #fixed wall,pf Te,i if istewcon=1, etc bbb.recycp[0] = 0.8 #hydrogen recycling coeff at plates # Transport coefficients bbb.difni[0] = 1. #D for radial hydrogen diffusion bbb.kye = 1. #chi_e for radial elec energy diffusion bbb.kyi = 1. #chi_i for radial ion energy diffusion bbb.travis[0] = 1. #eta_a for radial ion momentum diffusion # Flux limits bbb.flalfe = 0.21 #electron parallel thermal conduct. coeff bbb.flalfi = 0.21 #ion parallel thermal conduct. coeff bbb.flalfv = 1. #ion parallel viscosity coeff bbb.flalfgx = 1.e20 #neut. gas in poloidal direction bbb.flalfgy = 1.e20 #neut. gas in radial direction # Solver package bbb.svrpkg = "nksol" #Newton solver using Krylov method #bbb.svrpkg = "petsc" #Newton solver using Krylov method bbb.premeth = "banded" #Solution method for precond. Jacobian matrix # Restart from a pfb savefile bbb.restart = 1 #Begin from savefile, not estimated profiles bbb.allocate() #allocate space for savevariables bbb.nis.shape #pdb_restore('./pfd3d_ex.16x8') #read in the solution from pfb file hdf5_restore('./h5d3d_ex.16x8') #hdf5_restore('./testhdf5') #hdf5_save('testhdf5') #restore('pfd3d_ex.16x8') #read in the solution from pfb file os.system('ln -s ~/Uedge/uedge/in/aph aph6') com.istabon = 10 aph.isaphdir = 0 #=0 specifies atomic data file is in run directory # Execute uedge bbb.exmain() # Print out a few variables across outer midplane print'' print'*** Printing variables versus radial index at outer midplane' print'' print '************\nradial position relative to separatrix [m]' print(com.yyc) print '************\n ion density, ni [m**-3] = ' print(bbb.ni[bbb.ixmp,]) print '************\n parallel ion velocity, up [m/s] = ' print(bbb.up[bbb.ixmp,]) print '************\n electron temp, te [eV] = ' print(bbb.te[bbb.ixmp,]/bbb.ev) print '************\n ion temp, ti [eV] = ' print(bbb.ti[bbb.ixmp,]/bbb.ev)
3,496
36.202128
74
py
UEDGE
UEDGE-master/test/Forthon_cases/Forthon_case6/rdtimedpd_1.py
#Coarse mesh (nx=16, ny=8) for DIII-D MHD equilibrium #Uses diffusive neutrals, so five variables (ni,upi,Te,Ti,ng) # ##package flx;package grd;package bbb # Initialize pyuedge from uedge import * # Set the geometry bbb.mhdgeo = 1 #use MHD equilibrium os.system('rm -f aeqdsk neqdsk') #change names of MHD eqil. files os.system('cp aeqdskd3d aeqdsk') # (Cannot tab or indent these 3 lines) os.system('cp neqdskd3d neqdsk') flx.psi0min1 = 0.98 #normalized flux on core bndry flx.psi0min2 = 0.98 #normalized flux on pf bndry flx.psi0sep = 1.00001 #normalized flux at separatrix flx.psi0max = 1.07 #normalized flux on outer wall bndry bbb.ngrid = 1 #number of meshes (always set to 1) com.nxleg[0,0] = 4 #pol. mesh pts from inner plate to x-point com.nxcore[0,0] = 4 #pol. mesh pts from x-point to top on inside com.nxcore[0,1] = 4 #pol. mesh pts from top to x-point on outside com.nxleg[0,1] = 4 #pol. mesh pts from x-point to outer plate com.nysol[0] = 6 #rad. mesh pts in SOL com.nycore[0] =2 #rad. mesh pts in core # Finite-difference algorithms (upwind, central diff, etc.) bbb.methn = 33 #ion continuty eqn bbb.methu = 33 #ion parallel momentum eqn bbb.methe = 33 #electron energy eqn bbb.methi = 33 #ion energy eqn bbb.methg = 33 #neutral gas continuity eqn # Boundary conditions bbb.ncore[0] = 2.5e19 #hydrogen ion density on core ## iflcore = 0 #flag; =0, fixed Te,i; =1, fixed power on core bbb.tcoree = 100. #core Te bbb.tcorei = 100. #core Ti bbb.tedge = 2. #fixed wall,pf Te,i if istewcon=1, etc bbb.recycp[0] = 0.8 #hydrogen recycling coeff at plates # Transport coefficients bbb.difni[0] = 1. #D for radial hydrogen diffusion bbb.kye = 1. #chi_e for radial elec energy diffusion bbb.kyi = 1. #chi_i for radial ion energy diffusion bbb.travis[0] = 1. #eta_a for radial ion momentum diffusion # Flux limits bbb.flalfe = 0.21 #electron parallel thermal conduct. coeff bbb.flalfi = 0.21 #ion parallel thermal conduct. coeff bbb.flalfv = 1. #ion parallel viscosity coeff bbb.flalfgx = 1.e20 #neut. gas in poloidal direction bbb.flalfgy = 1.e20 #neut. gas in radial direction # Solver package bbb.svrpkg = "nksol" #Newton solver using Krylov method bbb.premeth = "banded" #Solution method for precond. Jacobian matrix # Restart from a pfb savefile bbb.restart = 1 #Begin from savefile, not estimated profiles bbb.allocate() #allocate space for savevariables restore('pfd3d_ex.16x8') #read in the solution from pfb file ###os.system('ln -s ~/Uedge/uedge/in/aph aph6') com.istabon = 10 aph.isaphdir = 0 #=0 specifies atomic data file is in run directory # Set initial time-step and execute uedge bbb.dtreal=1e-4 bbb.exmain() # Read the initialization file for time-dependent parameters, and run to S.S. execfile("rdinitdt.py") execfile("rdcontdt.py")
2,813
35.545455
77
py
UEDGE
UEDGE-master/test/Forthon_cases/Forthon_case6/rdtimedpd_2.py
#Coarse mesh (nx=16, ny=8) for DIII-D MHD equilibrium #Uses diffusive neutrals, so five variables (ni,upi,Te,Ti,ng) # ##package flx;package grd;package bbb # Initialize pyuedge from uedge import * # Set the geometry bbb.mhdgeo = 1 #use MHD equilibrium os.system('rm -f aeqdsk neqdsk') #change names of MHD eqil. files os.system('cp aeqdskd3d aeqdsk') # (Cannot tab or indent these 3 lines) os.system('cp neqdskd3d neqdsk') flx.psi0min1 = 0.98 #normalized flux on core bndry flx.psi0min2 = 0.98 #normalized flux on pf bndry flx.psi0sep = 1.00001 #normalized flux at separatrix flx.psi0max = 1.07 #normalized flux on outer wall bndry bbb.ngrid = 1 #number of meshes (always set to 1) com.nxleg[0,0] = 4 #pol. mesh pts from inner plate to x-point com.nxcore[0,0] = 4 #pol. mesh pts from x-point to top on inside com.nxcore[0,1] = 4 #pol. mesh pts from top to x-point on outside com.nxleg[0,1] = 4 #pol. mesh pts from x-point to outer plate com.nysol[0] = 6 #rad. mesh pts in SOL com.nycore[0] =2 #rad. mesh pts in core # Finite-difference algorithms (upwind, central diff, etc.) bbb.methn = 33 #ion continuty eqn bbb.methu = 33 #ion parallel momentum eqn bbb.methe = 33 #electron energy eqn bbb.methi = 33 #ion energy eqn bbb.methg = 33 #neutral gas continuity eqn # Boundary conditions bbb.ncore[0] = 10.e19 #hydrogen ion density on core ## iflcore = 0 #flag; =0, fixed Te,i; =1, fixed power on core bbb.tcoree = 100. #core Te bbb.tcorei = 100. #core Ti bbb.tedge = 2. #fixed wall,pf Te,i if istewcon=1, etc bbb.recycp[0] = 0.8 #hydrogen recycling coeff at plates # Transport coefficients bbb.difni[0] = 1. #D for radial hydrogen diffusion bbb.kye = 1. #chi_e for radial elec energy diffusion bbb.kyi = 1. #chi_i for radial ion energy diffusion bbb.travis[0] = 1. #eta_a for radial ion momentum diffusion # Flux limits bbb.flalfe = 0.21 #electron parallel thermal conduct. coeff bbb.flalfi = 0.21 #ion parallel thermal conduct. coeff bbb.flalfv = 1. #ion parallel viscosity coeff bbb.flalfgx = 1.e20 #neut. gas in poloidal direction bbb.flalfgy = 1.e20 #neut. gas in radial direction # Solver package bbb.svrpkg = "nksol" #Newton solver using Krylov method bbb.premeth = "banded" #Solution method for precond. Jacobian matrix # Restart from a pfb savefile bbb.restart = 1 #Begin from savefile, not estimated profiles bbb.allocate() #allocate space for savevariables restore('pfd3d_ex.16x8') #read in the solution from pfb file ###os.system('ln -s ~/Uedge/uedge/in/aph aph6') com.istabon = 10 aph.isaphdir = 0 #=0 specifies atomic data file is in run directory # Set initial time-step and execute uedge bbb.dtreal=1e-2 bbb.itermx=10 bbb.exmain() # Read the initialization file for time-dependent parameters, and run to S.S. execfile("rdinitdt.py") execfile("rdcontdt.py")
2,827
35.25641
77
py
UEDGE
UEDGE-master/test/Forthon_cases/Forthon_case5/plate.iter-feat.py
# define divertor plate for ITER-FEAT # derived from equilibrium from Kukushkin Jan 2002 ###integer oldecho=echo ###echo=no # for inboard half of mesh: grd.nplate1=15 grd.gchange("Mmod",0) grd.rplate1=[\ 3.96500E+00, 4.06360E+00, 4.29140E+00, 4.44480E+00, 4.48730E+00, \ 4.40860E+00, 4.23980E+00, 4.07090E+00, 4.34190E+00, 4.42400E+00, \ 4.47660E+00, 4.52350E+00, 4.55570E+00, 4.76130E+00, 4.96790E+00] grd.zplate1=[\ 2.94670E+00, 2.93000E+00, 2.83000E+00, 2.63410E+00, 2.38890E+00, \ 2.15280E+00, 1.88820E+00, 1.62250E+00, 1.37820E+00, 1.40330E+00, \ 1.61430E+00, 1.80290E+00, 1.84040E+00, 1.88310E+00, 1.84590E+00] # for outboard half of mesh: grd.nplate2=14 grd.gchange("Mmod",0) grd.rplate2=[\ 4.96790E+00, 5.14560E+00, 5.26870E+00, 5.26710E+00, \ 5.15840E+00, 5.04220E+00, 5.07300E+00, 5.56490E+00, 5.56450E+00, \ 5.56350E+00, 5.61140E+00, 5.74950E+00, 5.95820E+00, 6.20770E+00] grd.zplate2=[\ 1.84590E+00, 1.73420E+00, 1.56400E+00, 1.51460E+00, \ 1.31340E+00, 1.09830E+00, 1.02620E+00, 8.91300E-01, 1.24570E+00, \ 1.59910E+00, 1.85160E+00, 2.06830E+00, 2.21830E+00, 2.28010E+00] ###echo=oldecho
1,132
31.371429
68
py
UEDGE
UEDGE-master/test/Forthon_cases/Forthon_case5/rd_forthon_case5.py
###character*6 probname="itfa40" #Case of full toroidal equilibrium for ITER-FEAT with multi-species Carbon # ####package flx;package grd;package bbb # Initial pyuedge from uedge import * from pdb_restore import * bbb.mhdgeo=1 bbb.isfixlb=0 bbb.isfixrb=0 os.system('rm -f aeqdsk neqdsk') os.system('cp aeq_iter-feat aeqdsk') os.system('cp geq_iter-feat neqdsk') ###character*9 machine="bbb.iter-feat" # Set the geometry bbb.ngrid = 1 grd.kxmesh = 1 #=4 for exponential grid in leg regions grd.dxgas[0] = 1.2e-03 grd.dxgas[1] = 1.2e-03 grd.nxgas[0] = 11 grd.nxgas[1] = 11 grd.alfx[0] = .64 grd.alfx[1] = .64 com.nxleg[0,0]=17 com.nxleg[0,1]=17 com.nxcore[0,0]=14 com.nxcore[0,1]=14 com.nycore[0]=10 com.nysol[0]=16 flx.psi0min1 = 0.95 #core minimum psi flx.psi0min2 = 0.992 #private flux minimum psi flx.psi0max = 1.035 #maximum flux at wall flx.alfcy = 2.0 #nonuniformity factor for radial mesh grd.slpxt=1.2 # Mesh construction--non orthogonal mesh com.ismmon=3 grd.istream=0 grd.iplate=1 grd.nsmooth=3 grd.wtmesh1=0.75 grd.dmix0=1.0 execfile('plate.iter-feat.py') com.isnonog=1 # non orthogonal differencing # Boundary conditions bbb.isnicore[0] = 1 #=1uses ncore for density BC bbb.isngcore[0]=2 # use ionization scale length for gas bbb.ncore[0] = 6.0e19 #value of core density if isnicore=1 bbb.curcore = 0. #core particle current if isnicore=0 bbb.iflcore = 1 #specify core power bbb.pcoree = 5.0e7 #electron power across core bbb.pcorei = 5.0e7 #ion power across core bbb.recycp=1.0 bbb.istepfc=3;bbb.istipfc=3 #priv. flux has fixed temperature scale length. bbb.istewc=3;bbb.istiwc=3 #wall has fixed temperature scale length. bbb.isnwcono=3;bbb.isnwconi=3 #walls have fixed density scale length bbb.lyni=0.05;bbb.lyte=0.05;bbb.lyti=0.05 # set walls into 7 zones bbb.nwsor=7 bbb.xgaso[0:7]=[1.533E-01, 4.599E-01, 3.506E+00, 9.291E+00, 1.508E+01, 1.817E+01, 1.858E+01] bbb.wgaso[0:7]=[3.066E-01, 3.066E-01, 5.790E+00, 5.790E+00, 5.790E+00, 5.000E-01, 4.089E-01] bbb.albdso[0:7]=1. bbb.matwso[0:7]=1. bbb.xgasi[0:7]=[1.213E-01, 3.638E-01, 6.063E-01, 9.024E-01, 1.246E+00, 1.541E+00, 1.837E+00] bbb.wgasi[0:7]=[2.425E-01, 2.425E-01, 2.425E-01, 3.705E-01, 2.955E-01, 2.955E-01, 2.955E-01] bbb.albdsi[0:7]=[0.98,0.98,1.0,1.0,1.0,0.98,0.98] bbb.matwsi[0:7]=1 bbb.recycw[0]=1.0 bbb.bcee = 5.; bbb.bcei = 3.5 #energy transmission coeffs. bbb.isupss = 1 #parallel vel can be supersonic # Transport coefficients bbb.difni[0] = 0.3 bbb.kye = 1.; bbb.kyi = 1. bbb.travis[0]=1.;bbb.parvis[0]=1. # Flux limits bbb.flalfe=0.21;bbb.flalfi=0.21;bbb.flalfgx=1.;bbb.flalfgy=1.;bbb.flalfgxy=1.;bbb.flalfv=0.5 bbb.lgmax=0.05 bbb.isplflxl=0 # Finite difference algorithms bbb.methe=33;bbb.methu=33;bbb.methg=66 bbb.methn=33;bbb.methi=33 # Solver package bbb.svrpkg="nksol" bbb.mfnksol = 3 bbb.epscon1 = .005 bbb.ftol = 1.e-8 bbb.iscolnorm = 3 # set to 3 for nksol bbb.premeth="ilut" bbb.lfililut = 100 bbb.lenpfac=75 bbb.runtim=1.e-7 bbb.rlx=0.9 ###bbb.del=1.e-8 # Neutral gas properties bbb.tfcx=5.;bbb.tfcy=5. #Franck-Condon temperatures bbb.eion = 5. #F-C energy to each born ion bbb.ediss = 10. #diss. energy from elec. (ediss=2*eion) bbb.isrecmon = 1 #e-i recombination (=1 is on) bbb.ngbackg=1.e12 # minimum floor neutral density bbb.ingb=4 # parameter used to force floor density # Inertial neutral model bbb.isupgon[0]=1;bbb.isngon[0]=0;com.ngsp=1;com.nhsp=2;bbb.ziin[com.nhsp-1]=0 bbb.cngmom=0;bbb.cmwall=0;bbb.cngtgx=0;bbb.cngtgy=0;bbb.cngflox=0;bbb.cngfloy=0;bbb.cfbgt=0 bbb.kxn=0;bbb.kyn=0 bbb.flalftgx=10.0;bbb.flalftgy=10.0 # Currents and potential parameters bbb.isphion=0 bbb.rsigpl=1.e-8 #anomalous cross-field conductivity bbb.cfjhf=0. #turn-on heat flow from current (fqp) bbb.jhswitch=0 #Joule Heating switch # Hydrogenic ions bbb.minu[0:2] = 2.5 # Atomic physics packages com.istabon=10 #Stotler's rates for istabon=10 aph.isaphdir = 0 #=0 specifies atomic data file is in run directory ## Impurity gas com.ngsp = 2 #total number of gas species bbb.isngon[1] = 1 #turns on impurity gas bbb.ngbackg[1] = 1.e9 bbb.ingb = 2 bbb.istgcon[1] = 1 #=1 for constant tg(2) at tgas(2) bbb.tgas[1] = 1. #value for tg when istgcon=1 bbb.rcxighg = 0. # best value; ratio of imp cx to hyd cx bbb.kelighi[1] = 5.e-16 #elastic sig_v for imp_gas/h_ion bbb.kelighg[1] = 5.e-16 #elastic sig_v for imp_gas/h_gas bbb.n0g[1] = 1.e16 #imp. gas density normalization # Impurity gas boundary conditions bbb.recycp[1] = 0.01 #plate recycling of impurities bbb.recycw[1] = 1e-4 #wall recycling; matwsi,o set above for hyd bbb.isch_sput[1]=7;bbb.isph_sput[1]=3 # Haasz/Davis sputtering model bbb.t_wall=300;bbb.t_plat=500 #wall and plate temperatures bbb.crmb=bbb.minu[0] #mass of bombarding particles bbb.allocate() #allocate chemywi,o etc bbb.fchemywi=1.;bbb.fchemywo=1. #scaling factor for chem sputt walls bbb.fchemylb=1.;bbb.fchemyrb=1. #scaling factor for chem sputt plates ## Impurity ions bbb.isimpon = 6 #Use force-balance only com.nzsp[0] = 6 #number chrg states impurity isotope #1 bbb.csfaclb[2:8] = 2.191 #at plate csout=sqrt(mi_imp/m_dt) for up=up_imp bbb.csfacrb[2:8] = 2.191 #at plate csout=sqrt(mi_imp/m_dt) for up=up_imp bbb.minu[2:8] = 12. #mass in AMU; python doesn't fill last place of [2:8] bbb.ziin[2:8] = [1,2,3,4,5,6] #charge per ion bbb.znuclin[0:2] = 1 #nuclear charge for hydrogen species (python fills 2-1) bbb.znuclin[2:8] = 6 #nuclear charge for impurity (python fills 8-1) bbb.n0[2:8] = 1.e17 #global density normalization bbb.nzbackg = 1.e9 #background density for impurities bbb.inzb = 2 #exponent for switching on nzbackg bbb.ismctab = 2 # use Braams' rate tables com.mcfilename[0] = "C" # Imp rate file name com.mcfilename[1] = "_" com.mcfilename[2] = "r" com.mcfilename[3] = "a" com.mcfilename[4] = "t" com.mcfilename[5] = "e" com.mcfilename[6] = "s" com.mcfilename[7] = "." com.mcfilename[8] = "a" com.mcfilename[9] = "d" com.mcfilename[10] = "a" com.mcfilename[11] = "s" com.mcfilename = 'C_rates_adas\0' # Impurity ion boundary conditions ###bbb.isnicore(com.nhsp+com.nzsp[0])= 3 #=3 for flux=curcore & constant ni ###bbb.curcore(com.nhsp+com.nzsp[0]) = 0. #Rcurrent for isnicore=3 bbb.isnicore[com.nhsp+com.nzsp[0]-1]= 3 #=3 for flux=curcore & constant ni bbb.curcore[com.nhsp+com.nzsp[0]-1] = 0. #Rcurrent for isnicore=3 bbb.isnwcono[2:8] = 3 #use lyni scale-length (set above); outer wall bbb.isnwconi[2:8] = 3 #use lyni scale-length (set above); inner wall bbb.nwomin[2:8] = 1e7 #minimum ni at outer sidewall bbb.nwimin[2:8] = 1e7 #minimum ni at inner sidewall # Restart from a save file bbb.restart = 1 bbb.allocate() pdb_restore ('pfiter_msC.15') # Execute UEDGE for this case bbb.exmain() # Print out a few variables across outer midplane print'' print'*** Printing variables versus radial index at outer midplane' print'' print '************\nradial position relative to separatrix [m]' print(com.yyc) print '************\n ion densities, ni [m**-3] = ' print(bbb.ni[bbb.ixmp,]) print '************\n parallel ion velocity, up [m/s] = ' print(bbb.up[bbb.ixmp,]) print '************\n electron temp, te [eV] = ' print(bbb.te[bbb.ixmp,]/bbb.ev) print '************\n ion temp, ti [eV] = ' print(bbb.ti[bbb.ixmp,]/bbb.ev) print '************\n gas densities, ng [m**-3] = ' print(bbb.ng[bbb.ixmp,])
7,524
34.163551
92
py
UEDGE
UEDGE-master/test/Forthon_cases/Forthon_case1/rd_forthon_case1.py
# This input file sets the parameters for the parallel test case # from uedge import * # flags also set in the workstation version. bbb.mhdgeo=-1 bbb.isnion=1 bbb.isupon=1 bbb.isteon=1 bbb.istion=1 bbb.isngon=0 bbb.svrpkg="vodpk" bbb.premeth="ilut" bbb.epscon1=1.e-2 bbb.ireorder=0 bbb.ncore=2.e19 bbb.tcoree=100 bbb.tcorei=100 bbb.isfixlb=2 bbb.recycp=.9 bbb.runtim=1.e-7 ##bbb.trange=4.e6 bbb.trange=4.e3 bbb.nsteps=30 bbb.n0g=1.e16 bbb.difni=1. bbb.kye=1. bbb.flalfe=0.21 bbb.flalfi=0.21 bbb.flalfgx=1.e10 bbb.flalfgy=1.e10 com.nycore=0 com.nysol=10 com.nxleg[0,0]=0 com.nxleg[0,1]=2 com.nxcore[0,0]=0 com.nxcore[0,1]=4 grd.zax=1. grd.zaxpt=.75 grd.alfyt=-1.e-5 print "Finished setting variables" print "Allocate Storage." bbb.allocate () bbb.restart=0 bbb.exmain()
773
14.795918
64
py
UEDGE
UEDGE-master/test/Forthon_cases/Forthon_case4/plate.iter-feat.py
# define divertor plate for ITER-FEAT # derived from equilibrium from Kukushkin Jan 2002 ###integer oldecho=echo ###echo=no # for inboard half of mesh: grd.nplate1=15 grd.gchange("Mmod",0) grd.rplate1=[\ 3.96500E+00, 4.06360E+00, 4.29140E+00, 4.44480E+00, 4.48730E+00, \ 4.40860E+00, 4.23980E+00, 4.07090E+00, 4.34190E+00, 4.42400E+00, \ 4.47660E+00, 4.52350E+00, 4.55570E+00, 4.76130E+00, 4.96790E+00] grd.zplate1=[\ 2.94670E+00, 2.93000E+00, 2.83000E+00, 2.63410E+00, 2.38890E+00, \ 2.15280E+00, 1.88820E+00, 1.62250E+00, 1.37820E+00, 1.40330E+00, \ 1.61430E+00, 1.80290E+00, 1.84040E+00, 1.88310E+00, 1.84590E+00] # for outboard half of mesh: grd.nplate2=14 grd.gchange("Mmod",0) grd.rplate2=[\ 4.96790E+00, 5.14560E+00, 5.26870E+00, 5.26710E+00, \ 5.15840E+00, 5.04220E+00, 5.07300E+00, 5.56490E+00, 5.56450E+00, \ 5.56350E+00, 5.61140E+00, 5.74950E+00, 5.95820E+00, 6.20770E+00] grd.zplate2=[\ 1.84590E+00, 1.73420E+00, 1.56400E+00, 1.51460E+00, \ 1.31340E+00, 1.09830E+00, 1.02620E+00, 8.91300E-01, 1.24570E+00, \ 1.59910E+00, 1.85160E+00, 2.06830E+00, 2.21830E+00, 2.28010E+00] ###echo=oldecho
1,132
31.371429
68
py
UEDGE
UEDGE-master/test/Forthon_cases/Forthon_case4/rd_forthon_case4.py
###character*6 probname="itfa40" #Case of full toroidal equilibrium for ITER-FEAT # ####package flx;package grd;package bbb # Initialize pyuedge from uedge import * bbb.mhdgeo=1 bbb.isfixlb=0 bbb.isfixrb=0 os.system('rm -f aeqdsk neqdsk') os.system('cp aeq_iter-feat aeqdsk') os.system('cp geq_iter-feat neqdsk') ###character*9 machine="bbb.iter-feat" # Set the geometry bbb.ngrid = 1 grd.kxmesh = 1 #=4 for exponential grid in leg regions grd.dxgas[0] = 1.2e-03 grd.dxgas[1] = 1.2e-03 grd.nxgas[0] = 11 grd.nxgas[1] = 11 grd.alfx[0] = .64 grd.alfx[1] = .64 com.nxleg[0,0]=17 com.nxleg[0,1]=17 com.nxcore[0,0]=14 com.nxcore[0,1]=14 com.nycore[0]=10 com.nysol[0]=16 flx.psi0min1 = 0.95 #core minimum psi flx.psi0min2 = 0.992 #private flux minimum psi flx.psi0max = 1.035 #maximum flux at wall flx.alfcy = 2.0 #nonuniformity factor for radial mesh grd.slpxt=1.2 # Mesh construction--non orthogonal mesh com.ismmon=3 grd.istream=0 grd.iplate=1 grd.nsmooth=3 grd.wtmesh1=0.75 grd.dmix0=1.0 execfile('plate.iter-feat.py') com.isnonog=1 # non orthogonal differencing # Boundary conditions bbb.isnicore[0] = 1 #=1uses ncore for density BC bbb.isngcore[0]=2 # use ionization scale length for gas bbb.ncore[0] = 6.0e19 #value of core density if isnicore=1 bbb.curcore = 0. #core particle current if isnicore=0 bbb.iflcore = 1 #specify core power bbb.pcoree = 5.0e7 #electron power across core bbb.pcorei = 5.0e7 #ion power across core bbb.recycp=1.0 bbb.istepfc=3;bbb.istipfc=3 #priv. flux has fixed temperature scale length. bbb.istewc=3;bbb.istiwc=3 #wall has fixed temperature scale length. bbb.isnwcono=3;bbb.isnwconi=3 #walls have fixed density scale length bbb.lyni=0.05;bbb.lyte=0.05;bbb.lyti=0.05 # set walls into 7 zones bbb.nwsor=7 bbb.xgaso[0:7]=[1.533E-01, 4.599E-01, 3.506E+00, 9.291E+00, 1.508E+01, 1.817E+01, 1.858E+01] bbb.wgaso[0:7]=[3.066E-01, 3.066E-01, 5.790E+00, 5.790E+00, 5.790E+00, 5.000E-01, 4.089E-01] bbb.albdso[0:7]=1. bbb.matwso[0:7]=1. bbb.xgasi[0:7]=[1.213E-01, 3.638E-01, 6.063E-01, 9.024E-01, 1.246E+00, 1.541E+00, 1.837E+00] bbb.wgasi[0:7]=[2.425E-01, 2.425E-01, 2.425E-01, 3.705E-01, 2.955E-01, 2.955E-01, 2.955E-01] bbb.albdsi[0:7]=[0.98,0.98,1.0,1.0,1.0,0.98,0.98] bbb.matwsi[0:7]=1 bbb.recycw[0]=1.0 bbb.bcee = 5.; bbb.bcei = 3.5 #energy transmission coeffs. bbb.isupss = 1 #parallel vel can be supersonic # Transport coefficients bbb.difni[0] = 0.3 bbb.kye = 1.; bbb.kyi = 1. bbb.travis[0]=1.;bbb.parvis[0]=1. # Flux limits bbb.flalfe=0.21;bbb.flalfi=0.21;bbb.flalfgx=1.;bbb.flalfgy=1.;bbb.flalfgxy=1.;bbb.flalfv=0.5 bbb.lgmax=0.05 bbb.isplflxl=0 # Finite difference algorithms bbb.methe=33;bbb.methu=33;bbb.methg=66 bbb.methn=33;bbb.methi=33 # Solver package bbb.svrpkg="nksol" bbb.mfnksol = 3 bbb.epscon1 = .005 bbb.ftol = 1.e-8 bbb.iscolnorm = 3 # set to 3 for nksol bbb.premeth="ilut" bbb.lfililut = 100 bbb.lenpfac=75 bbb.runtim=1.e-7 bbb.rlx=0.9 ###bbb.del=1.e-8 # Neutral gas properties bbb.tfcx=5.;bbb.tfcy=5. #Franck-Condon temperatures bbb.eion = 5. #F-C energy to each born ion bbb.ediss = 10. #diss. energy from elec. (ediss=2*eion) bbb.isrecmon = 1 #e-i recombination (=1 is on) bbb.ngbackg=1.e12 # minimum floor neutral density bbb.ingb=4 # parameter used to force floor density # Inertial neutral model bbb.isupgon[0]=1;bbb.isngon[0]=0;com.ngsp=1;com.nhsp=2;bbb.ziin[com.nhsp-1]=0 bbb.cngmom=0;bbb.cmwall=0;bbb.cngtgx=0;bbb.cngtgy=0;bbb.cngflox=0;bbb.cngfloy=0;bbb.cfbgt=0 bbb.kxn=0;bbb.kyn=0 bbb.flalftgx=10.0;bbb.flalftgy=10.0 # Currents and potential parameters bbb.isphion=0 bbb.rsigpl=1.e-8 #anomalous cross-field conductivity bbb.cfjhf=0. #turn-on heat flow from current (fqp) bbb.jhswitch=0 #Joule Heating switch # Hydrogenic ions bbb.minu[0:2] = 2.5 # Atomic physics packages aph.isaphdir = 0 #=0 specifies atomic data file is in run directory com.istabon=10 #Stotler's rates for istabon=10 # turn on impurities bbb.isimpon=2 #=2 for fixed fraction model bbb.allocate() # set impurity concentration bbb.afracs = 0.03 # Restart from a save file bbb.restart = 1 restore ('pfiter_ffC.5') # Run this case bbb.exmain() # Print out a few variables across outer midplane print'' print'*** Printing variables versus radial index at outer midplane' print'' print '************\nradial position relative to separatrix [m]' print(com.yyc) print '************\n ion density, ni [m**-3] = ' print(bbb.ni[bbb.ixmp,]) print '************\n parallel ion velocity, up [m/s] = ' print(bbb.up[bbb.ixmp,]) print '************\n electron temp, te [eV] = ' print(bbb.te[bbb.ixmp,]/bbb.ev) print '************\n ion temp, ti [eV] = ' print(bbb.ti[bbb.ixmp,]/bbb.ev) print '************\n gas density, ng [m**-3] = ' print(bbb.ng[bbb.ixmp,])
4,740
28.08589
92
py
UEDGE
UEDGE-master/test/facets/example.py
# This is a simple example exercising the methods in uefacets.py from uefacets import * from Numeric import * print "Creating UEDGE instance" ue = Uedge() print "Reading input file 'testin.py' which is in the working directory" print "(Note: alternatively give the full path)" ue.ReadParams("testin.py") print "Set up MPI. This does nothing for now since UEDGE is running serial" ue.SetComm() print "Set some boundary conditions different from those in testin.py" ni = array([3.e19]) ng = array([1.e15]) ti = array([100.]) ui = array([0.]) print "Specifically, ni = ",ni[0], ", ng = ",ng[0],"," print " Te=Ti=", ti[0], ", ui = ",ui[0] ue.SetData(ni,ui,ng,ti,ti,"val","val","val","val") print "Take a step to time t=0.001" niflux,uiflux,ngflux,tiflux,teflux = ue.Advance(.001) print "The output fluxes are:" print "niflux = ",niflux print "uiflux = ",uiflux print "ngflux = ",ngflux print "tiflux = ",tiflux print "teflux = ",teflux print "Let's not accept this result and reset" ue.Reset() print "Change n_i to 2e19 and retake step" ni = array([2.e19]) ue.SetData(ni,ui,ng,ti,ti,"val","val","val","val") niflux,uiflux,ngflux,tiflux,teflux = ue.Advance(.001) print "The output fluxes are:" print "niflux = ",niflux print "uiflux = ",uiflux print "ngflux = ",ngflux print "tiflux = ",tiflux print "teflux = ",teflux print "create dumpfiles dump.pdb and dump.hdf" print "Note this requires that data to be dumped has been flagged in dot v files" ue.Dump("dump") print "Illustrating use of 'DoCommand' to print the sum of tiflux and teflux" ue.DoCommand("print 'tiflux+teflux = ',tiflux+teflux")
1,598
31.632653
81
py
UEDGE
UEDGE-master/test/facets/testin.py
#Coarse mesh (nx=16, ny=8) for DIII-D MHD equilibrium #Uses diffusive neutrals, so five variables (ni,upi,Te,Ti,ng) # ##package flx;package grd;package bbb # Initialize pyuedge from uedge import * # Set the geometry bbb.mhdgeo = 1 #use MHD equilibrium os.system('rm -f aeqdsk neqdsk') #change names of MHD eqil. files os.system('cp aeqdskd3d aeqdsk') # (Cannot tab or indent these 3 lines) os.system('cp neqdskd3d neqdsk') flx.psi0min1 = 0.98 #normalized flux on core bndry flx.psi0min2 = 0.98 #normalized flux on pf bndry flx.psi0sep = 1.00001 #normalized flux at separatrix flx.psi0max = 1.07 #normalized flux on outer wall bndry bbb.ngrid = 1 #number of meshes (always set to 1) com.nxleg[0,0] = 4 #pol. mesh pts from inner plate to x-point com.nxcore[0,0] = 4 #pol. mesh pts from x-point to top on inside com.nxcore[0,1] = 4 #pol. mesh pts from top to x-point on outside com.nxleg[0,1] = 4 #pol. mesh pts from x-point to outer plate com.nysol[0] = 6 #rad. mesh pts in SOL com.nycore[0] =2 #rad. mesh pts in core # Finite-difference algorithms (upwind, central diff, etc.) bbb.methn = 33 #ion continuty eqn bbb.methu = 33 #ion parallel momentum eqn bbb.methe = 33 #electron energy eqn bbb.methi = 33 #ion energy eqn bbb.methg = 33 #neutral gas continuity eqn # Boundary conditions bbb.ncore[0] = 2.5e19 #hydrogen ion density on core ## iflcore = 0 #flag; =0, fixed Te,i; =1, fixed power on core bbb.tcoree = 100. #core Te bbb.tcorei = 100. #core Ti bbb.tedge = 2. #fixed wall,pf Te,i if istewcon=1, etc bbb.recycp[0] = 0.8 #hydrogen recycling coeff at plates # Transport coefficients bbb.difni[0] = 1. #D for radial hydrogen diffusion bbb.kye = 1. #chi_e for radial elec energy diffusion bbb.kyi = 1. #chi_i for radial ion energy diffusion bbb.travis[0] = 1. #eta_a for radial ion momentum diffusion # Flux limits bbb.flalfe = 0.21 #electron parallel thermal conduct. coeff bbb.flalfi = 0.21 #ion parallel thermal conduct. coeff bbb.flalfv = 1. #ion parallel viscosity coeff bbb.flalfgx = 1.e20 #neut. gas in poloidal direction bbb.flalfgy = 1.e20 #neut. gas in radial direction # Solver package bbb.svrpkg = "nksol" #Newton solver using Krylov method bbb.premeth = "banded" #Solution method for precond. Jacobian matrix # Restart from a pfb savefile bbb.restart = 1 #Begin from savefile, not estimated profiles bbb.allocate() #allocate space for savevariables restore('pfd3d_ex.16x8') #read in the solution from pfb file ###os.system('ln -s ~/Uedge/uedge/in/aph aph6') com.istabon = 10 aph.isaphdir = 0 #=0 specifies atomic data file is in run directory # Execute uedge #bbb.exmain() """ # Print out a few variables across outer midplane print'' print'*** Printing variables versus radial index at outer midplane' print'' print '************\nradial position relative to separatrix [m]' print(com.yyc) print '************\n ion density, ni [m**-3] = ' print(bbb.ni[bbb.ixmp,]) print '************\n parallel ion velocity, up [m/s] = ' print(bbb.up[bbb.ixmp,]) print '************\n electron temp, te [eV] = ' print(bbb.te[bbb.ixmp,]/bbb.ev) print '************\n ion temp, ti [eV] = ' print(bbb.ti[bbb.ixmp,]/bbb.ev) """
3,181
35.574713
74
py
UEDGE
UEDGE-master/pytests/test_template.py
import unittest import os import numpy as np import sys,getopt def prhelp(): print(""" Usage: python test_slab.py [-r|--ref] [-h] or pytest --forked test_slab.py -r|--ref to produce reference files for future test runs """) # # Note the numbering of the tests. Purists will say that # tests should be order independent. Because Uedge variables # are in shared objects it is naturally stateful. Because of # this stateful feature the order needs to be controlled. # class TestRun(unittest.TestCase): def setUp(self): """ This is run pre-test. """ def test_one(self): """ Test that uedge will import """ global ftol try: import uedge as ue assert True except: assert False if __name__ == '__main__': try: opts, args = getopt.getopt(sys.argv[1:],'hr',['ref']) except getopt.GetoptError: prhelp() sys.exit(2) for opt,arg in opts: if opt in ('-h'): prhelp() sys.exit(2) elif opt in ('-r','--ref'): """ Do something that will store reference data """ sys.exit(2) unittest.main()
1,282
18.439394
72
py
UEDGE
UEDGE-master/pytests/testscripts/rdinitdt.py
# Setup file to run time-dependently using dtreal # Change dtreal for starting dt and savefname to change pfb file name # Once variables are set, read rdrundt to execute a time-dependent run # IMPORT UEDGE (needed to define group name bbb) from uedge import * i_stor = 0 nfe_tot = 0 savefn = "savedt.hdf5" # name of hdf5 savefile written every timestep bbb.rdtphidtr = 1e20 # ratio dtphi/dtreal bbb.ismfnkauto = 1 # if =1, mfnksol=3 for dtreal<dtmfnk3, otherwise=-3 bbb.dtmfnk3 = 5.e-4 # dtreal for mfnksol sign change if ismfnkauto=1 bbb.mult_dt = 3.4 # factor expanding dtreal after ii2max steps bbb.ii1max = 500 # number of changes to dtreal bbb.ii2max = 5 # number of timesteps at current dtreal bbb.itermxrdc = 7 # value of itermx used by rdcontdt bbb.incpset = 7 # iterations until Jacobian is recomputed bbb.ftol_dt = 1.e-5 # fnrm tolerance for the time-dependent steps bbb.ftol_min = 1e-9 # value of fnrm where time advance will stop bbb.dt_tot = 0. # tot time accumulated for run (output, not input) bbb.t_stop = 100. # value of dt_tot (sec) where calculation will stop bbb.dt_max = 100. # maximum time step for dtreal bbb.dt_kill = 1e-14 # min allowed time step; rdcontdt stops if reached bbb.deldt_min = 0.04 # minimum relative change allowed for model_dt > 0 bbb.initjac = 0 # if=1, calc initial Jac upon reading rdcontdt bbb.numrevjmax = 2 # number of dt reductions before Jac recalculated bbb.numfwdjmax = 1 # number of dt increases before Jac recalculated ###bbb.ismmaxuc = 1 # =1 for intern calc mmaxu; =0,set mmaxu & dont chng bbb.irev = -1 # flag to allow reduced dt advance after cutback bbb.rlx = 0.9 # max. change in variable at each linear iteration bbb.itermx = 7 # max. number of linear iterations allowed bbb.tstor_s = 1e-5 # beginning time for storing solution bbb.tstor_e = 1e-3 # ending time for storing solution bbb.n_stor = 0 # number of linearly spaced storage points bbb.ipt = 1 # index of variable; value printed at step # if ipt not reset from unity, ipt=idxte(nx,iysptrx+1)
2,100
51.525
76
py
UEDGE
UEDGE-master/pytests/testscripts/uetests.py
import unittest import os import numpy as np import sys,getopt import uuid from importlib import import_module as im from multiprocessing import Process def saverefs(filename): import uedge as ue import uedge.hdf5 as h5 fnrm = ue.bbb.get_fnrm(ue.bbb.dtreal) nfe = ue.bbb.nfe[0][0] sclyl = ue.bbb.yldot * ue.bbb.sfscal nodeid = uuid.getnode() h5.hdf5_dump(filename,vars=['nodeid','fnrm','nfe','sclyl'],globals=locals()) def prhelp(): print(""" Usage: python test_slab.py [-r|--ref] [-h|--help] or pytest --forked test_slab.py -r|--ref to produce reference files for future test runs """) def startup(case,ftol=1.e-9): """ startup("<case>",ftol=1.e-9) Reconverge specified case. Done for every test. """ import uedge as ue im(case) ue.bbb.ftol = ftol ue.bbb.exmain() def perturb(case,ftol=1.e-9): """ perturn("<case>",ftol=1.e-9) Perturb solution and Reconverge specified case. """ import uedge as ue ue.bbb.ngs = ue.bbb.ngs*1.02 ue.bbb.nis = ue.bbb.nis*1.02 ue.bbb.phis = ue.bbb.phis*1.02 ue.bbb.tes = ue.bbb.tes*1.02 ue.bbb.tis = ue.bbb.tis*1.02 ue.bbb.ups = ue.bbb.ups*1.02 ue.bbb.ftol = ftol def steadystate(): """ steadystate() Evolve current case to steady state with rdcontdt. """ import uedge as ue import rdinitdt import rdcontdt def makeperturb(filename,case,ftol=1.e-9): import uedge as ue startup(case,ftol=ftol) perturb(case,ftol=ftol) steadystate() saverefs(filename) def makeref(filename,case,ftol=1.e-9): """ Produce and save reference data for comparison in future test runs. """ import uedge as ue startup(case,ftol=ftol) saverefs(filename) identical=1.e-10 close=0.5 def check_fnorm(name,filename,case,doassert=True): import uedge as ue import uedge.hdf5 as h5 ref={} h5.hdf5_restore_dump(filename,scope=ref) fnrm = ue.bbb.get_fnrm(ue.bbb.dtreal) nfe = ue.bbb.nfe[0][0] if np.isclose(fnrm,ref['fnrm'],atol=0.0,rtol=identical): # # If it makes it here then fnorm is basically identical. # Most likely running on the same os/compiler versions # as when the reference files were produced. # if doassert: print() print(name,' fnorm identical.') assert True else: return True elif fnrm > ue.bbb.ftol: if doassert: print('Relative change in Fnorm too large. Threshold is ',close) print('fnrm: ',fnrm,' ref: ',ref['fnrm']) print(' rel change: ',np.abs(fnrm - ref['fnrm'])/np.abs(ref['fnrm'])) print(' abs change: ',np.abs(fnrm - ref['fnrm'])) print() sclyl = ue.bbb.yldot * ue.bbb.sfscal rsclyl = ref['sclyl'] iva = np.abs(rsclyl - sclyl) / (np.abs(rsclyl) + np.abs(sclyl) + 1e-20) ind = np.where(iva == np.max(iva)) iv = ind[0][0] (ix,iy) = ue.bbb.igyl[iv,0:2] loc_troub_eqn = np.mod(iv,ue.bbb.numvar)+1 numvar = ue.bbb.numvar if doassert: print("** Number of variables is:") print("numvar = ", numvar) print(" ") print("** Troublemaker equation is:") print("loc_troub_eqn = ",loc_troub_eqn) print(" ") print("** Troublemaker cell (ix,iy) is:") print(ue.bbb.igyl[iv,:]) print(" ") print("** Timestep for troublemaker equation:") print(ue.bbb.dtuse[iv]) print(" ") print("** yl for troublemaker equation:") print(ue.bbb.yl[iv]) print(" ") assert False else: return False else: if doassert: assert True else: return True def check_nfe(name,filename,case,doassert=None): import uedge as ue import uedge.hdf5 as h5 ref={} h5.hdf5_restore_dump(filename,scope=ref) fnrm = ue.bbb.get_fnrm(ue.bbb.dtreal) nfe = ue.bbb.nfe[0][0] if np.isclose(fnrm,ref['fnrm'],atol=0.0,rtol=identical): # # If it makes it here then fnorm is basically identical. # Most likely running on the same os/compiler versions # as when the reference files were produced. # if doassert: print() print(name,' fnorm identical.') assert True else: return True elif not np.isclose(nfe,ref['nfe'],atol=0.0,rtol=0.02): if doassert: print("The number of Krylov iterations for a 2% perturbation is ",nfe/ref['nfe']," times the ref case") assert False else: return False else: if doassert: assert True else: return True
4,934
24.050761
114
py
UEDGE
UEDGE-master/pytests/testscripts/rdcontdt.py
# This file runs a time-dependent case using dtreal. First, obtain a converged # solution for a (usually small) dtreal; xuedge must report iterm=1 at the end. # Then adjust control parameters in rdinitdt; read this file, which reads rdinitdt. # If a mistake is made, to restart this file without a Jacobian evaluation, # be sure to reset iterm=1 (=> last step was successful) # IMPORT UEDGE (assuming starting from ipython before any imports) from uedge import * from numpy import zeros,sqrt # IMPORT HDF5 routines for saving solutions below from uedge.hdf5 import * i_stor = 0 nfe_tot = 0 savefn = "savedt.hdf5" # name of hdf5 savefile written every timestep no = 0;yes = 1 echo = no # Set precisions of floating point output ###import print_options ###print_options.set_float_precision(4) nx=com.nx;ny=com.ny;nisp=com.nisp;ngsp=com.ngsp;numvar=bbb.numvar isteon=bbb.isteon if (i_stor==0): ni_stor = zeros((bbb.n_stor,nx+1+1,ny+1+1,nisp),"d") # set time storage arrays up_stor = zeros((bbb.n_stor,nx+1+1,ny+1+1,nisp),"d") te_stor = zeros((bbb.n_stor,nx+1+1,ny+1+1),"d") ti_stor = zeros((bbb.n_stor,nx+1+1,ny+1+1),"d") ng_stor = zeros((bbb.n_stor,nx+1+1,ny+1+1,ngsp),"d") phi_stor = zeros((bbb.n_stor,nx+1+1,ny+1+1),"d") tim_stor = zeros((bbb.n_stor),"d") dtreal_stor = zeros((bbb.n_stor),"d") nfe_stor = zeros((bbb.n_stor),"l") dt_stor = (bbb.tstor_e - bbb.tstor_s)/(bbb.n_stor - 1) i_stor = max(i_stor,1) # set counter for storage arrays bbb.dt_tot = max(bbb.dt_tot,0.) nfe_tot = max(nfe_tot,0) deldt_0 = bbb.deldt isdtsf_sav = bbb.isdtsfscal if (bbb.ipt==1 and bbb.isteon==1): # set ipt to te(nx,iysptrx+1) if no user value ipt = bbb.idxte[nx-1,com.iysptrx] #note: ipt is local, bbb.ipt global bbb.irev = -1 # forces second branch of irev in ii1 loop below if (bbb.iterm == 1): # successful initial run with dtreal bbb.dtreal = bbb.dtreal/bbb.mult_dt # gives same dtreal after irev loop else: # unsuccessful initial run; reduce dtreal bbb.dtreal = bbb.dtreal/(3*bbb.mult_dt) # causes dt=dt/mult_dt after irev loop if (bbb.initjac == 0): bbb.newgeo=0 dtreal_sav = bbb.dtreal bbb.itermx = bbb.itermxrdc bbb.dtreal = bbb.dtreal/bbb.mult_dt #adjust for mult. to follow; mult_dt in rdinitdt bbb.dtphi = bbb.rdtphidtr*bbb.dtreal neq=bbb.neq svrpkg=bbb.svrpkg.tostring().strip() # bbb.ylodt = bbb.yl bbb.pandf1 (-1, -1, 0, bbb.neq, 1., bbb.yl, bbb.yldot) fnrm_old = sqrt(sum((bbb.yldot[0:neq]*bbb.sfscal[0:neq])**2)) if (bbb.initjac == 1): fnrm_old=1.e20 print(( "initial fnrm =",fnrm_old)) for ii1 in range( 1, bbb.ii1max+1): if (bbb.ismfnkauto==1): bbb.mfnksol = 3 # adjust the time-step if (bbb.irev == 0): # Only used after a dt reduc. success. completes loop ii2 for fixed dt bbb.dtreal = min(3*bbb.dtreal,bbb.t_stop) #first move forward after reduction bbb.dtphi = bbb.rdtphidtr*bbb.dtreal if (bbb.ismfnkauto==1 and bbb.dtreal > bbb.dtmfnk3): bbb.mfnksol = -3 bbb.deldt = 3*bbb.deldt else: # either increase or decrease dtreal; depends on mult_dt bbb.dtreal = min(bbb.mult_dt*bbb.dtreal,bbb.t_stop) bbb.dtphi = bbb.rdtphidtr*bbb.dtreal if (bbb.ismfnkauto==1 and bbb.dtreal > bbb.dtmfnk3): bbb.mfnksol = -3 bbb.deldt = bbb.mult_dt*bbb.deldt bbb.dtreal = min(bbb.dtreal,bbb.dt_max) bbb.dtphi = bbb.rdtphidtr*bbb.dtreal if (bbb.ismfnkauto==1 and bbb.dtreal > bbb.dtmfnk3): bbb.mfnksol = -3 bbb.deldt = min(bbb.deldt,deldt_0) bbb.deldt = max(bbb.deldt,bbb.deldt_min) nsteps_nk=1 print('--------------------------------------------------------------------') print('--------------------------------------------------------------------') print(' ') print(('*** Number time-step changes = ',ii1,' New time-step = ', bbb.dtreal)) print('--------------------------------------------------------------------') bbb.itermx = bbb.itermxrdc if (ii1>1 or bbb.initjac==1): # first time calc Jac if initjac=1 if (bbb.irev == 1): # decrease in bbb.dtreal if (bbb.numrev < bbb.numrevjmax and \ bbb.numrfcum < bbb.numrevjmax+bbb.numfwdjmax): #dont recom bbb.jac bbb.icntnunk = 1 bbb.numrfcum = bbb.numrfcum + 1 else: # force bbb.jac calc, reset numrev bbb.icntnunk = 0 bbb.numrev = -1 # yields api.zero in next statement bbb.numrfcum = 0 bbb.numrev = bbb.numrev + 1 bbb.numfwd = 0 else: # increase in bbb.dtreal if (bbb.numfwd < bbb.numfwdjmax and \ bbb.numrfcum < bbb.numrevjmax+bbb.numfwdjmax): #dont recomp bbb.jac bbb.icntnunk = 1 bbb.numrfcum = bbb.numrfcum + 1 else: bbb.icntnunk = 0 #recompute jacobian for increase dt bbb.numfwd = -1 bbb.numrfcum = 0 bbb.numfwd = bbb.numfwd + 1 bbb.numrev = 0 #bbb.restart counter for dt reversals bbb.isdtsfscal = isdtsf_sav bbb.ftol = min(bbb.ftol_dt, 0.01*fnrm_old) bbb.exmain() # take a single step at the present bbb.dtreal if (bbb.iterm == 1): bbb.dt_tot = bbb.dt_tot + bbb.dtreal nfe_tot = nfe_tot + bbb.nfe[0,0] bbb.ylodt = bbb.yl bbb.pandf1 (-1, -1, 0, bbb.neq, 1., bbb.yl, bbb.yldot) fnrm_old = sqrt(sum((bbb.yldot[0:neq-1]*bbb.sfscal[0:neq-1])**2)) if (bbb.dt_tot>=0.9999999*bbb.t_stop or fnrm_old<bbb.ftol_min): print(' ') print('*****************************************************') print('** SUCCESS: frnm < bbb.ftol; or dt_tot >= t_stop **') print('*****************************************************') break bbb.icntnunk = 1 bbb.isdtsfscal = 0 for ii2 in range( 1, bbb.ii2max+1): #take ii2max steps at the present time-step if (bbb.iterm == 1): bbb.itermx = bbb.itermxrdc bbb.ftol = min(bbb.ftol_dt, 0.01*fnrm_old) bbb.exmain() if (bbb.iterm == 1): bbb.ylodt = bbb.yl bbb.pandf1 (-1, -1, 0, bbb.neq, 1., bbb.yl, bbb.yldot) fnrm_old = sqrt(sum((bbb.yldot[0:neq-1]*bbb.sfscal[0:neq-1])**2)) print("Total time = ",bbb.dt_tot,"; Timestep = ",bbb.dtreal) print("variable index ipt = ",ipt, " bbb.yl[ipt] = ",bbb.yl[ipt]) dtreal_sav = bbb.dtreal bbb.dt_tot = bbb.dt_tot + bbb.dtreal nfe_tot = nfe_tot + bbb.nfe[0,0] if (bbb.dt_tot>=0.999999999999*bbb.t_stop or fnrm_old<bbb.ftol_min): break ## Store variables if a storage time has been crossed if (bbb.dt_tot >= dt_stor*i_stor and i_stor<=bbb.n_stor): i_stor1 = i_stor-1 ni_stor[i_stor1,:,:,:] = ni up_stor[i_stor1,:,:,:] = up te_stor[i_stor1,:,:] = te ti_stor1[i_stor1,:,:] = ti ng_stor[i_stor1,:,:,:] = ng phi_stor1[i_stor1,:,:] = phi tim_stor[i_stor1] = bbb.dt_tot nfe_stor[i_stor1] = nfe_tot dtreal_stor[i_stor1] = bbb.dtreal i_stor = i_stor + 1 ## End of storage section if (bbb.dt_tot>=bbb.t_stop or fnrm_old<bbb.ftol_min): break # need for both loops bbb.irev = bbb.irev-1 if (bbb.iterm != 1): #print bad eqn, cut dtreal by 3, set irev flag ####### a copy of idtroub script ######################## oldecho=echo echo=no # integer ii # real8 ydmax scalfac = bbb.sfscal if (svrpkg != "nksol"): scalfac = 1/(bbb.yl + 1.e-30) # for time-dep calc. ydmax = 0.999999999*max(abs(bbb.yldot*scalfac)) itrouble = 0 for ii in range(neq): if (abs(bbb.yldot[ii]*scalfac[ii]) > ydmax): itrouble=ii print("** Fortran index of trouble making equation is:") print(itrouble+1) break print("** Number of variables is:") print("numvar = ", numvar) print(" ") iv_t = (itrouble).__mod__(numvar) + 1 print("** Troublemaker equation is:") print("iv_t = ",iv_t) print(" ") print("** Troublemaker cell (ix,iy) is:") print(bbb.igyl[itrouble,]) print(" ") print("** Timestep for troublemaker equation:") print(bbb.dtuse[itrouble]) print(" ") print("** yl for troublemaker equation:") print(bbb.yl[itrouble]) print(" ") echo=oldecho ######## end of idtroub script ############################## if (bbb.dtreal < bbb.dt_kill): print(' ') print('*************************************') print('** FAILURE: time-step < dt_kill **') print('*************************************') break bbb.irev = 1 print('*** Converg. fails for bbb.dtreal; reduce time-step by 3, try again') print('----------------------------------------------------------------- ') bbb.dtreal = bbb.dtreal/(3*bbb.mult_dt) bbb.dtphi = bbb.rdtphidtr*bbb.dtreal if (bbb.ismfnkauto==1 and bbb.dtreal > bbb.dtmfnk3): bbb.mfnksol = -3 bbb.deldt = bbb.deldt/(3*bbb.mult_dt) bbb.iterm = 1 echo = yes
9,290
41.040724
88
py
UEDGE
UEDGE-master/pytests/fast_tests/d3d_snull_D_only/j.py
import uedge import rd_d3dHsm_in import uedge uedge.bbb.exmain() import uedge.uedgeplots as up up.plotmesh()
110
12.875
29
py
UEDGE
UEDGE-master/pytests/fast_tests/d3d_snull_D_only/rd_d3dHsm_in.py
# # ########################################################################### # DESCRIPTION OF PROBLEM (d3dHsm) from FACETS test suite: # DIII-D single-null geometry with 5 variables (ni,upi,te,ti,ng) and a # (16+2)*(8+2)=18x10 [poloidal*radial] mesh yielding 900 variables. # Solver used is Newton Krylov (svrpkg="nksol") and preconditioner uses a # direct banded solver for the LU decomposition (premeth="banded"). Iterates # to steady-state solution from an initial profile file (HF5). ########################################################################### import uedge from uedge import * # Set the geometry bbb.mhdgeo = 1 #=1 use MHD equilibrium files ##flx.aeqdskfname = "aeqdskd3d" #name of EFIT 'a' file for flux-surface mesh ##flx.geqdskfname = "neqdskd3d" #name of EFIT 'g' or 'n' file for flux-sur mesh flx.psi0min1 = 0.98 #normalized flux on core bndry flx.psi0min2 = 0.98 #normalized flux on pf bndry flx.psi0sep = 1.00001 #normalized flux at separatrix flx.psi0max = 1.07 #normalized flux on outer wall bndry bbb.ngrid = 1 #number of mesh sequenc. (always set to 1) com.nxleg[0,0] = 4 #pol. mesh pts from inner plate to x-point com.nxcore[0,0] = 4 #pol. mesh pts from x-point to top on inside com.nxcore[0,1] = 4 #pol. mesh pts from top to x-point on outside com.nxleg[0,1] = 4 #pol. mesh pts from x-point to outer plate com.nysol[0] = 6 #rad. mesh pts in SOL com.nycore[0] = 2 #rad. mesh pts in core # Finite-difference algorithms (upwind, central diff, etc.) bbb.methn = 33 #ion continuty eqn bbb.methu = 33 #ion parallel momentum eqn bbb.methe = 33 #electron energy eqn bbb.methi = 33 #ion energy eqn bbb.methg = 33 #neutral gas continuity eqn # Boundary conditions bbb.ncore[0] = 2.5e19 #hydrogen ion density on core ## iflcore = 0 #flag; =0, fixed Te,i; =1, fixed power on core bbb.tcoree = 100. #core Te bbb.tcorei = 100. #core Ti bbb.tedge = 2. #fixed wall,pf Te,i if istewcon=1, etc bbb.recycp[0] = 0.8 #hydrogen recycling coeff at plates # Transport coefficients (m**2/s) bbb.difni[0] = 1. #D for radial hydrogen diffusion bbb.kye = 1. #chi_e for radial elec energy diffusion bbb.kyi = 1. #chi_i for radial ion energy diffusion bbb.travis[0] = 1. #eta_a for radial ion momentum diffusion # Flux limits bbb.flalfe = 0.21 #electron parallel thermal conduct. coeff bbb.flalfi = 0.21 #ion parallel thermal conduct. coeff bbb.flalfv = 1. #ion parallel viscosity coeff bbb.flalfgx = 1.e20 #neut. gas in poloidal direction bbb.flalfgy = 1.e20 #neut. gas in radial direction # Solver package bbb.svrpkg = "nksol" #Newton solver using Krylov method bbb.premeth = "banded" #Solution method for precond. Jacobian matrix # Restart from a HDF5 or PDB savefile bbb.restart = 1 #Begin from savefile, not estimated profiles bbb.allocate() #allocates storage for arrays from uedge.hdf5 import * hdf5_restore("d3dHsm.h5") # Atomic data switches com.istabon = 10 #=10 specifics hydrogen data file ehr2.dat
2,997
41.828571
81
py
UEDGE
UEDGE-master/pytests/fast_tests/d3d_snull_D_only/test_d3d_snull_D_only.py
import unittest import os import numpy as np import sys,getopt from multiprocessing import Process thisfile=os.path.realpath(__file__) thispath=os.path.dirname(thisfile) sys.path.insert(0,thispath) sys.path.insert(0,os.path.dirname(os.path.dirname(thispath))+'/testscripts') import uetests as uet ftol = 1.e-9 name = 'd3d_snull_D_only' restart = 'rd_d3dHsm_in' def prhelp(): print(""" Usage: python test_slab.py [-r|--ref] [-h|--help] or pytest --forked test_slab.py -r|--ref to produce reference files for future test runs """) # # Note the numbering of the tests. Purists will say that # tests should be order independent. Because Uedge variables # are in shared objects it is naturally stateful. Because of # this stateful feature the order needs to be controlled. # class TestRun(unittest.TestCase): def setUp(self): """ This is run pre-test. Reads in restart data and re-converges solution. """ import uedge as ue os.chdir(thispath) def test_reconv(self): """ Test that initial re-converged solution fnrm is low. """ global ftol import uedge as ue uet.startup(restart,ftol=ftol) uet.check_fnorm(name+' Reconverge','ref_reconv.h5',restart,doassert=True) def test_perturb(self): """ Test that initial re-converged solution fnrm is low. """ global ftol import uedge as ue uet.startup(restart,ftol=ftol) uet.perturb(restart,ftol=ftol) uet.steadystate() uet.check_nfe(name+' Perturb','ref_perturb.h5',restart,doassert=True) if __name__ == '__main__': try: opts, args = getopt.getopt(sys.argv[1:],'hr',['help,ref']) except getopt.GetoptError: prhelp() sys.exit(2) for opt,arg in opts: if opt in ('-h','--help'): prhelp() sys.exit(2) elif opt in ('-r','--ref'): kargs = {'ftol':ftol} p1 = Process(target=uet.makeref,args=('ref_reconv.h5',restart),kwargs=kargs) p1.start() p2 = Process(target=uet.makeperturb,args=('ref_perturb.h5',restart),kwargs=kargs) p2.start() p1.join() p2.join() sys.exit(2) unittest.main()
2,339
25.292135
92
py
UEDGE
UEDGE-master/pytests/fast_tests/d3d_snull_D+C_orthog/test_d3d_snull_D+C_orthog.py
import unittest import os import numpy as np import sys,getopt from multiprocessing import Process thisfile=os.path.realpath(__file__) thispath=os.path.dirname(thisfile) sys.path.insert(0,thispath) sys.path.insert(0,os.path.dirname(os.path.dirname(thispath))+'/testscripts') import uetests as uet ftol = 1.e-9 name = 'd3d_snull_D+C_orthog' restart = 'rd_d3dHmsCog_strahl_Rp95' def prhelp(): print(""" Usage: python test_slab.py [-r|--ref] [-h|--help] or pytest --forked test_slab.py -r|--ref to produce reference files for future test runs """) # # Note the numbering of the tests. Purists will say that # tests should be order independent. Because Uedge variables # are in shared objects it is naturally stateful. Because of # this stateful feature the order needs to be controlled. # class TestRun(unittest.TestCase): def setUp(self): """ This is run pre-test. Reads in restart data and re-converges solution. """ import uedge as ue os.chdir(thispath) def test_reconv(self): """ Test that initial re-converged solution fnrm is low. """ global ftol import uedge as ue uet.startup(restart,ftol=ftol) uet.check_fnorm(name+' Reconverge','ref_reconv.h5',restart,doassert=True) def test_perturb(self): """ Test that initial re-converged solution fnrm is low. """ global ftol import uedge as ue uet.startup(restart,ftol=ftol) uet.perturb(restart,ftol=ftol) uet.steadystate() uet.check_nfe(name+' Perturb','ref_perturb.h5',restart,doassert=True) if __name__ == '__main__': try: opts, args = getopt.getopt(sys.argv[1:],'hr',['help,ref']) except getopt.GetoptError: prhelp() sys.exit(2) for opt,arg in opts: if opt in ('-h','--help'): prhelp() sys.exit(2) elif opt in ('-r','--ref'): kargs = {'ftol':ftol} p1 = Process(target=uet.makeref,args=('ref_reconv.h5',restart),kwargs=kargs) p1.start() p2 = Process(target=uet.makeperturb,args=('ref_perturb.h5',restart),kwargs=kargs) p2.start() p1.join() p2.join() sys.exit(2) unittest.main()
2,355
25.47191
92
py
UEDGE
UEDGE-master/pytests/fast_tests/d3d_snull_D+C_orthog/rd_d3dHmsCog_strahl_Rp95.py
# # ########################################################################### # DESCRIPTION OF PROBLEM (d3dHmsCnog) using Strahl imp data; from FACETS : # DIII-D single-null geometry with 6 hydrogen variables (ni,ng,upi,upg,te,ti) # and 7 carbon variables (six charge-state densities ni and one ng) on a # (16+2)*(8+2)=18x10 [poloidal*radial] mesh yielding 2340 variables. # Solver used is Newton Krylov (svrpkg="nksol") and preconditioner uses an # iterative solver ILUT for Jacobian LU decomposition. Also includes tilted # divertor plates wrt flux-surface normal, thus testing the nonorthogonal # finite-volume difference stencil. Iterates to steady-state solution from # an initial profile file (HF5). ########################################################################### # Import uedge into python and make variables active import uedge from uedge import * # Begin uedge parameter input # Set the geometry bbb.mhdgeo = 1 #=1 use MHD equilibrium files ##flx.aeqdskfname = "a110465.03500" #EFIT "a" file for flux-surface mesh ##flx.geqdskfname = "g110465.03500" #EFIT "g" or "n" file for flux-sur mesh flx.psi0min1 = 0.96 #normalized flux on core bndry flx.psi0min2 = 0.98 #normalized flux on pf bndry flx.psi0sep = 1.00001 #normalized flux at separatrix flx.psi0max = 1.07 #normalized flux on outer wall bndry bbb.ngrid = 1 #number of mesh sequenc. (always set to 1) bbb.recycm = 0.1 com.nxleg[0,0] = 4 #pol. mesh pts from inner plate to x-point com.nxcore[0,0] = 4 #pol. mesh pts from x-point to top on inside com.nxcore[0,1] = 4 #pol. mesh pts from top to x-point on outside com.nxleg[0,1] = 4 #pol. mesh pts from x-point to outer plate com.nysol[0] = 4 #rad. mesh pts in SOL com.nycore[0] = 4 #rad. mesh pts in core flx.alfcy = 2 #>0 concentrates y-mesh near separatrix # Mesh construction--non orthogonal mesh ##com.ismmon = 3 #controls transition from nonorthog to orthog mesh ##com.isnonog = 1 # non orthogonal differencing ##grd.istream = 0 ##grd.iplate = 1 ##grd.nsmooth = 3 ##grd.wtmesh1 = 0.75 ##grd.dmix0 = 1.0 # Set params for line-segments defining inner(1) and outer(2) plots ##grd.nplate1 = 2 ##grd.nplate2 = 2 # Finite-difference algorithms (upwind, central diff, etc.) bbb.methn = 33 #ion continuty eqn bbb.methu = 33 #ion parallel momentum eqn bbb.methe = 33 #electron energy eqn bbb.methi = 33 #ion energy eqn ##bbb.methg = 66 bbb.methg = 33 # Boundary conditions bbb.ncore[0] = 2.e19 #hydrogen ion density on core bbb.iflcore = 0 #flag; =0, fixed Te,i; =1, fixed power on core bbb.tcoree = 600. #core Te bbb.tcorei = 600. #core Ti bbb.tedge = 2. #fixed wall,pf Te,i if istewcon=1, etc bbb.recycp[0] = 0.95 #hydrogen recycling coeff at plates # Transport coefficients (m**2/s) bbb.difni[0] = 1. #D for radial hydrogen diffusion bbb.kye = 1. #chi_e for radial elec energy diffusion bbb.kyi = 1. #chi_i for radial ion energy diffusion bbb.travis[0] = 1. #eta_a for radial ion momentum diffusion # Flux limits bbb.flalfe = 0.21 #electron parallel thermal conduct. coeff bbb.flalfi = 0.21 #ion parallel thermal conduct. coeff bbb.flalfv = 0.5 #ion parallel viscosity coeff bbb.flalfgx = 1. #neut. gas part. flux in poloidal direction bbb.flalfgy = 1. #neut. gas part. flux in radial direction bbb.flalfgxy = 1. #neut. gas part. flux in mixed derivatives bbb.flalftgx = 1. #neut. gas thermal flux, poloidal direction bbb.flalftgy = 1. #neut. gas thermal flux, radial direction bbb.lgmax = 0.1 #max scale length for flalfgx,y bbb.lgtmax = 0.1 #max scale length for flalftgx,y bbb.lgvmax = 0.1 #max scale length for flalfvgx,y # Solver package bbb.svrpkg = "nksol" #Newton solver using Krylov method bbb.premeth = "ilut" #Solution method for precond. Jacobian matrix # Parallel neutral momentum equation bbb.isupgon[0] = 1 bbb.ineudif = 2 #=2 for evolving pg=ng*tg variable bbb.isngon[0] = 0 com.ngsp = 1 com.nhsp = 2 bbb.ziin[1] = 0 ## Impurity gas basics com.ngsp = 2 #total number of gas species bbb.isngon[1] = 1 #turns on impurity gas bbb.ngbackg[1] = 1.e9 #neutral impurity background for added source bbb.ingb = 2 #exponent for strength of ngbackg turn-on bbb.istgcon[1] = 1 #=1 for constant tg(2) at tgas(2) bbb.tgas[1] = 1. #value for tg when istgcon=1 bbb.rcxighg = 0. # best value; ratio of imp cx to hyd cx bbb.kelighi[1] = 5.e-16 #elastic sig_v for imp_gas/h_ion bbb.kelighg[1] = 5.e-16 #elastic sig_v for imp_gas/h_gas bbb.n0g[1] = 1.e16 #imp. gas density normalization # Impurity gas boundary conditions bbb.recycp[1] = 0.01 #plate recycling of impurities bbb.recycw[1] = 1e-4 #wall recycling; matwsi,o set above for hyd bbb.isch_sput[1]=7 # Haasz/Davis chemical sputtering model bbb.isph_sput[1]=3 # physical sputtering model bbb.t_wall = 300. bbb.t_plat = 500. bbb.crmb = 2. ## Impurity ions bbb.isimpon = 6 #Use force-balance only com.nzsp[0] = 6 #number chrg states impurity isotope #1 bbb.csfaclb[2,0] = 2.191 bbb.csfaclb[3,0] = 2.191 bbb.csfaclb[4,0] = 2.191 bbb.csfaclb[5,0] = 2.191 bbb.csfaclb[6,0] = 2.191 bbb.csfaclb[7,0] = 2.191 bbb.csfacrb[2,0] = 2.191 bbb.csfacrb[3,0] = 2.191 bbb.csfacrb[4,0] = 2.191 bbb.csfacrb[5,0] = 2.191 bbb.csfacrb[6,0] = 2.191 bbb.csfacrb[7,0] = 2.191 bbb.csfaclb[2,1] = 2.191 bbb.csfaclb[3,1] = 2.191 bbb.csfaclb[4,1] = 2.191 bbb.csfaclb[5,1] = 2.191 bbb.csfaclb[6,1] = 2.191 bbb.csfaclb[7,1] = 2.191 bbb.csfacrb[2,1] = 2.191 bbb.csfacrb[3,1] = 2.191 bbb.csfacrb[4,1] = 2.191 bbb.csfacrb[5,1] = 2.191 bbb.csfacrb[6,1] = 2.191 bbb.csfacrb[7,1] = 2.191 bbb.minu[2] = 12. bbb.minu[3] = 12. bbb.minu[4] = 12. bbb.minu[5] = 12. bbb.minu[6] = 12. bbb.minu[7] = 12. bbb.ziin[2] = 1 bbb.ziin[3] = 2 bbb.ziin[4] = 3 bbb.ziin[5] = 4 bbb.ziin[6] = 5 bbb.ziin[7] = 6 bbb.znuclin[0] = 1 bbb.znuclin[1] = 1 bbb.znuclin[2] = 6 bbb.znuclin[3] = 6 bbb.znuclin[4] = 6 bbb.znuclin[5] = 6 bbb.znuclin[6] = 6 bbb.znuclin[7] = 6 bbb.n0[2] = 1.e17 bbb.n0[3] = 1.e17 bbb.n0[4] = 1.e17 bbb.n0[5] = 1.e17 bbb.n0[6] = 1.e17 bbb.n0[7] = 1.e17 bbb.nzbackg = 1.e9 #background density for impurities bbb.inzb = 2 #exponent for switching on nzbackg bbb.ismctab = 2 # use Braams" rate tables com.mcfilename[0] = "C_rates.strahl" # Imp rate file name bbb.isnicore[7] = 3 bbb.curcore[7] = 0. bbb.isnwcono[2] = 3 bbb.isnwcono[3] = 3 bbb.isnwcono[4] = 3 bbb.isnwcono[5] = 3 bbb.isnwcono[6] = 3 bbb.isnwcono[7] = 3 bbb.isnwconi[2] = 3 bbb.isnwconi[3] = 3 bbb.isnwconi[4] = 3 bbb.isnwconi[5] = 3 bbb.isnwconi[6] = 3 bbb.isnwconi[7] = 3 bbb.nwomin[2] = 1.e7 bbb.nwomin[3] = 1.e7 bbb.nwomin[4] = 1.e7 bbb.nwomin[5] = 1.e7 bbb.nwomin[6] = 1.e7 bbb.nwomin[7] = 1.e7 bbb.nwimin[2] = 1.e7 bbb.nwimin[3] = 1.e7 bbb.nwimin[4] = 1.e7 bbb.nwimin[5] = 1.e7 bbb.nwimin[6] = 1.e7 bbb.nwimin[7] = 1.e7 bbb.restart = 1 #Begin from savefile, not estimated profiles # Filling newly allocated arrays as desired bbb.ftol = 1.e-8 bbb.allocate() #allocates storage for arrays from uedge.hdf5 import * hdf5_restore("d3dHm_Cog_strahl_Rp95.h5") # Atomic data switches com.istabon = 10 #=10 specifics hydrogen data file ehr2.dat # Scale factor converting (upi-upg)**2 energy to thermal energy bbb.cfnidh = 0.2
7,614
32.995536
77
py
UEDGE
UEDGE-master/pytests/fast_tests/Slab_geometry/test_sgeom.py
import unittest import os import numpy as np import sys,getopt from multiprocessing import Process thisfile=os.path.realpath(__file__) thispath=os.path.dirname(thisfile) sys.path.insert(0,thispath) sys.path.insert(0,os.path.dirname(os.path.dirname(thispath))+'/testscripts') import uetests as uet ftol = 2.e-9 def prhelp(): print(""" Usage: python test_slab.py [-r|--ref] [-h|--help] or pytest --forked test_slab.py -r|--ref to produce reference files for future test runs """) # # Note the numbering of the tests. Purists will say that # tests should be order independent. Because Uedge variables # are in shared objects it is naturally stateful. Because of # this stateful feature the order needs to be controlled. # class TestRun(unittest.TestCase): def setUp(self): """ This is run pre-test. Reads in restart data and re-converges solution. """ import uedge as ue os.chdir(thispath) def test_reconv(self): """ Test that initial re-converged solution fnrm is low. """ global ftol import uedge as ue uet.startup('rd_slabH_in_w_h5',ftol=ftol) uet.check_fnorm('Slab Geometry Reconverge','ref_reconv.h5','rd_slabH_in_w_h5',doassert=True) def test_perturb(self): """ Test that initial re-converged solution fnrm is low. """ global ftol import uedge as ue uet.startup('rd_slabH_in_w_h5',ftol=ftol) uet.perturb('rd_slabH_in_w_h5',ftol=ftol) uet.steadystate() uet.check_nfe('Slab Geometry Perturb','ref_perturb.h5','rd_slabH_in_w_h5',doassert=True) if __name__ == '__main__': try: opts, args = getopt.getopt(sys.argv[1:],'hr',['help,ref']) except getopt.GetoptError: prhelp() sys.exit(2) for opt,arg in opts: if opt in ('-h','--help'): prhelp() sys.exit(2) elif opt in ('-r','--ref'): kargs = {'ftol':ftol} p1 = Process(target=uet.makeref,args=('ref_reconv.h5','rd_slabH_in_w_h5'),kwargs=kargs) p1.start() p2 = Process(target=uet.makeperturb,args=('ref_perturb.h5','rd_slabH_in_w_h5'),kwargs=kargs) p2.start() p1.join() p2.join() sys.exit(2) unittest.main()
2,382
26.079545
103
py
UEDGE
UEDGE-master/pytests/fast_tests/Slab_geometry/rd_slabH_in_w_h5.py
# # ########################################################################### # DESCRIPTION OF PROBLEM (slabH): # Slab model with 4 hydrogen variables (ni,upi,te,ti) and a (6+2)*(10+2) # = 8x12 [poloidal*radial] mesh yielding 384 variables. The +2 in the mesh # size description arises from one guard-cell at each end of the domain used # to set boundary conditions. This case starts from generic internal initial # profiles (bbb.restart=0), which is generally the "hard way" to start if a # similar solution has been saved previous as an HDF5 (or PDB) file; see # ../tokgeo_H case. Also, this case has the neutrals frozen (bbb.isngon=0) # and uses a simple internal hydrogen ionization function for any residual # ionization from the frozen neutrals (small). # Solver is the time-dependent, Newton-Krylov ODE routine VODPK # (svrpkg="vodpk") and the preconditioner approximate LU decomposition is # provided by the ILUT sparse solver (premeth="ilut"). ########################################################################### import uedge from uedge import * bbb.mhdgeo=-1 bbb.isnion=1 bbb.isupon=1 bbb.isteon=1 bbb.istion=1 bbb.isngon=0 bbb.svrpkg="nksol" bbb.premeth="ilut" bbb.epscon1=1.e-2 bbb.ireorder=0 bbb.ncore=2.e19 bbb.tcoree=100 bbb.tcorei=100 bbb.isfixlb=2 bbb.recycp=.9 bbb.trange=4.e3 bbb.nsteps=1 bbb.n0g=1.e16 bbb.difni=1. bbb.kye=1. bbb.flalfe=0.21 bbb.flalfi=0.21 bbb.flalfgx=1.e10 bbb.flalfgy=1.e10 com.nycore=0 com.nysol=10 com.nxleg[0,0]=0 com.nxleg[0,1]=2 com.nxcore[0,0]=0 com.nxcore[0,1]=4 grd.zax=1. grd.zaxpt=.75 grd.alfyt=-1.e-5 bbb.restart=0 bbb.ftol = 1.e-5 bbb.dtreal = 1.e2 bbb.allocate() #allocates storage for arrays bbb.restart = 1 #use initial solution in slabH.h5 from uedge.hdf5 import * hdf5_restore("slabH.h5") # perform on time-step bbb.exmain() #now should have starting profile at dtreal=1.e-9; restart from that soln bbb.restart = 1 # Increase bbb.dtreal for more time-steps or read bbb.rdinitdt and bbb.rdcontdt
1,982
26.929577
79
py
UEDGE
UEDGE-master/pytests/fast_tests/d3d_snull_D+C_nonorthog/plate_d3d_2.py
# Set params for inner (1) and outer (2) plate line-segments from uedge import * #needed to define variables within this file grd.nplate1 = 5 grd.gchange("Mmod",0) grd.rplate1=[\ 1.600E+00, 1.27300E+00, 1.15310E+00, 1.01600E+00, 1.01600E+00] grd.zplate1=[\ 2.34100E-01, 2.34100E-01, 2.34100E-01, 3.71200E-01, 1.0000] grd.nplate2 = 10 grd.gchange("Mmod,0") grd.rplate2=[\ 2.13690E+00, 1.78570E+00, 1.76800E+00, 1.76800E+00, 1.68100E+00, 1.67500E+00, 1.67200E+00, 1.67200E+00, 1.55500E+00, 1.21200E+00] grd.zplate2=[\ 6.28600E-01, 4.25600E-01, 3.89300E-01, 3.46000E-01, 3.46000E-01, 3.43000E-01, 3.37000E-01, 2.34100E-01, 2.34100E-01, 2.34100E-01]
712
25.407407
66
py
UEDGE
UEDGE-master/pytests/fast_tests/d3d_snull_D+C_nonorthog/test_d3d_snull_D+C_nonorthog.py
import unittest import os import numpy as np import sys,getopt from multiprocessing import Process thisfile=os.path.realpath(__file__) thispath=os.path.dirname(thisfile) sys.path.insert(0,thispath) sys.path.insert(0,os.path.dirname(os.path.dirname(thispath))+'/testscripts') import uetests as uet ftol = 1.e-9 name = 'd3d_snull_D+C_nonorthog' restart = 'rd_d3dHmsCnog_strahl_Rp95' def prhelp(): print(""" Usage: python test_slab.py [-r|--ref] [-h|--help] or pytest --forked test_slab.py -r|--ref to produce reference files for future test runs """) # # Note the numbering of the tests. Purists will say that # tests should be order independent. Because Uedge variables # are in shared objects it is naturally stateful. Because of # this stateful feature the order needs to be controlled. # class TestRun(unittest.TestCase): def setUp(self): """ This is run pre-test. Reads in restart data and re-converges solution. """ import uedge as ue os.chdir(thispath) def test_reconv(self): """ Test that initial re-converged solution fnrm is low. """ global ftol import uedge as ue uet.startup(restart,ftol=ftol) uet.check_fnorm(name+' Reconverge','ref_reconv.h5',restart,doassert=True) def test_perturb(self): """ Test that initial re-converged solution fnrm is low. """ global ftol import uedge as ue uet.startup(restart,ftol=ftol) uet.perturb(restart,ftol=ftol) uet.steadystate() uet.check_nfe(name+' Perturb','ref_perturb.h5',restart,doassert=True) if __name__ == '__main__': try: opts, args = getopt.getopt(sys.argv[1:],'hr',['help,ref']) except getopt.GetoptError: prhelp() sys.exit(2) for opt,arg in opts: if opt in ('-h','--help'): prhelp() sys.exit(2) elif opt in ('-r','--ref'): kargs = {'ftol':ftol} p1 = Process(target=uet.makeref,args=('ref_reconv.h5',restart),kwargs=kargs) p1.start() p2 = Process(target=uet.makeperturb,args=('ref_perturb.h5',restart),kwargs=kargs) p2.start() p1.join() p2.join() sys.exit(2) unittest.main()
2,359
25.516854
92
py
UEDGE
UEDGE-master/pytests/fast_tests/d3d_snull_D+C_nonorthog/rd_d3dHmsCnog_strahl_Rp95.py
# # ########################################################################### # DESCRIPTION OF PROBLEM (d3dHmsCnog) from FACETS test suite: # DIII-D single-null geometry with 6 hydrogen variables (ni,ng,upi,upg,te,ti) # and 7 carbon variables (six charge-state densities ni and one ng) on a # (16+2)*(8+2)=18x10 [poloidal*radial] mesh yielding 2340 variables. # Solver used is Newton Krylov (svrpkg="nksol") and preconditioner uses an # iterative solver ILUT for Jacobian LU decomposition. Also includes tilted # divertor plates wrt flux-surface normal, thus testing the nonorthogonal # finite-volume difference stencil. Iterates to steady-state solution from # an initial profile file (HF5). ########################################################################### # Import uedge into python and make variables active import uedge from uedge import * # Begin uedge parameter input # Set the geometry bbb.mhdgeo = 1 #=1 use MHD equilibrium files ##flx.aeqdskfname = "a110465.03500" #EFIT "a" file for flux-surface mesh ##flx.geqdskfname = "g110465.03500" #EFIT "g" or "n" file for flux-sur mesh flx.psi0min1 = 0.96 #normalized flux on core bndry flx.psi0min2 = 0.98 #normalized flux on pf bndry flx.psi0sep = 1.00001 #normalized flux at separatrix flx.psi0max = 1.07 #normalized flux on outer wall bndry bbb.ngrid = 1 #number of mesh sequenc. (always set to 1) bbb.recycm = 0.1 com.nxleg[0,0] = 4 #pol. mesh pts from inner plate to x-point com.nxcore[0,0] = 4 #pol. mesh pts from x-point to top on inside com.nxcore[0,1] = 4 #pol. mesh pts from top to x-point on outside com.nxleg[0,1] = 4 #pol. mesh pts from x-point to outer plate com.nysol[0] = 4 #rad. mesh pts in SOL com.nycore[0] = 4 #rad. mesh pts in core flx.alfcy = 2 #>0 concentrates y-mesh near separatrix # Mesh construction--non orthogonal mesh com.ismmon = 3 #controls transition from nonorthog to orthog mesh com.isnonog = 1 # non orthogonal differencing grd.istream = 0 grd.iplate = 1 grd.nsmooth = 3 grd.wtmesh1 = 0.75 grd.dmix0 = 1.0 # Define divertor-plate geometry from line-segment file from plate_d3d_2 import * # Finite-difference algorithms (upwind, central diff, etc.) bbb.methn = 33 #ion continuty eqn bbb.methu = 33 #ion parallel momentum eqn bbb.methe = 33 #electron energy eqn bbb.methi = 33 #ion energy eqn bbb.methg = 66 # Boundary conditions bbb.ncore[0] = 2.e19 #hydrogen ion density on core bbb.iflcore = 0 #flag; =0, fixed Te,i; =1, fixed power on core bbb.tcoree = 401. #core Te bbb.tcorei = 401. #core Ti bbb.tedge = 2. #fixed wall,pf Te,i if istewcon=1, etc bbb.recycp[0] = 0.95 #hydrogen recycling coeff at plates # Transport coefficients (m**2/s) bbb.difni[0] = 1. #D for radial hydrogen diffusion bbb.kye = 1. #chi_e for radial elec energy diffusion bbb.kyi = 1. #chi_i for radial ion energy diffusion bbb.travis[0] = 1. #eta_a for radial ion momentum diffusion # Flux limits bbb.flalfe = 0.21 #electron parallel thermal conduct. coeff bbb.flalfi = 0.21 #ion parallel thermal conduct. coeff bbb.flalfv = 0.5 #ion parallel viscosity coeff bbb.flalfgx = 1. #neut. gas part. flux in poloidal direction bbb.flalfgy = 1. #neut. gas part. flux in radial direction bbb.flalfgxy = 1. #neut. gas part. flux in mixed derivatives bbb.flalftgx = 1. #neut. gas thermal flux, poloidal direction bbb.flalftgy = 1. #neut. gas thermal flux, radial direction bbb.lgmax = 0.1 #max scale length for flalfgx,y bbb.lgtmax = 0.1 #max scale length for flalftgx,y bbb.lgvmax = 0.1 #max scale length for flalfvgx,y # Solver package bbb.svrpkg = "nksol" #Newton solver using Krylov method bbb.premeth = "ilut" #Solution method for precond. Jacobian matrix # Parallel neutral momentum equation bbb.isupgon[0] = 1 bbb.ineudif = 2 #=2 for evolving pg=ng*tg variable bbb.isngon[0] = 0 com.ngsp = 1 com.nhsp = 2 bbb.ziin[1] = 0 ## Impurity gas basics com.ngsp = 2 #total number of gas species bbb.isngon[1] = 1 #turns on impurity gas bbb.ngbackg[1] = 1.e9 #neutral impurity background for added source bbb.ingb = 2 #exponent for strength of ngbackg turn-on bbb.istgcon[1] = 1 #=1 for constant tg(2) at tgas(2) bbb.tgas[1] = 1. #value for tg when istgcon=1 bbb.rcxighg = 0. # best value; ratio of imp cx to hyd cx bbb.kelighi[1] = 5.e-16 #elastic sig_v for imp_gas/h_ion bbb.kelighg[1] = 5.e-16 #elastic sig_v for imp_gas/h_gas bbb.n0g[1] = 1.e16 #imp. gas density normalization # Impurity gas boundary conditions bbb.recycp[1] = 0.01 #plate recycling of impurities bbb.recycw[1] = 1e-4 #wall recycling; matwsi,o set above for hyd bbb.isch_sput[1]=7 # Haasz/Davis chemical sputtering model bbb.isph_sput[1]=3 # physical sputtering model bbb.t_wall = 300. bbb.t_plat = 500. ## Impurity ions bbb.isimpon = 6 #Use force-balance only com.nzsp[0] = 6 #number chrg states impurity isotope #1 bbb.csfaclb[2,0] = 2.191 bbb.csfaclb[3,0] = 2.191 bbb.csfaclb[4,0] = 2.191 bbb.csfaclb[5,0] = 2.191 bbb.csfaclb[6,0] = 2.191 bbb.csfaclb[7,0] = 2.191 bbb.csfacrb[2,0] = 2.191 bbb.csfacrb[3,0] = 2.191 bbb.csfacrb[4,0] = 2.191 bbb.csfacrb[5,0] = 2.191 bbb.csfacrb[6,0] = 2.191 bbb.csfacrb[7,0] = 2.191 bbb.csfaclb[2,1] = 2.191 bbb.csfaclb[3,1] = 2.191 bbb.csfaclb[4,1] = 2.191 bbb.csfaclb[5,1] = 2.191 bbb.csfaclb[6,1] = 2.191 bbb.csfaclb[7,1] = 2.191 bbb.csfacrb[2,1] = 2.191 bbb.csfacrb[3,1] = 2.191 bbb.csfacrb[4,1] = 2.191 bbb.csfacrb[5,1] = 2.191 bbb.csfacrb[6,1] = 2.191 bbb.csfacrb[7,1] = 2.191 bbb.minu[2] = 12. bbb.minu[3] = 12. bbb.minu[4] = 12. bbb.minu[5] = 12. bbb.minu[6] = 12. bbb.minu[7] = 12. bbb.ziin[2] = 1 bbb.ziin[3] = 2 bbb.ziin[4] = 3 bbb.ziin[5] = 4 bbb.ziin[6] = 5 bbb.ziin[7] = 6 bbb.znuclin[0] = 1 bbb.znuclin[1] = 1 bbb.znuclin[2] = 6 bbb.znuclin[3] = 6 bbb.znuclin[4] = 6 bbb.znuclin[5] = 6 bbb.znuclin[6] = 6 bbb.znuclin[7] = 6 bbb.n0[2] = 1.e17 bbb.n0[3] = 1.e17 bbb.n0[4] = 1.e17 bbb.n0[5] = 1.e17 bbb.n0[6] = 1.e17 bbb.n0[7] = 1.e17 bbb.nzbackg = 1.e9 #background density for impurities bbb.inzb = 2 #exponent for switching on nzbackg bbb.ismctab = 2 # use Braams" rate tables com.mcfilename[0] = "C_rates.strahl" # Imp rate file name bbb.isnicore[7] = 3 bbb.curcore[7] = 0. bbb.isnwcono[2] = 3 bbb.isnwcono[3] = 3 bbb.isnwcono[4] = 3 bbb.isnwcono[5] = 3 bbb.isnwcono[6] = 3 bbb.isnwcono[7] = 3 bbb.isnwconi[2] = 3 bbb.isnwconi[3] = 3 bbb.isnwconi[4] = 3 bbb.isnwconi[5] = 3 bbb.isnwconi[6] = 3 bbb.isnwconi[7] = 3 bbb.nwomin[2] = 1.e7 bbb.nwomin[3] = 1.e7 bbb.nwomin[4] = 1.e7 bbb.nwomin[5] = 1.e7 bbb.nwomin[6] = 1.e7 bbb.nwomin[7] = 1.e7 bbb.nwimin[2] = 1.e7 bbb.nwimin[3] = 1.e7 bbb.nwimin[4] = 1.e7 bbb.nwimin[5] = 1.e7 bbb.nwimin[6] = 1.e7 bbb.nwimin[7] = 1.e7 bbb.restart = 1 #Begin from savefile, not estimated profiles # Filling newly allocated arrays as desired bbb.ftol = 1.e-8 bbb.allocate() #allocates storage for arrays from uedge.hdf5 import * hdf5_restore("d3dHm_Cnog_strahl_Rp95.h5") # Atomic data switches com.istabon = 10 #=10 specifics hydrogen data file ehr2.dat # Scale factor converting (upi-upg)**2 energy to thermal energy bbb.cfnidh = 0.2
7,525
33.054299
77
py
UEDGE
UEDGE-master/jupyter/case_setup.py
#Coarse mesh [com.nx=8, com.ny=4] for DIII-D MHD equilibrium #Uses diffusive neutrals, so five variables [bbb.ni,bbb.upi,Te,Ti,bbb.ng] # import sys,os from uedge import * # Set the com.geometry bbb.mhdgeo = 1 #use MHD equilibrium #com.rm -flx.f aeqdsk neqdsk #change names of MHD eqil. files flx.psi0min1 = 0.98 #normalized flux on core bndry flx.psi0min2 = 0.98 #normalized flux on pf bndry flx.psi0sep = 1.00001 #normalized flux at separatrix flx.psi0max = 1.07 #normalized flux on outer wall bndry bbb.ngrid = 1 #number of meshes [always set to 1] com.nxleg[0,0] = 4 #pol. mesh pts from inner plate to flx.x-point com.nxcore[0,0] = 4 #pol. mesh pts from flx.x-point to top on inside com.nxcore[0,1] = 4 #pol. mesh pts from top to flx.x-point on outside com.nxleg[0,1] = 4 #pol. mesh pts from flx.x-point to outer plate com.nysol[0] = 6 #rad. mesh pts in SOL com.nycore[0] =2 #rad. mesh pts in core # Finite-difference algorithms [upwind, central diff, etc.] bbb.methn = 33 #ion continuty eqn bbb.methu = 33 #ion parallel momentum eqn bbb.methe = 33 #electron energy eqn bbb.methi = 33 #ion energy eqn bbb.methg = 33 #neutral gas continuity eqn # Boundary conditions bbb.ncore[0] = 2.e19 #hydrogen ion density on core bbb.iflcore = 0 #flag =0, fixed Te,i =1, fixed power on core bbb.tcoree = 100. #core Te bbb.tcorei = 100. #core Ti bbb.tedge = 2. #fixed wall,pf Te,i if istewcon=1, etc bbb.istepfc = 3 #=3 sets scale length bbb.lyte bbb.lyte = 0.03 #radial scale-length for Te on PF boundary bbb.recycp[0] = 0.98 #hydrogen recycling grd.coeff at plates bbb.recycw[0] = 0.9 #wall recycling if bbb.matwso,i=1 bbb.matwso[0] = 1 #recycle on main-chamber wall bbb.isnwcono = 1 #if=1, set bbb.ni[,,com.ny+1]=bbb.nwallo bbb.isnwconi = 1 #if=1, set PF wall bbb.ni=bbb.nwalli bbb.allocate() #bbb.allocate() space of bbb.nwallo,i bbb.nwallo = 1.e18 bbb.nwalli = 1.e18 # Transport coefficients bbb.difni[0] = 1. #D for radial hydrogen diffusion bbb.kye = 1. #chi_e for radial elec energy diffusion bbb.kyi = 1. #chi_i for radial ion energy diffusion bbb.travis[0] = 1. #eta_a for radial ion momentum diffusion # Flux limits bbb.flalfe = 0.21 #electron parallel thermal conduct. grd.coeff bbb.flalfi = 0.21 #ion parallel thermal conduct. grd.coeff bbb.flalfv = 1. #ion parallel viscosity grd.coeff bbb.flalfgx = 1. #neut. density poloidal diffusion bbb.flalfgy = 1. #neut. density radial diffusion bbb.flalfvgx = 1. #neut. momentum poloidal viscosity bbb.flalfvgy = 1. #neut. momentum radial viscosity bbb.flalftgx = 1. #neut. particle poloidal heat diffusion bbb.flalftgy = 1. #neut. particle radial heat diffusion # Solver package bbb.svrpkg = "nksol" #Newton solver using Krylov method bbb.premeth = "ilut" #Solution method for precond. Jacobian matrix bbb.mfnksol = 3 # Parallel neutral momentum equation bbb.ineudif = 2 bbb.isupgon[0]=1 if bbb.isupgon[0] == 1: bbb.isngon[0]=0 com.ngsp=1 com.nhsp=2 bbb.ziin[com.nhsp-1]=0 bbb.travis[1] = 0. #shouldn't be used for neutrals - to be sure # The following are probably default, set them anyway to be sure bbb.cngmom=0 bbb.cmwall=0 bbb.cngtgx=0 bbb.cngtgy=0 bbb.kxn=0 bbb.kyn=0 ## bbb.ingb = 0 # Currents and potenial parameters bbb.isphion = 1 bbb.b0 = 1. # =1 for normal direction of B-field bbb.rsigpl=1.e-8 #anomalous cross-field conductivity bbb.cfjhf=1. #turn-on heat flow from current [bbb.fqp] bbb.cfjve = 1. #makes bbb.vex = vix - bbb.cfjve*bbb.fqx bbb.jhswitch=1 #Joule Heating switch bbb.cfjpy = 0. #diamag. cur. in flx.y-direction bbb.cfjp2 = 0. #diamag. cur. in 2-direction bbb.newbcl=1 bbb.newbcr=1 #Sheath BC [bee,i] from current equation bbb.isfdiax =1. #Factor to turn on diamag. contrib. to sheath bbb.cfyef = 1.0 #ExB drift in flx.y-dir. bbb.cf2ef = 1.0 #ExB drift in 2-dir. bbb.cfydd = 0. #Diamag. drift in flx.y-dir. [always=0] bbb.cf2dd = 0. #Diamag. drift in 2-dir. [always=0] bbb.cfrd = 0. #Resistive drift in flx.y and 2 dirs. bbb.cfbgt = 0. #Diamag. energy drift [always=0] bbb.cfybf = 1. #turns on bbb.vycb - radial grad_B drift bbb.cf2bf = 1. #turns on bbb.v2cb - perp grad_B drift [nearly pol] bbb.cfqybf = 1. #turns on bbb.vycb contrib to radial current bbb.cfq2bf = 1. #turns on bbb.v2cb contrib to perp["2"] current bbb.isnewpot = 1 bbb.rnewpot = 1. bbb.iphibcc = 3 #set bbb.phi[,1] uniform poloidally # Restart from bbb.a pfb savefile bbb.restart = 1 #Begin from savefile, not estimated profiles bbb.allocate() #bbb.allocate() space for savevariables
4,590
38.239316
79
py
StyleFusion
StyleFusion-master/src/main.py
from shared import * from tf_lib import * from dataset import * from model import * #from dialog_gui import * from classifier import load_classifier """ AUTHOR: Xiang Gao (xiag@microsoft.com) at Microsoft Research """ def run_master(mode, args): if mode not in ['train','continue'] and args.restore != '': aa = args.restore.split('/') bb = [] for a in aa: if len(a) > 0: bb.append(a) fld = '/'.join(bb[:-1]) if mode in ['vali','vis']: vocab_only = False fld_data, _, _ = get_model_fld(args) path_bias_vocab = fld_data + '/vocab_bias.txt' else: vocab_only = True fld_data = fld path_bias_vocab = fld + '/vocab_bias.txt' else: vocab_only = False fld_data, fld_model, subfld = get_model_fld(args) fld = fld_model + '/' + subfld path_bias_vocab = fld_data + '/vocab_bias.txt' if os.path.exists(path_bias_vocab): allowed_words = [line.strip('\n').strip('\r') for line in open(path_bias_vocab, encoding='utf-8')] else: allowed_words = None model_class = args.model_class.lower() if model_class.startswith('fuse'): Master = StyleFusion elif model_class == 'mtask': Master = VanillaMTask elif model_class == 's2s': Master = Seq2Seq elif model_class == 'lm': Master = LanguageModel elif model_class == 's2s+lm': pass else: raise ValueError if model_class == 's2s+lm': master = Seq2SeqLM(args, allowed_words) else: dataset = Dataset(fld_data, max_ctxt_len=args.max_ctxt_len, max_resp_len=args.max_resp_len, vocab_only=vocab_only, noisy_vocab=args.noisy_vocab) master = Master(dataset, fld, args, new=(mode=='train'), allowed_words=allowed_words) if mode != 'train': if args.restore.endswith('.npz') or model_class == 's2s+lm': restore_path = args.restore else: restore_path = master.fld + '/models/%s.npz'%args.restore master.load_weights(restore_path) if mode in ['vis', 'load']: return master if args.clf_name.lower() == 'holmes': CLF_NAMES = ['classifier/Reddit_vs_Holmes/neural', 'classifier/Reddit_vs_Holmes/ngram'] elif args.clf_name.lower() == 'arxiv': CLF_NAMES = ['classifier/Reddit_vs_arxiv/neural', 'classifier/Reddit_vs_arxiv/ngram'] else: CLF_NAMES = [args.clf_name] print('loading classifiers '+str(CLF_NAMES)) master.clf_names = CLF_NAMES master.classifiers = [] for clf_name in CLF_NAMES: master.classifiers.append(load_classifier(clf_name)) print('\n'+fld+'\n') if mode in ['continue', 'train']: ss = ['', mode + ' @ %i'%time.time()] for k in sorted(args.__dict__.keys()): ss.append('%s = %s'%(k, args.__dict__[k])) with open(master.fld + '/args.txt', 'a') as f: f.write('\n'.join(ss)+'\n') if args.debug: batch_per_load = 1 else: if PHILLY: n_sample = 1280 # philly unstable for large memory else: n_sample = 2560 batch_per_load = int(n_sample/BATCH_SIZE) if mode == 'continue': master.vali() master.train(batch_per_load) elif 'summary' == mode: print(master.model.summary()) elif mode in ['cmd', 'test', 'vali']: classifiers = [] for clf_name in CLF_NAMES: classifiers.append(load_classifier(clf_name)) if 'vali' == mode: data = master.get_vali_data() s_decoded = eval_decoded(master, data, classifiers=classifiers, corr_by_tgt=True, r_rand=args.r_rand, calc_retrieval=('holmes' in args.data_name.lower()) )[0] s_surrogate = eval_surrogate(master, data)[0] print(restore_path) print() print(s_decoded) print() print(s_surrogate) return """ if model_class != 's2s+lm': with tf.variable_scope('base_rankder', reuse=tf.AUTO_REUSE): fld_base_ranker = 'restore/%s/%s/pretrained/'%(args.model_class.replace('fuse1','fuse'), args.data_name) dataset_base_ranker = Dataset(fld_base_ranker, max_ctxt_len=args.max_ctxt_len, max_resp_len=args.max_resp_len, vocab_only=True, noisy_vocab=False) base_ranker = Master(dataset_base_ranker, fld_base_ranker, args, new=False, allowed_words=master.allowed_words) path = fld_base_ranker + '/' + open(fld_base_ranker+'/base_ranker.txt').readline().strip('\n') base_ranker.load_weights(path) print('*'*10 + ' base_ranker loaded from: '+path) else: """ base_ranker = None def print_results(results): ss = ['total', 'logP', 'logP_c', 'logP_b', 'rep', 'len',] + ['clf%i'%i for i in range(len(CLF_NAMES))] print('; '.join([' '*(6-len(s))+s for s in ss])) for score, resp, terms in results: print('%6.3f; '%score + '; '.join(['%6.3f'%x for x in terms]) + '; ' + resp) if 'cmd' == mode: while True: print('\n---- please input ----') inp = input() infer_args = parse_infer_args() if inp == '': break results = infer_rank(inp, master, infer_args, base_ranker=base_ranker) print_results(results) elif 'test' == mode: infer_args = parse_infer_args() path_out = args.path_test+'.hyp' open(path_out, 'w', encoding='utf-8') for line in open(args.path_test, encoding='utf-8'): line = line.strip('\n') inp = line.split('\t')[0] results = infer_rank(inp, master, infer_args, base_ranker=base_ranker) lines = [] for _, hyp, _ in results[:min(10, len(results))]: lines.append(line + '\t' + hyp.replace(' _EOS_','').strip()) with open(path_out, 'a', encoding='utf-8') as f: f.write('\n'.join(lines) + '\n') """ path_in = DATA_PATH + '/test/' + args.test_fname if not PHILLY: fld_out = master.fld + '/eval2/' else: fld_out = OUT_PATH makedirs(fld_out) npz_name = args.restore.split('/')[-1].replace('.npz','') path_out = fld_out + '/' + args.test_fname + '_' + npz_name test_master(master, path_in, path_out, max_n_src=args.test_n_max, base_ranker=base_ranker, baseline=args.baseline, r_rand=args.r_rand) """ else: raise ValueError def get_model_fld(args): data_name = args.data_name if PHILLY: data_name = data_name.replace('+','').replace('_','') fld_data = DATA_PATH +'/' + data_name master_config = 'width%s_depth%s'%( (args.token_embed_dim, args.rnn_units), (args.encoder_depth, args.decoder_depth)) if args.max_ctxt_len != 90 or args.max_resp_len != 30: master_config += '_len' + str((args.max_ctxt_len, args.max_resp_len)) master_config = master_config.replace("'",'') fld_model = OUT_PATH if args.debug: fld_model += '/debug' fld_model += '/' + args.data_name.replace('../','') + '_' + master_config subfld = [] """ if args.randmix: s_mix = 'randmix' if args.ratio05 > 0: s_mix += '(0.5=%.2f)'%args.ratio05 else: """ s_mix = 'mix' model_class = args.model_class.lower() if model_class == 's2s': subfld = ['s2s_%s(%.2f)'%(s_mix, args.conv_mix_ratio)] # no conv data else: subfld = ['%s_%s(%.2f,%.2f)'%(model_class, s_mix, args.conv_mix_ratio, args.nonc_mix_ratio)] if args.noisy_vocab > 0: subfld.append('unk%.1fk'%(args.noisy_vocab/1000)) if model_class.startswith('fuse'): subfld.append('std%.1f'%args.stddev) if args.reld: subfld.append('reld') subfld.append('lr'+str(args.lr)) if len(args.fld_suffix) > 0: subfld.append(args.fld_suffix) subfld = '_'.join(subfld) return fld_data, fld_model.replace(' ',''), subfld.replace(' ','') if __name__ == '__main__': parser.add_argument('mode') parser.add_argument('--skip', type=float, default=0.0) parser.add_argument('--test_fname', default='') parser.add_argument('--r_rand', '-r', type=float, default=-1) parser.add_argument('--test_n_max', '-n', type=int, default=2000) args = parser.parse_args() run_master(args.mode, args)
7,523
28.73913
137
py
StyleFusion
StyleFusion-master/src/decode.py
from shared import * from nltk.translate.bleu_score import SmoothingFunction """ AUTHOR: Sean Xiang Gao (xiag@microsoft.com) at Microsoft Research """ class Decoder: def __init__(self, dataset, model, decoder_depth, latent_dim, allowed_words=None): self.dataset = dataset self.model = model self.decoder_depth = decoder_depth self.latent_dim = latent_dim if allowed_words is None: self.mask = np.array([1.] * (self.dataset.num_tokens + 1)) else: self.mask = np.array([0.] * (self.dataset.num_tokens + 1)) for word in allowed_words: ix = self._ix(word) if ix is not None: self.mask[ix] = 1. print('allowed words %i/%i'%(sum(self.mask), len(self.mask))) default_forbid = [UNK_token, '(', '__url__', ')', EQN_token, CITE_token, IX_token] #+ ['queer', 'holmes', 'sherlock', 'john', 'watson', 'bannister'] for word in default_forbid: ix = self._ix(word) if ix is not None: self.mask[ix] = 0. # in either case, UNK is not allowed def _ix(self, token): return self.dataset.token2index.get(token, None) def predict(self, latents, sampling=False, softmax_temperature=1, lm_wt=None): # autoregressive in parallel, greedy or softmax sampling latents = np.reshape(latents, (-1, self.latent_dim)) # (n, dim) n = latents.shape[0] n_vocab = len(self.mask) prev = np.zeros((n, 1)) + self._ix(SOS_token) states = [latents] * self.decoder_depth # list of state, each is [n, dim] mask = np.repeat(np.reshape(self.mask, (1, -1)), n, axis=0) # (n, vocab) logP = [0.] * n stop = [False] * n hyp = [] for _ in range(n): hyp.append([]) def sample_token_index_softmax(prob): if softmax_temperature != 1: prob = np.exp(np.log(prob) * softmax_temperature) return np.random.choice(n_vocab, 1, p=prob/sum(prob))[0] def sample_token_index_greedy(prob): return np.argmax(prob) if sampling: sample_token_index = sample_token_index_softmax else: sample_token_index = sample_token_index_greedy for _ in range(self.dataset.max_resp_len): out = self.model.predict([prev] + states) states = out[1:] tokens_proba = np.squeeze(out[0]) * mask # squeeze: (n, 1, vocab) => (n, vocab) prev = [0] * n for i in range(n): if stop[i]: continue prob = tokens_proba[i,:].ravel() ix = sample_token_index(prob) logP[i] += np.log(prob[ix]) hyp[i].append(ix) prev[i] = ix if ix == self._ix(EOS_token): stop[i] = True prev = np.reshape(prev, (n, 1)) return [logP[i]/len(hyp[i]) for i in range(n)], hyp def evaluate(self, latents, tgt_seqs): # teacher-forcing in parallel latents = np.reshape(latents, (-1, self.latent_dim)) # (n, dim) n = latents.shape[0] states = [latents] * self.decoder_depth # list of state, each is [n, dim] logP = [0.] * n prev = np.zeros((n, 1)) + self._ix(SOS_token) lens = [len(seq) for seq in tgt_seqs] epsilon = 1e-6 for t in range(self.dataset.max_resp_len): out = self.model.predict([prev] + states) states = out[1:] tokens_proba = np.reshape(out[0], (n, -1)) # squeeze: (n, 1, vocab) => (n, vocab) prev = [0] * n for i in range(n): if t < lens[i]: ix = tgt_seqs[i][t] logP[i] += np.log(max(epsilon, tokens_proba[i, ix])) prev[i] = ix prev = np.reshape(prev, (n, 1)) return [logP[i]/lens[i] for i in range(n)] #return [logP[i]/self.dataset.max_resp_len for i in range(n)] def predict_beam(self, latents, beam_width=10, n_child=3, max_n_hyp=100): # multi-head beam search, not yet parallel prev = np.atleast_2d([self._ix(SOS_token)]) beam = [] for latent in latents: latent = np.atleast_2d(latent) states = [latent] * self.decoder_depth node = {'states':states[:], 'prev':prev, 'logP':0, 'hyp':[]} beam.append(node) print('beam search initial n = %i'%len(beam)) results = queue.PriorityQueue() t = 0 while True: t += 1 if t > 20:#self.dataset.max_tgt_len: break if len(beam) == 0: break pq = queue.PriorityQueue() for node in beam: out = self.model.predict([node['prev']] + node['states']) tokens_proba = out[0].ravel() states = out[1:] tokens_proba = tokens_proba * self.mask tokens_proba = tokens_proba/sum(tokens_proba) top_tokens = np.argsort(-tokens_proba) for ix in top_tokens[:n_child]: logP = node['logP'] + np.log(tokens_proba[ix]) hyp = node['hyp'][:] + [ix] if ix == self._ix(EOS_token): results.put((logP/t, hyp)) if results.qsize() > max_n_hyp: results.get() # pop the hyp of lowest logP/t continue pq.put(( logP, # no need to normalize to logP/t as every node is at the same t np.random.random(), # to avoid the case logP is the same { 'states':states, 'prev':np.atleast_2d([ix]), 'logP':logP, 'hyp':hyp, } )) if pq.qsize() > beam_width: pq.get() # pop the node of lowest logP to maintain at most beam_width nodes => but this will encourage bland response beam = [] while not pq.empty(): _, _, node = pq.get() beam.append(node) logPs = [] hyps = [] while not results.empty(): logP, hyp = results.get() logPs.append(logP) hyps.append(hyp) return logPs, hyps def rank_nbest(hyps, logP, logP_center, master, inp, infer_args=dict(), base_ranker=None): # make sure hyps are list of str, and inp is str # as base_ranker, master, and clf may not share the same vocab assert(isinstance(hyps, list)) assert(isinstance(hyps[0], str)) assert(isinstance(inp, str)) hyps_no_ie = [] for hyp in hyps[:]: hyps_no_ie.append((' '+hyp+' ').replace(' i . e . ,',' ').replace(' i . e. ',' ').strip()) hyps = hyps_no_ie[:] wt_clf = infer_args.get('wt_clf', 0) / len(master.classifiers) wt_rep = infer_args.get('wt_rep', 0) wt_len = infer_args.get('wt_len', 0) wt_center = infer_args.get('wt_center', 0) wt_base = infer_args.get('wt_base', 0) n = len(logP) clf_score = [] max_tgt_len = 30 for clf in master.classifiers: clf_score.append(clf.predict(hyps).ravel()) if base_ranker is not None: hyp_seqs_base = [base_ranker.dataset.txt2seq(hyp) for hyp in hyps] inp_seq_base = base_ranker.dataset.txt2seq(inp) latent_base = base_ranker.model_encoder['S2S'].predict(np.atleast_2d(inp_seq_base)) logP_base = base_ranker.decoder.evaluate([latent_base]*n, hyp_seqs_base) else: logP_base = [0] * n pq = queue.PriorityQueue() for i in range(n): hyp = hyps[i] rep = repetition_penalty(hyp) l = min(max_tgt_len, len(hyp.split()))/max_tgt_len score = logP[i] + wt_center * logP_center[i] + wt_rep * rep + wt_len * l + wt_base * logP_base[i] clf_score_ = [] for k in range(len(master.classifiers)): s = clf_score[k][i] score += wt_clf * s clf_score_.append(s) pq.put((-score, hyp, (logP[i], logP_center[i], logP_base[i], rep, l) + tuple(clf_score_))) results = [] while not pq.empty(): neg_score, hyp, terms = pq.get() #if len(set(['queer', 'holmes', 'sherlock', 'john', 'watson', 'bannister']) & set(hyp.split())) > 0: # continue hyp = (' ' + hyp + ' ').replace(' to day ',' today ').replace(' to morrow ',' tomorrow ')#.replace('mr barker','') results.append((-neg_score, hyp, terms)) return results def repetition_penalty(hyp): # simplified from https://sunlamp.visualstudio.com/sunlamp/_git/sunlamp?path=%2Fsunlamp%2Fpython%2Fdynamic_decoder_custom.py&version=GBmaster # ratio of unique 1-gram ww = hyp.split() return np.log(min(1.0, len(set(ww)) / len(ww))) def infer(latent, master, method='greedy', beam_width=10, n_rand=20, r_rand=1.5, softmax_temperature=1, lm_wt=0.5): if method == 'greedy': return master.decoder.predict(latent, lm_wt=lm_wt) elif method == 'softmax': return master.decoder.predict([latent] * n_rand, sampling=True, lm_wt=lm_wt) elif method == 'beam': return master.decoder.predict_beam([latent], beam_width=beam_width) elif method.startswith('latent'): latents = [] if r_rand >= 0: rr = [r_rand] * n_rand else: rr = np.linspace(0, 5, n_rand) for r in rr: latents.append(rand_latent(latent, r, limit=True)) if 'beam' in method: return master.decoder.predict_beam(latents, beam_width=beam_width) else: return master.decoder.predict(latents, sampling=('softmax' in method), softmax_temperature=softmax_temperature, lm_wt=lm_wt) else: raise ValueError def infer_comb(inp, master): inp_seq = master.dataset.txt2seq(inp) latent = master.model_encoder['S2S'].predict(np.atleast_2d(inp_seq)) reset_rand() logP, hyp_seqs = infer(latent, master, method='latent', n_rand=10, r_rand=-1) logP, hyp_seqs = remove_duplicate_unfished(logP, hyp_seqs, master.dataset.token2index[EOS_token]) results = sorted(zip(logP, hyp_seqs), reverse=True) s = '-'*10 + '\n' + inp + '\n' for i, (logP, seq) in enumerate(results): hyp = master.dataset.seq2txt(seq) s += '%.3f'%logP + '\t' + hyp + '\n' if i == 4: break s += '-'*5 + '\n' return s def remove_duplicate_unfished(logP, hyp_seqs, ix_EOS): d = dict() for i in range(len(logP)): k = tuple(hyp_seqs[i]) if k[-1] != ix_EOS: continue if k not in d or logP[i] > d[k]: d[k] = logP[i] logP0, hyp0 = logP[0], hyp_seqs[0][:] logP = [] hyp_seqs = [] for k in d: logP.append(d[k]) hyp_seqs.append(list(k)) if len(logP) == 0: return [logP0], [hyp0] else: return logP, hyp_seqs def parse_infer_args(): arg = {'prefix':'S2S'} for line in open('src/infer_args.csv'): if line.startswith('#'): continue if ',' not in line: continue k, v = line.strip('\n').split(',') if k != 'method': if k in ['beam_width', 'n_rand']: v = int(v) else: v = float(v) arg[k] = v return arg def infer_rank(inp, master, infer_args, base_ranker=None, unique=True, verbose=True): if verbose: print('infer_args = '+str(infer_args)) inp_seq = master.dataset.txt2seq(inp) latent = master.model_encoder['S2S'].predict(np.atleast_2d(inp_seq)) reset_rand() if verbose: print('infering...') t0 = datetime.datetime.now() logP, hyp_seqs = infer(latent, master, method=infer_args['method'], beam_width=infer_args.get('beam_width'), n_rand=infer_args.get('n_rand'), r_rand=infer_args.get('r_rand'), softmax_temperature=infer_args.get('softmax_temperature'), lm_wt=infer_args.get('lm_wt')) t1 = datetime.datetime.now() if verbose: print('*'*10 + ' infer spent: '+str(t1-t0)) n_raw = len(logP) logP, hyp_seqs = remove_duplicate_unfished(logP, hyp_seqs, master.dataset.token2index[EOS_token]) if verbose: print('kept %i/%i after remove deuplication/unfisihed'%(len(logP), n_raw)) hyps = [master.dataset.seq2txt(seq) for seq in hyp_seqs] if len(hyps) == 0: return [] n_results = len(logP) if infer_args['method'] == 'latent' and infer_args['r_rand'] > 0: if verbose: print('calculating tf_logP...') logP_center = master.decoder.evaluate([latent]*n_results, hyp_seqs) else: logP_center = logP t2 = datetime.datetime.now() if verbose: print('*'*10 + ' logP_center spent: '+str(t2-t1)) wts_classifier = [] for clf_name in master.clf_names: wts_classifier.append(infer_args.get(clf_name, 0)) if verbose: print('ranking...') results = rank_nbest(hyps, logP, logP_center, master, inp, infer_args, base_ranker) t3 = datetime.datetime.now() if verbose: print('*'*10 + ' ranking spent: '+str(t3-t2)) return results
11,359
29.455764
150
py
StyleFusion
StyleFusion-master/src/vis.py
from shared import * from tf_lib import * from main import run_master, get_model_fld from scipy.optimize import fmin_powell as fmin from mpl_toolkits.mplot3d import Axes3D from sklearn import manifold import scipy """ AUTHOR: Sean Xiang Gao (xiag@microsoft.com) at Microsoft Research """ def dist_mat(coord): n = coord.shape[0] dist_T2T = np.zeros((n, n)) for i in range(n): for j in range(i + 1, n): d = euc_dist(coord[i, :], coord[j, :]) dist_T2T[i, j] = d dist_T2T[j, i] = d return dist_T2T def interp(master, model_name, fld_save, type_='resp'): n = 1000 print('building data...') _, d_inp_enc, d_inp_dec, d_out_dec, _ = master.dataset.feed_data('test', max_n=n, check_src=True) if type_ == 'resp': vec_u0 = master.model_encoder['S2S'].predict(d_inp_enc['ctxt']) vec_u1 = master.model_encoder['AE'].predict(d_inp_enc['resp']) elif type_ == 'stry': vec_u0 = master.model_encoder['AE'].predict(d_inp_enc['resp']) vec_u1 = master.model_encoder['AE'].predict(d_inp_enc['stry']) else: raise ValueError print('evaluating...') uu = np.linspace(0, 1, 11) NLL = [] for u in uu: latent = vec_u0 + u * np.ones(vec_u0.shape) * (vec_u1 - vec_u0) NLL_resp = master.model_decoder_tf.evaluate( [latent, d_inp_dec['resp']], d_out_dec['resp'], verbose=0) if type_ == 'resp': NLL_ = NLL_resp else: NLL_stry = master.model_decoder_tf.evaluate( [latent, d_inp_dec['stry']], d_out_dec['stry'], verbose=0) NLL_ = NLL_resp * (1. - u) + u * NLL_stry print('u = %.3f, NLL = %.3f'%(u, NLL_)) NLL.append(NLL_) fig = plt.figure(figsize=(6,3)) ax = fig.add_subplot(111) ax.plot(uu, NLL,'k.-') print(uu) print(NLL) ax.plot(0, NLL[0], 'ro') ax.plot(1, NLL[-1], 'bo') ax.text(0, NLL[0] + 0.5, ' '+r'$S$', color='r') ax.text(1, NLL[-1], ' '+r'$T$', color='b') plt.xlabel(r'$u$') plt.ylabel('NLL') plt.title(model_name+'\nNLL of interpolation: '+r'$S+u(T-S)$') plt.subplots_adjust(top=0.8) plt.subplots_adjust(bottom=0.2) plt.savefig(fld_save+'/interp_%s.png'%type_) with open(fld_save+'/interp_%s.tsv'%type_,'w') as f: f.write('\t'.join(['u'] + ['%.3f'%u for u in uu])+'\n') f.write('\t'.join(['NLL'] + ['%.3f'%l for l in NLL])+'\n') plt.show() def clusters(master, model_name, fld_save, D=2, use_bias=True, n_batch=1): n_sample = BATCH_SIZE * n_batch method = 'MDS' #method = 'tSNE' #method = 'isomap' latent_d = dict() colors = { 'base_conv': 'y', 'base_resp': 'r', 'bias_conv': 'k', 'bias_nonc': 'b', } print('building data...') d_inp_enc = master.dataset.feed_data('test', max_n=n_sample, check_src=True, mix_ratio=(0.,1.))['inp_enc'] latent_d['base_conv'] = master.model_encoder['S2S'].predict(d_inp_enc['ctxt']) if use_bias and 'AE' in master.prefix: latent_d['bias_nonc'] = master.model_encoder['AE'].predict(d_inp_enc['nonc']) #if use_bias and 'bias_conv' in master.dataset.files['test']: # d_inp_enc = master.dataset.feed_data('test', max_n=n_sample, check_src=True, mix_ratio=(1.,0.))['inp_enc'] # latent_d['bias_conv'] = master.model_encoder['S2S'].predict(d_inp_enc['ctxt']) #else: d_inp_enc = master.dataset.feed_data('test', max_n=n_sample, check_src=True, mix_ratio=(0.,0.))['inp_enc'] if 'AE' in master.prefix: #latent_d['base_nonc'] = master.model_encoder['AE'].predict(d_inp_enc['nonc']) latent_d['base_resp'] = master.model_encoder['AE'].predict(d_inp_enc['resp']) labels = list(sorted(latent_d.keys())) fname_suffix = args.restore.split('/')[-1].replace('.npz','') if use_bias: fname_suffix += '_wbias' n_labels = len(labels) latent = np.concatenate([latent_d[k] for k in labels], axis=0) print('latent.shape',latent.shape) print('plotting bit hist...') bins = np.linspace(-1,1,31) for k in latent_d: l = latent_d[k].ravel() freq, _, _ = plt.hist(l, bins=bins, color='w') plt.plot(bins[:-1], 100.*freq/sum(freq), colors[k]+'.-') plt.ylim([0,50]) plt.savefig(fld_save+'/hist_%s.png'%fname_suffix) plt.close() print('plotting dist mat...') d_norm = np.sqrt(latent.shape[1]) f, ax = plt.subplots() cax = ax.imshow(dist_mat(latent)/d_norm, cmap='bwr') #ax.set_title(model_name) f.colorbar(cax) ticks = [] ticklabels = [] n_prev = 0 for i in range(n_labels): ticks.append(n_prev + n_sample/2) ticklabels.append(labels[i]+'\n') ticks.append(n_prev + n_sample) ticklabels.append('%i'%(n_sample * (i+1))) n_prev = n_prev + n_sample ax.set_xticks(ticks) ax.set_xticklabels(ticklabels) ax.xaxis.tick_top() ax.set_yticks(ticks) ax.set_yticklabels([s.strip('\n') for s in ticklabels]) plt.savefig(fld_save+'/dist_%s.png'%fname_suffix) plt.close() if method == 'tSNE': approx = manifold.TSNE(init='pca', verbose=1).fit_transform(latent) elif method == 'MDS': approx = manifold.MDS(D, verbose=1, max_iter=500, n_init=1).fit_transform(latent) elif method == 'isomap': approx = manifold.Isomap().fit_transform(latent) else: raise ValueError f, ax = plt.subplots() for k in labels: ax.plot(np.nan, np.nan, colors[k]+'.', label=k) jj = list(range(approx.shape[0])) np.random.shuffle(jj) for j in jj: i_label = int(j/n_sample) ax.plot(approx[j, 0], approx[j, 1], colors[labels[i_label]]+'.') #plt.legend(loc='best') plt.title(model_name) #ax.set_xticks([]) #ax.set_yticks([]) plt.savefig(fld_save+'/%s_%s.png'%(method, fname_suffix)) plt.show() def cos_sim(a, b): #return 1. - scipy.spatial.distance.cosine(a, b) return np.inner(a, b)/np.linalg.norm(a)/np.linalg.norm(b) def angel_hist(master, model_name, fld_save): from rand_decode import load_1toN_data data = load_1toN_data(master.dataset.generator['test']) angel = [] n_sample = 1000 extra_info = [] for i in range(n_sample): if i%10 == 0: print(i) d = data[i] src_seq = np.reshape(d['src_seq'], [1,-1]) latent_src = np.ravel(master.model_encoder['dial'].predict(src_seq)) diff = [] for ref_seq in d['ref_seqs']: ref_seq = np.reshape(ref_seq, [1,-1]) latent_ref = np.ravel(master.model_encoder['auto'].predict(ref_seq)) diff.append(latent_ref - latent_src) for i in range(len(diff) - 1): for j in range(i + 1, len(diff)): if str(d['ref_seqs'][i]) == str(d['ref_seqs'][j]): continue angel.append(cos_sim(diff[i], diff[j])) extra_info.append('%i\t%i'%(i, len(d['ref_seqs']))) with open(fld_save+'/angel.txt', 'w') as f: f.write('\n'.join([str(a) for a in angel])) with open(fld_save+'/angel_extra.txt', 'w') as f: f.write('\n'.join(extra_info)) plt.hist(angel, bins=30) plt.title(model_name) plt.savefig(fld_save+'/angel.png') plt.show() def plot_history(paths, labels, k, ix=-1, ax=None): if isinstance(paths, str): paths = [paths] import matplotlib.pyplot as plt def MA(y): window = 30 ret = [np.nan] * len(y) for i in range(window, len(y)): ret[i] = np.mean(y[max(0, i - window + 1): i + 1]) return ret def read_log(path, k): trained = np.nan xx = [] yy = [[] for _ in range(4)] m = None for line in open(path): if line.startswith('***** trained '): trained = float(line.split(',')[0].split()[-2]) if line.startswith(k): vv = [float(v) for v in line.replace(':','=').split('=')[-1].split(',')] if m is None: m = len(vv) print('expecting %i values'%m) else: if m!=len(vv): continue xx.append(trained) for i in range(len(vv)): yy[i].append(vv[i]) return xx, yy[:m] if ax is None: _, ax = plt.subplots() color = ['r','b','k','m'] if len(paths) > 0: for i, path in enumerate(paths): xx, yy = read_log(path, k) ss = path.split('/') label = ss[-1].replace('.txt','') ax.plot(xx, yy[ix], color=color[i], linestyle=':', alpha=0.5) ax.plot(xx, MA(yy[ix]), color=color[i], label=labels[i]) ax.set_title(k + '[%i]'%ix) else: xx, yy = read_log(paths[0], k) for i in range(len(yy)): ax.plot(xx, yy[i], color=color[i], linestyle=':') ax.plot(xx, MA(yy[i]), color=color[i], label=str(i + 1)) def plot_multiple(kk, paths, labels): n_col = 4 n_row = int(len(kk)/n_col) n_row = int(np.ceil(len(kk)/n_col)) print('n_row = %i'%n_row) _, axs = plt.subplots(n_row, n_col, sharex=True) for i in range(len(kk)): k = kk[i] col = i%n_col row = int(i/n_col) ax = axs[row][col] if k.startswith('bleu') or k.startswith('corr'): ix = 2 else: ix = -1 plot_history(paths, labels, k, ix, ax=ax) #if i == 0: # ax.legend(loc='best') ax.grid(True) plt.show() if __name__ == '__main__': parser.add_argument('--vis_tp', default='clusters') parser.add_argument('--use_bias', type=int, default=1) parser.add_argument('--n_batch', type=int, default=5) args = parser.parse_args() print('>>>>> Not using GPU') os.environ["CUDA_VISIBLE_DEVICES"]="-1" master = run_master('vis', args) #if args.cpu_only: #fld = os.path.join(fld_model, model_name, 'vis') model_name = '' fld = master.fld + '/vis' print(fld) makedirs(fld) if args.vis_tp.startswith('interp'): if 'stry' in args.vis_tp: interp(master, model_name, fld, type_='stry') else: interp(master, model_name, fld, type_='resp') elif args.vis_tp == 'clusters': clusters(master, model_name, fld, use_bias=bool(args.use_bias), n_batch=args.n_batch) elif args.vis_tp == 'angel': angel_hist(master, model_name, fld) else: raise ValueError
9,316
25.850144
109
py
StyleFusion
StyleFusion-master/src/model.py
from shared import * from tf_lib import * from dataset import * from decode import * from evaluate import * """ AUTHOR: Xiang Gao (xiag@microsoft.com) at Microsoft Research """ class ModelBase: def __init__(self): self.fld = None # str self.n_trained = None # int self.max_n_trained = None # int self.dataset = None # Dataset obj self.extra = None # list of str self.vali_data = None # dict of list self.layers = None def init_log(self, new, args): # deal with existing fld if new and os.path.exists(self.fld): if PHILLY: suffix = 0 while True: fld = self.fld + '_%i'%suffix if not os.path.exists(fld): self.fld = fld break else: if not PHILLY and not self.debug: print('%s\nalready exists, do you want to delete the folder? (y/n)'%self.fld) ans = input() if not ans.lower() == 'y': exit() print('deleting fld: '+self.fld) shutil.rmtree(self.fld) time.sleep(0.1) print('fld deleted') self.log_train = self.fld + '/train.txt' if new or PHILLY or hostname != 'MININT-3LHNLKS': makedirs(os.path.join(self.fld, 'models')) open(self.log_train, 'w') if not os.path.exists(self.fld + '/vocab.txt'): shutil.copyfile(self.dataset.path_vocab, self.fld + '/vocab.txt') ss = [] for k in sorted(args.__dict__.keys()): ss.append('%s = %s'%(k, args.__dict__[k])) with open(self.fld + '/args.txt', 'w') as f: f.write('\n'.join(ss)) if PHILLY: with open(self.log_train, 'a') as f: f.write('hostname: %s\n'%hostname) f.write('data_path: %s\n'%DATA_PATH) f.write('out_path: %s\n'%OUT_PATH) def train(self, batch_per_load=100): self.vali() while self.n_trained < self.max_n_trained: s = '\n***** trained %.3f M'%(self.n_trained/1e6) for tp in self.dataset.n_reset['train']: s += ', %s = %i'%(tp, self.dataset.n_reset['train'][tp]) s += ' *****' write_log(self.log_train, s) self.train_a_load(batch_per_load) if self.debug: exit() def load_weights(self, path): self.prev_wt_fuse = None print('loading weights from %s'%path) npz = np.load(path, encoding='latin1', allow_pickle=True) print(npz.files) weights = npz['layers'].item() for k in weights: s = ' '*(20-len(k)) + k + ': %i params: '%len(weights[k]) for wt in weights[k]: s += str(wt.shape) + ', ' print(s) for attr in self.extra: if attr in npz: if attr not in ['name']: setattr(self, attr, npz[attr]) else: print('WARNING! attr %s not in npz'%attr) self.build_model(weights) self.build_model_test() def extract_weights(self): weights = dict() if self.layers is None: return weights for k in self.layers: weights[k] = self.layers[k].get_weights() return weights def save_weights(self): path = self.fld + '/models/%.1fM.npz'%(self.n_trained/1e6) weights = self.extract_weights() to_save = {'layers':weights} for attr in self.extra: to_save[attr] = getattr(self, attr) n_try = 0 while n_try < 3: try: np.savez(path, **to_save) print('saved to: '+path) break except: n_try += 1 print('cannot save, try %i'%n_try) return path def build_model_test(self): pass def build_model(self, weights=dict()): pass def train_a_load(self, batch_per_load): pass def set_extra(self, npz): pass class Seq2SeqBase(ModelBase): def __init__(self, dataset, fld, args, new=False, allowed_words=None): self.dataset = dataset self.fld = fld self.allowed_words = allowed_words self.layers = None self.history = LossHistory() self.vali_data = None self.classifiers = [] self.n_batch = 0 self.prev_n_batch = 0 self.dn_batch_vali = 100 self.bias_conv = False # hasattr(self.dataset, 'files') and ('bias_conv' in self.dataset.files['train']) self.debug = args.debug self.token_embed_dim = args.token_embed_dim self.rnn_units = args.rnn_units self.encoder_depth = args.encoder_depth self.decoder_depth = args.decoder_depth self.lr = args.lr self.max_n_trained = args.max_n_trained self.randmix = False self.mix_ratio = (args.conv_mix_ratio, args.nonc_mix_ratio) if not self.bias_conv: assert(args.conv_mix_ratio == 0.) self.extra = ['name'] self.init_extra(args) if hasattr(args, 'skip'): skip = int(1e6*args.skip) else: skip = 0 self.dataset.skip(skip, self.mix_ratio, conv_only=(self.name=='s2s')) self.n_trained = skip self.init_log(new, args) self.build_model() def get_mix_ratio(self): if self.randmix: ret = [] for ratio in self.mix_ratio: p = [1. - ratio, ratio] ret.append(np.random.choice([0.,1.], 1, p=p)[0]) return tuple(ret) else: return self.mix_ratio def fit(self, inputs, outputs): n_try = 0 if self.debug: self.model.fit( inputs, outputs, batch_size=BATCH_SIZE, callbacks=[self.history], verbose=FIT_VERBOSE) return while n_try < 3: try: self.model.fit( inputs, outputs, batch_size=BATCH_SIZE, callbacks=[self.history], verbose=FIT_VERBOSE) return except Exception as e: print('got error, sleeping') print('E'*20) print(e) print('E'*20) time.sleep(1) n_try += 1 def _stacked_rnn(self, rnns, inputs, initial_states=None): if initial_states is None: initial_states = [None] * len(rnns) outputs, state = rnns[0](inputs, initial_state=initial_states[0]) states = [state] for i in range(1, len(rnns)): outputs, state = rnns[i](outputs, initial_state=initial_states[i]) states.append(state) return outputs, states def _build_encoder(self, inputs, prefix): _, encoder_states = self._stacked_rnn( [self.layers['%s_encoder_rnn_%i'%(prefix, i)] for i in range(self.encoder_depth)], self.layers['embedding'](inputs)) latent = encoder_states[-1] return latent def _build_decoder(self, input_seqs, input_states): """ for auto-regressive, states are returned and used as input for the generation of the next token for teacher-forcing, token already given, so only need init states """ decoder_outputs, decoder_states = self._stacked_rnn( [self.layers['decoder_rnn_%i'%i] for i in range(self.decoder_depth)], self.layers['embedding'](input_seqs), input_states) decoder_outputs = self.layers['decoder_softmax'](decoder_outputs) return decoder_outputs, decoder_states def _create_layers(self, weights=dict()): layers = dict() name = 'embedding' params = _params(name, weights, {'mask_zero':True}) layers[name] = Embedding( self.dataset.num_tokens + 1, # +1 as mask_zero self.token_embed_dim, **params) for i in range(self.decoder_depth): name = 'decoder_rnn_%i'%i params = _params(name, weights, {'return_state':True, 'return_sequences':True}) layers[name] = GRU( self.rnn_units, **params) for prefix in self.prefix: for i in range(self.encoder_depth): name = '%s_encoder_rnn_%i'%(prefix, i) params = _params(name, weights, {'return_state':True, 'return_sequences':True}) layers[name] = GRU( self.rnn_units, **params) name = 'decoder_softmax' params = _params(name, weights, {'activation':'softmax'}) layers[name] = Dense( self.dataset.num_tokens + 1, # +1 as mask_zero **params) return layers def build_model_test(self): #self.refresh_session() decoder_inputs = Input(shape=(None,), name='decoder_inputs') # encoder self.model_encoder = dict() self.model_tf = dict() self.tf_history = dict() for prefix in self.prefix: encoder_inputs = Input(shape=(None,), name=prefix+'_encoder_inputs') latent = self._build_encoder(encoder_inputs, prefix=prefix) self.model_encoder[prefix] = Model(encoder_inputs, latent) self.model_encoder[prefix]._make_predict_function() decoder_outputs, _ = self._build_decoder(decoder_inputs, [latent]*self.decoder_depth) self.model_tf[prefix] = Model([encoder_inputs, decoder_inputs], decoder_outputs) for layer in self.model_tf[prefix].layers: layer.trainable = False self.model_tf[prefix].compile(Adam(lr=0.), loss=_dec_loss) # lr = 0 to use '.fit', which has callbacks, as '.evaluate' self.tf_history[prefix] = LossHistory() # decoder: autoregressive decoder_inital_states = [] for i in range(self.decoder_depth): decoder_inital_states.append(Input(shape=(self.rnn_units,), name="decoder_inital_state_%i"%i)) decoder_outputs, decoder_states = self._build_decoder(decoder_inputs, decoder_inital_states) model_decoder = Model( [decoder_inputs] + decoder_inital_states, [decoder_outputs] + decoder_states) model_decoder._make_predict_function() self.decoder = Decoder(self.dataset, model_decoder, self.decoder_depth, self.rnn_units, allowed_words=self.allowed_words) def get_vali_data(self): if self.vali_data is not None: #print('returning self.vali_data', self.vali_data) return self.vali_data print('getting vali data...') def _feed_vali(k): self.dataset.reset('vali') d = self.dataset.feed_data('vali', max_n=vali_size, check_src=True, mix_ratio=k, conv_only=(self.name=='s2s')) self.dataset.reset('vali') return d if self.debug: vali_size = BATCH_SIZE else: vali_size = 1000 self.vali_data = _feed_vali((0, 1)) """ self.vali_data['base'] = _feed_vali((0, 0)) self.vali_data['mix'] = _feed_vali(self.mix_ratio) if self.bias_conv: self.vali_data['bias'] = _feed_vali((1, 1)) else: self.vali_data['bias'] = _feed_vali((0, 1)) """ return self.vali_data def vali(self): self.build_model_test() ss = [] for inp in ['who is he ?', 'do you like this game ?', 'good morning .']: ss.append(infer_comb(inp, self)) write_log(self.log_train, '\n'.join(ss)) """ data = self.get_vali_data() if self.name.startswith('fuse'): r_rand = 0.1 * np.sqrt(self.rnn_units) else: r_rand = 0. #s_decoded = ''#eval_decoded(self, data, self.classifiers, r_rand=r_rand)[0] #s_surrogate = eval_surrogate(self, data)[0] #write_log(self.log_train, '\n' + s_decoded + '\n\n' + s_surrogate + '\n') """ self.prev_n_batch = self.n_batch # save -------------------- self.save_weights() def init_extra(self, args): pass def train_a_load(self, batch_per_load): mix_ratio = self.get_mix_ratio() data = self.dataset.feed_data('train', BATCH_SIZE * batch_per_load, mix_ratio=mix_ratio, conv_only=(self.name == 's2s')) n_sample, inputs, outputs = self._inp_out_data(data) t0 = datetime.datetime.now() t0_str = str(t0).split('.')[0] write_log(self.log_train, 'start: %s'%t0_str + ', mix_ratio = '+str(mix_ratio)) print('fitting...') self.fit(inputs, outputs) self.n_trained += n_sample self.n_batch += batch_per_load dt = (datetime.datetime.now() - t0).seconds loss = np.mean(self.history.losses) write_log(self.log_train, 'n_batch: %i, prev %i'%(self.n_batch, self.prev_n_batch)) ss = ['spent: %i sec'%dt, 'train: %.4f'%loss] write_log(self.log_train, '\n'.join(ss)) if not self.debug and (self.n_batch - self.prev_n_batch < self.dn_batch_vali): return # vali -------------------- self.vali() def print_loss(self, loss_weights): s = 'loss: '+'-'*20 + '\n' for i in range(len(self.loss)): loss_name = str(self.loss[i]) if loss_name.startswith('<func'): loss_name = loss_name.split()[1] s += '%6.2f '%loss_weights[i] + loss_name + '\n' s += '-'*20 + '\n' write_log(self.log_train, s) class Seq2Seq(Seq2SeqBase): def init_extra(self, args): self.name = 's2s' self.prefix = ['S2S'] def build_model(self, weights=dict()): self.layers = self._create_layers(weights) # create new encoder_inputs = Input(shape=(None,), name='encoder_inputs') decoder_inputs = Input(shape=(None,), name='decoder_inputs') # connections: teacher forcing latent = self._build_encoder(encoder_inputs, self.prefix[0]) decoder_outputs, _ = self._build_decoder(decoder_inputs, [latent]*self.decoder_depth) # models self.model = Model( [encoder_inputs, decoder_inputs], # [input sentences, ground-truth target sentences], decoder_outputs) # shifted ground-truth sentences self.model.compile(Adam(lr=self.lr), loss=_dec_loss) def _inp_out_data(self, data): inputs = [data['inp_enc']['ctxt'], data['inp_dec']['resp']] outputs = data['out_dec']['resp'] return data['n_sample'], inputs, outputs class VanillaMTask(Seq2SeqBase): def init_extra(self, args): self.name = 'mtask' self.loss = [ _dec_loss, # logP(resp | S2S), just the seq2seq loss _dec_loss, # logP(resp | AE_resp) _dec_loss, # logP(resp | AE_nonc) ] self.prefix = ['AE','S2S'] def build_model(self, weights=dict()): loss_weights = [1., 0.5, 0.5] self.layers = self._create_layers(weights) # create new # inputs inp_enc_ctxt = Input(shape=(None,), name='inp_enc_ctxt') inp_enc_resp = Input(shape=(None,), name='inp_enc_resp') inp_dec_resp = Input(shape=(None,), name='inp_dec_resp') inp_enc_nonc = Input(shape=(None,), name='inp_enc_nonc') inp_dec_nonc = Input(shape=(None,), name='inp_dec_nonc') inps_enc = [inp_enc_ctxt, inp_enc_resp, inp_enc_nonc] inps_dec = [inp_dec_resp, inp_dec_nonc] inputs = inps_enc + inps_dec # hiddens vec_s2s = self._build_encoder(inp_enc_ctxt, prefix='S2S') vec_ae_resp = self._build_encoder(inp_enc_resp, prefix='AE') vec_ae_nonc = self._build_encoder(inp_enc_nonc, prefix='AE') # outputs out_s2s, _ = self._build_decoder(inp_dec_resp, [vec_s2s]*self.decoder_depth) out_ae_resp, _ = self._build_decoder(inp_dec_nonc, [vec_ae_resp]*self.decoder_depth) out_ae_nonc, _ = self._build_decoder(inp_dec_nonc, [vec_ae_nonc]*self.decoder_depth) outputs = [out_s2s, out_ae_resp, out_ae_nonc] # compile self.print_loss(loss_weights) self.model = Model(inputs, outputs) self.model.compile(Adam(lr=self.lr), loss=self.loss, loss_weights=loss_weights) def _inp_out_data(self, data, u=None): n_sample = data['n_sample'] if n_sample == 0: return n_sample, [], [] inps_enc = [data['inp_enc']['ctxt'], data['inp_enc']['resp'], data['inp_enc']['nonc']] inps_dec = [data['inp_dec']['resp'], data['inp_dec']['nonc']] outs_dec = [data['out_dec']['resp'], data['out_dec']['resp'], data['out_dec']['nonc']] return n_sample, inps_enc + inps_dec, outs_dec class StyleFusion(Seq2SeqBase): def init_extra(self, args): self.name = args.model_class.lower() assert(self.name in ['fuse','fuse1']) self.max_wt_dist = args.wt_dist self.stddev = args.stddev self.v1 = (self.name == 'fuse1') self.ablation = args.ablation if self.v1: # roughly, not exactly, follow SpaceFusion v1, as in https://arxiv.org/abs/1902.11205 _dec_loss_ae = _dec_loss _dist_loss = _absdiff_dist_v1 else: # v2, consider fuse with nonc _dec_loss_ae = _dec_loss_u # interp(ae_resp, ae_nonc) if args.reld: _dist_loss = _relative_dist # consider all these terms d(s2s,resp), d(s2s,nonc), d(resp), d(nonc), d(s2s) else: _dist_loss = _absdiff_dist self.randmix = True # binary batch mix self.loss = [ _dec_loss, # logP(resp | S2S), just the seq2seq loss _dec_loss, # logP(resp | interp), interp is between ctxt and resp, i.e. the 3rd term in Eq.3 in NAACL _dec_loss_ae, _dist_loss] self.prefix = ['AE','S2S'] """ def refresh_session(self): K.clear_session() # avoid building graph over and over to slow down everything config = tf.ConfigProto() config.gpu_options.allow_growth = True K.set_session(tf.Session(config=config)) for clf in self.classifiers: clf.load() """ def build_model(self, weights=dict()): loss_weights = [1., 1., 1., 1.] if self.ablation: loss_weights = [1., 1., 0., 1.] # disable L_{smooth,style} self.layers = self._create_layers(weights) # create new noisy = Lambda(_add_noise, arguments={'stddev':self.stddev}, name='noisy') concat = Concatenate(name='concat_1', axis=-1) # inputs inp_enc_ctxt = Input(shape=(None,), name='inp_enc_ctxt') inp_enc_resp = Input(shape=(None,), name='inp_enc_resp') inp_dec_resp = Input(shape=(None,), name='inp_dec_resp') inps_enc = [inp_enc_ctxt, inp_enc_resp] inps_dec = [inp_dec_resp] inp_enc_nonc = Input(shape=(None,), name='inp_enc_nonc') inp_dec_nonc = Input(shape=(None,), name='inp_dec_nonc') inps_enc.append(inp_enc_nonc) inps_dec.append(inp_dec_nonc) inp_u = [Input(shape=(None,), name='inp_u')] # rand drawn from U(0,1). each batch has the same value, see _inp_out_data inputs = inps_enc + inps_dec + inp_u # match _inp_out_data # hiddens vec_s2s = self._build_encoder(inp_enc_ctxt, prefix='S2S') vec_ae_resp = self._build_encoder(inp_enc_resp, prefix='AE') vec_ae_nonc = self._build_encoder(inp_enc_nonc, prefix='AE') vec_interp_resp = noisy(Lambda(_interp, name='interp_resp')([vec_s2s, vec_ae_resp] + inp_u)) # outputs out_s2s, _ = self._build_decoder(inp_dec_resp, [vec_s2s]*self.decoder_depth) out_interp_resp, _ = self._build_decoder(inp_dec_resp, [vec_interp_resp]*self.decoder_depth) if self.v1: out_ae, _ = self._build_decoder(inp_dec_nonc, [vec_ae_nonc]*self.decoder_depth) else: vec_interp_ae = noisy(Lambda(_interp, name='interp_ae')([vec_ae_resp, vec_ae_nonc] + inp_u)) out_interp_ae_resp, _ = self._build_decoder(inp_dec_resp, [vec_interp_ae]*self.decoder_depth) out_interp_ae_nonc, _ = self._build_decoder(inp_dec_nonc, [vec_interp_ae]*self.decoder_depth) out_ae = concat([out_interp_ae_resp, out_interp_ae_nonc]) outs_dec = [out_s2s, out_interp_resp, out_ae] outs_dist = concat([vec_s2s, vec_ae_resp, vec_ae_nonc]) outputs = outs_dec + [outs_dist] # compile self.print_loss(loss_weights) self.model = Model(inputs, outputs) self.model.compile(Adam(lr=self.lr), loss=self.loss, loss_weights=loss_weights) def _inp_out_data(self, data, u=None): n_sample = data['n_sample'] if n_sample == 0: return n_sample, [], [] if u is None: u = np.random.random(n_sample) else: u = np.array([u] * n_sample) inps_enc = [data['inp_enc']['ctxt'], data['inp_enc']['resp']] inps_dec = [data['inp_dec']['resp']] outs_dec = [data['out_dec']['resp'], data['out_dec']['resp']] inps_enc.append(data['inp_enc']['nonc']) inps_dec.append(data['inp_dec']['nonc']) inputs = inps_enc + inps_dec + [u] if self.v1: outs_dec.append(data['out_dec']['nonc']) else: _, l, v = data['out_dec']['resp'].shape out_interp_nonc = np.zeros([n_sample, l, v*2+1]) out_interp_nonc[:,:,:v] = data['out_dec']['resp'] out_interp_nonc[:,:,v:v*2] = data['out_dec']['nonc'] for t in range(l): out_interp_nonc[:,t,-1] = u outs_dec.append(out_interp_nonc) outputs = outs_dec + [np.zeros((n_sample, 1))] return n_sample, inputs, outputs class LossHistory(Callback): def reset(self): self.losses = [] def on_train_begin(self, logs={}): self.reset() def on_batch_end(self, batch, logs={}): self.losses.append(logs.get('loss')) def _params(name, weights, extra=dict()): params = {'name':name} if name in weights: params['weights'] = weights[name] for k in extra: params[k] = extra[k] return params def write_log(path, s, PRINT=True, mode='a'): if PRINT: print(s) sys.stdout.flush() if not s.endswith('\n'): s += '\n' if PHILLY: n_try = 0 while n_try < 3: try: with open(path, mode) as f: f.write(s) break except:# PermissionError as e: #print(e) print('cannot write_log, sleeping...') time.sleep(2) n_try += 1 else: with open(path, mode) as f: f.write(s) # ------------------- customized loss -------------------- def _dist_1nn(a, b=None): n = BATCH_SIZE expanded_a = tf.expand_dims(a, 1) if b is None: b = a expanded_b = tf.expand_dims(b, 0) d_squared = tf.reduce_mean(tf.squared_difference(expanded_a, expanded_b), 2) mat = tf.sqrt(tf.maximum(0., d_squared)) wt = 1./(mat + tf.eye(n) * 1000 + 1e-6) sum_wt = tf.reshape(tf.reduce_sum(wt, axis=1), [n, 1]) sum_wt = tf.tile(sum_wt, [1,n]) wt = wt/sum_wt d1nn = tf.reduce_sum(mat * wt, axis=1) d1nn = tf.reduce_mean(d1nn) return d1nn def _cross_inner(vecs, v1=False): def sqrt_mse(a, b=None, shuffle=True, cap=None): if b is None: b = a if shuffle: #diff = a - tf.random_shuffle(b) _, d = a.shape n = BATCH_SIZE - 1 diff = tf.slice(a, [1,0], [n,d]) - tf.slice(b, [0,0], [n,d]) else: diff = a - b squared = tf.pow(diff, 2) if cap is not None: squared = tf.minimum(cap**2, squared) return tf.sqrt(tf.reduce_mean(squared)) vec_s2s, vec_ae_resp, vec_ae_nonc = tf.split(vecs, 3, axis=-1) cross_resp = sqrt_mse(vec_s2s, vec_ae_resp, shuffle=False) inner_s2s_resp = _dist_1nn(vec_s2s) inner_ae_nonc = _dist_1nn(vec_ae_nonc) if v1: print('*'*10 + ' [WARNING] Using v1 cross_inner ' + '*'*10) return cross_resp, inner_s2s_resp + inner_ae_nonc else: cross_s2s_nonc = _dist_1nn(vec_s2s, vec_ae_nonc) inner_ae_resp = _dist_1nn(vec_ae_resp) cross = 0.5 * (cross_resp + cross_s2s_nonc) inner = tf.minimum(tf.minimum(inner_s2s_resp, inner_ae_resp), inner_ae_nonc) return cross, inner def _relative_dist(_, y_pred): cross, inner = _cross_inner(y_pred) return cross / inner def _absdiff_dist(_, y_pred): cross, inner = _cross_inner(y_pred) return cross - inner def _absdiff_dist_v1(_, y_pred): cross, inner = _cross_inner(y_pred, v1=True) return cross - inner def _dec_loss(y_true, y_pred): # to compute - logP(resp|vec_interp_resp) return tf.reduce_mean(keras.losses.categorical_crossentropy(y_true, y_pred)) def _dec_loss_u(y_true, y_pred): # to compute u * logP(resp|vec_interp_ae) + (1-u) * logP(nonc|vec_interp_ae) # where vec_interp_ae = u * vec_resp_ae + (1-u) * vec_nonc_ae # y_true = concat([y_resp, y_nonc, u]), shape = [BATCH_SIZE, seq_len, 2 * vocab_size + 1], see out_interp_nonc in _in_out_data # y_pred = concat([y_resp_pred, y_nonc_pred]) y_resp_pred, y_nonc_pred = tf.split(y_pred, 2, axis=-1) vocab_size = tf.cast(y_resp_pred.shape[2], tf.int32) y_resp, y_nonc, u = tf.split(y_true, [vocab_size, vocab_size, 1], axis=-1) u = u[:,:,0] # like tf.squeeze, so [BATCH_SIZE, seq_len] loss_resp = keras.losses.categorical_crossentropy(y_resp, y_resp_pred) # [BATCH_SIZE, seq_len] loss_nonc = keras.losses.categorical_crossentropy(y_nonc, y_nonc_pred) loss = u * loss_resp + (1. - u) * loss_nonc # [BATCH_SIZE, seq_len] return tf.reduce_mean(loss) # ------------------- customized layers -------------------- def _add_noise(mu, stddev): eps = K.random_normal(shape=K.shape(mu)) return mu + tf.multiply(eps, stddev) def _interp(inp): if len(inp) == 2: a, b = inp u = K.random_uniform(shape=(K.shape(a)[0], 1)) else: a, b, u = inp u = K.tile(K.reshape(u, [-1,1]), [1, K.shape(a)[1]]) # repeat along axis=1 #return a + tf.multiply(b - a, u) return tf.multiply(a, u) + tf.multiply(b, 1 - u) def convert_model_vocab(path_npz_old, path_npz_new, path_vocab_old, path_vocab_new): if os.path.exists(path_npz_new): print('already exists: '+path_npz_new) return _, token2index_old = load_vocab(path_vocab_old) index2token_new, _ = load_vocab(path_vocab_new) n_old = max(token2index_old.values()) + 1 n_new = max(index2token_new.keys()) + 1 print('vocab: %i => %i'%(n_old, n_new)) new2old = dict() ix_unk_old = token2index_old[UNK_token] for ix in index2token_new: token = index2token_new[ix] new2old[ix] = token2index_old.get(token, ix_unk_old) print('loading from: '+str(path_npz_old)) npz = np.load(path_npz_old, encoding='latin1') weights = npz['layers'].item() embedding_old = weights['embedding'][0] softmax_wt_old = weights['decoder_softmax'][0] softmax_bias_old = weights['decoder_softmax'][1] n_old_loaded, dim = embedding_old.shape assert(n_old_loaded == n_old) embedding_new = np.zeros((n_new, dim)) softmax_wt_new = np.zeros((dim, n_new)) softmax_bias_new = np.zeros((n_new,)) print(' embedding: ' + str(embedding_old.shape) + ' => ' + str(embedding_new.shape)) print(' softmax_wt: ' + str(softmax_wt_old.shape) + ' => ' + str(softmax_wt_new.shape)) print('softmax_bias: ' + str(softmax_bias_old.shape) + ' => ' + str(softmax_bias_new.shape)) # PAD embedding_new[0,:] = embedding_old[0, :] softmax_wt_new[:, 0] = softmax_wt_old[:, 0] softmax_bias_new[0] = softmax_bias_old[0] for ix in index2token_new: embedding_new[ix, :] = embedding_old[new2old[ix], :] softmax_wt_new[:, ix] = softmax_wt_old[:, new2old[ix]] softmax_bias_new[ix] = softmax_bias_old[new2old[ix]] weights['embedding'] = [embedding_new] weights['decoder_softmax'] = [softmax_wt_new, softmax_bias_new] print('saving to: '+str(path_npz_new)) to_save = {'layers':weights} for k in npz.files: if k != 'layers' and 'mix' not in k: to_save[k] = npz[k] np.savez(path_npz_new, **to_save)
24,831
28.632458
127
py
StyleFusion
StyleFusion-master/src/tf_lib.py
from keras.models import Model, load_model, model_from_yaml from keras.layers import Input, GRU, Dense, Embedding, Dropout, Concatenate, Lambda, Add, Subtract, Multiply, GaussianNoise from keras.utils import plot_model from keras.callbacks import ModelCheckpoint from keras.optimizers import Adam, RMSprop from keras.callbacks import Callback from keras import backend as K import tensorflow as tf from keras.activations import hard_sigmoid import keras
455
37
123
py
StyleFusion
StyleFusion-master/src/classifier.py
from shared import * from tf_lib import * import json from dataset import load_vocab from sklearn import linear_model import pickle """ AUTHOR: Sean Xiang Gao (xiag@microsoft.com) at Microsoft Research """ class ClassifierNeural(): def __init__(self, fld): params = json.load(open(fld + '/args.json')) if params['tgt_only']: self.prefix = ['tgt'] else: self.prefix = ['src','tgt'] self.encoder_depth = params['encoder_depth'] self.rnn_units = params['rnn_units'] self.mlp_depth = params['mlp_depth'] self.mlp_units = params['mlp_units'] self.include_punc = params['include_punc'] self.index2token, self.token2index = load_vocab(fld + '/vocab.txt') self.fld = fld self.load() def load(self): self.build_model() self.model.load_weights(self.fld+'/model.h5') def _create_layers(self): layers = dict() layers['embedding'] = Embedding( max(self.index2token.keys()) + 1, # +1 as mask_zero self.rnn_units, mask_zero=True, name='embedding') for prefix in self.prefix: for i in range(self.encoder_depth): name = '%s_encoder_rnn_%i'%(prefix, i) layers[name] = GRU( self.rnn_units, return_state=True, return_sequences=True, name=name) for i in range(self.mlp_depth - 1): name = 'mlp_%i'%i layers[name] = Dense( self.mlp_units, activation='tanh', name=name) name = 'mlp_%i'%(self.mlp_depth - 1) layers[name] = Dense(1, activation='sigmoid', name=name) return layers def _stacked_rnn(self, rnns, inputs, initial_states=None): if initial_states is None: initial_states = [None] * len(rnns) outputs, state = rnns[0](inputs, initial_state=initial_states[0]) states = [state] for i in range(1, len(rnns)): outputs, state = rnns[i](outputs, initial_state=initial_states[i]) states.append(state) return outputs, states def _build_encoder(self, inputs, layers, prefix): _, encoder_states = self._stacked_rnn( [layers['%s_encoder_rnn_%i'%(prefix, i)] for i in range(self.encoder_depth)], layers['embedding'](inputs)) latent = encoder_states[-1] return latent def build_model(self): layers = self._create_layers() encoder_inputs = dict() latents = [] for prefix in self.prefix: encoder_inputs[prefix] = Input(shape=(None,), name=prefix+'_encoder_inputs') latents.append(self._build_encoder(encoder_inputs[prefix], layers, prefix=prefix)) if len(self.prefix) > 1: out = Concatenate()(latents) inp = [encoder_inputs['src'], encoder_inputs['tgt']] else: out = latents[0] inp = encoder_inputs[self.prefix[0]] for i in range(self.mlp_depth): out = layers['mlp_%i'%i](out) self.model = Model(inp, out) self.model.compile(optimizer=Adam(lr=0), loss='binary_crossentropy') def txt2seq(self, txt): tokens = txt.strip().split(' ') seq = [] ix_unk = self.token2index[UNK_token] for token in tokens: if self.include_punc or is_word(token): # skip punctuation if necessary seq.append(self.token2index.get(token, ix_unk)) return seq def seq2txt(self, seq): return ' '.join([self.index2token[i] for i in seq]) def txts2mat(self, txts, max_len=30): if isinstance(txts, str): txts = [txts] data = np.zeros((len(txts), max_len)) for j, txt in enumerate(txts): seq = self.txt2seq(txt.strip(EOS_token).strip()) # stripped EOS_token here for t in range(min(max_len, len(seq))): data[j, t] = seq[t] return data def predict(self, txts): mat = self.txts2mat(txts) return self.model.predict(mat).ravel() class ClassifierNgram: def __init__(self, fld, ngram, include_punc=False): self.fld = fld self.ngram2ix = dict() self.ngram = ngram self.include_punc = include_punc fname = '%igram'%ngram if include_punc: fname += '.include_punc' self.path_prefix = fld + '/' + fname for i, line in enumerate(open(self.path_prefix + '.txt', encoding='utf-8')): ngram = line.strip('\n') self.ngram2ix[ngram] = i assert(self.ngram == len(ngram.split())) self.vocab_size = i + 1 print('loaded %i %igram'%(self.vocab_size, self.ngram)) #self.model = LogisticRegression(solver='sag')#, max_iter=10) self.model = linear_model.SGDClassifier(loss='log', random_state=9, max_iter=1, tol=1e-3) def txts2mat(self, txts): X = np.zeros((len(txts), self.vocab_size)) for i, txt in enumerate(txts): ww = txt2ww(txt, self.include_punc) for t in range(self.ngram, len(ww) + 1): ngram = ' '.join(ww[t - self.ngram: t]) j = self.ngram2ix.get(ngram, None) if j is not None: X[i, j] = 1. return X def load(self): self.model = pickle.load(open(self.path_prefix + '.p', 'rb')) def predict(self, txts): data = self.txts2mat(txts) prob = self.model.predict_proba(data) return prob[:,1] class ClassifierNgramEnsemble: def __init__(self, fld, include_punc=False, max_ngram=4): self.fld = fld self.children = dict() self.wt = dict() for ngram in range(1, max_ngram + 1): self.children[ngram] = ClassifierNgram(fld, ngram, include_punc) self.children[ngram].load() acc = float(open(self.children[ngram].path_prefix + '.acc').readline().strip('\n')) self.wt[ngram] = 2. * max(0, acc - 0.5) def predict(self, txts): avg_scores = np.array([0.] * len(txts)) for ngram in self.children: scores = self.children[ngram].predict(txts) avg_scores += scores * self.wt[ngram] return avg_scores / sum(self.wt.values()) def is_word(token): for c in token: if c.isalpha(): return True return False def load_classifier(fld, args=None): if fld.endswith('ngram'): return ClassifierNgramEnsemble(fld) elif fld.endswith('neural'): return ClassifierNeural(fld) else: raise ValueError def clf_interact(fld): clf = load_classifier(fld) while True: print('\n---- please input ----') txt = input() if txt == '': break score = clf.predict([txt])[0] print('%.4f'%score) def clf_eval(clf_fld, path): # path is a tsv, last col is hyp clf = load_classifier(clf_fld) sum_score = 0 n = 0 for line in open(path, encoding='utf-8'): txt = line.strip('\n').split('\t')[-1].lower() sum_score += clf.predict([txt])[0] n += 1 if n % 100 == 0: print('eval %i lines'%n) print('finally %i samples'%n) print('avg style score: %.4f'%(sum_score/n)) def txt2ww(txt, include_punc): ww = [SOS_token] for w in txt.split(): if include_punc or is_word(w): ww.append(w) ww.append(EOS_token) return ww def score_file(path, name, col=1): clf = load_classifier(name) txts = [] for line in open(path, encoding='utf-8'): txts.append(line.strip('\n').split('\t')[col]) if len(txts) == 1500: break print('scoring...') print(np.mean(clf.predict(txts))) class Classifier1gramCount: def __init__(self, fld): self.fld = fld def fit(self, min_freq=60, max_n=1e5): scores = dict() n = 0 for line in open(self.fld + '/all.txt', encoding='utf-8'): n += 1 cells = line.strip('\n').split('\t') if len(cells) != 2: print(cells) exit() txt, score = cells for w in set(txt.strip().split()): if is_word(w): if w not in scores: scores[w] = [] scores[w].append(float(score)) if n == max_n: break lines = ['\t'.join(['word', 'avg', 'se', 'count'])] for w in scores: count = len(scores[w]) if count < min_freq: continue avg = np.mean(scores[w]) se = np.std(scores[w])/np.sqrt(count) lines.append('\t'.join([w, '%.4f'%avg, '%.4f'%se, '%i'%count])) with open(self.fld + '/count.tsv', 'w', encoding='utf-8') as f: f.write('\n'.join(lines)) def load(self): self.coef = dict() f = open(self.fld + '/count.tsv', encoding='utf-8') header = f.readline() for line in f: w, avg = line.strip('\n').split('\t')[:2] self.coef[w] = float(avg) def corpus_score(self, txts, kw=100): scores = [] coef_w = [] for w in self.coef: coef_w.append((self.coef[w], w)) coef_w = sorted(coef_w, reverse=True)[:kw] print('last:',coef_w[-1]) keywords = set([w for _, w in coef_w]) #total_joint = 0 #total = 0 for txt in txts: words = set() for w in txt.strip().split(): if is_word(w): words.add(w) joint = words & keywords scores.append(len(joint)/len(words)) #total_joint += len(joint) #total += len(words) return np.mean(scores), np.std(scores)/np.sqrt(len(scores)) #return total_joint/total def test(self, kw=100): import matplotlib.pyplot as plt txts = [] labels = [] for line in open(self.fld + '/sorted_avg.tsv', encoding='utf-8'): txt, label = line.strip('\n').split('\t') txts.append(txt) labels.append(float(label)) i0 = 0 human = [] pred = [] while True: i1 = i0 + 100 if i1 >= len(txts): break human.append(np.mean(labels[i0:i1])) pred.append(self.corpus_score(txts[i0:i1], kw=kw)) i0 = i1 plt.plot(human, pred, '.') plt.xlabel('human') plt.xlabel('metric (ratio of keywords)') plt.title('corr = %.4f'%np.corrcoef(human, pred)[0][1]) plt.savefig(self.fld + '/test_corr_kw%i.png'%kw) if __name__ == '__main__': # e.g. `python src/classifier.py classifier/Reddit_vs_arXiv/neural' for interaction # e.g. `python src/classifier.py classifier/Reddit_vs_arXiv/neural path/to/hyp/file.tsv' for evaluating a file fld_model = sys.argv[1] # e.g. if len(sys.argv) == 2: clf_interact(fld_model) elif len(sys.argv) == 3: path_hyp = sys.argv[2] clf_eval(fld_model, path_hyp)
10,493
27.594005
114
py
StyleFusion
StyleFusion-master/src/dataset.py
from shared import * """ AUTHOR: Sean Xiang Gao (xiag@microsoft.com) at Microsoft Research """ def load_vocab(path): with io.open(path, encoding='utf-8') as f: lines = f.readlines() index2token = dict() token2index = dict() for i, line in enumerate(lines): token = line.strip('\n').strip() index2token[i + 1] = token # start from 1, as 0 reserved for PAD token2index[token] = i + 1 assert(SOS_token in token2index) assert(EOS_token in token2index) assert(UNK_token in token2index) return index2token, token2index class Dataset: def __init__(self, fld_data, max_ctxt_len=93, max_resp_len=30, vocab_only=False, noisy_vocab=-1, noisy_AE_src=True, noisy_bias=True, # whether add UNK noise to bias data (conv and nonc, src and tgt) ): self.max_ctxt_len = max_ctxt_len self.max_resp_len = max_resp_len self.noisy_vocab = noisy_vocab self.noisy_AE_src = noisy_AE_src self.noisy_bias = noisy_bias types = ['base_conv','bias_conv', 'base_nonc', 'bias_nonc'] self.fld_data = fld_data self.path_vocab = fld_data + '/vocab.txt' self.index2token, self.token2index = load_vocab(self.path_vocab) self.num_tokens = len(self.token2index) # not including 0-th if self.noisy_vocab > 0: self.prob_keep = dict() for ix in self.index2token: self.prob_keep[ix] = np.exp(-ix/self.noisy_vocab) if vocab_only: return self.paths = dict() self.files = dict() self.n_reset = dict() for sub in ['train', 'vali', 'test']: self.paths[sub] = dict() self.files[sub] = dict() self.n_reset[sub] = dict() for tp in types: self.n_reset[sub][tp] = -1 self.paths[sub][tp] = fld_data + '/%s_%s.num'%(tp, sub) self.reset(sub, tp) for k in self.files: print(k, self.files[k].keys()) def reset(self, sub, tp=None): if tp is None: types = self.files[sub].keys() else: types = [tp] for tp in types: if os.path.exists(self.paths[sub][tp]): line = open(self.paths[sub][tp]).readline().strip('\n') if len(line) > 0: self.files[sub][tp] = open(self.paths[sub][tp]) self.n_reset[sub][tp] += 1 def seq2txt(self, seq): words = [] for j in seq: if j == 0: # skip PAD continue words.append(self.index2token[int(j)]) return ' '.join(words) def txt2seq(self, text): tokens = text.strip().split() seq = [] for token in tokens: seq.append(self.token2index.get(token, self.token2index[UNK_token])) return seq def seqs2enc(self, seqs, max_len): inp = np.zeros((len(seqs), max_len)) for i, seq in enumerate(seqs): for t in range(min(max_len, len(seq))): inp[i, t] = seq[t] return inp def seqs2dec(self, seqs, max_len): # len: +2 as will 1) add EOS and 2) shift to right by 1 time step # vocab: +1 as mask_zero (token_id == 0 means PAD) ix_SOS = self.token2index[SOS_token] ix_EOS = self.token2index[EOS_token] inp = np.zeros((len(seqs), max_len + 2)) out = np.zeros((len(seqs), max_len + 2, self.num_tokens + 1)) for i, seq in enumerate(seqs): seq = seq[:min(max_len, len(seq))] for t, token_index in enumerate(seq): inp[i, t + 1] = token_index # shift 1 time step out[i, t, token_index] = 1. inp[i, 0] = ix_SOS # inp starts with EOS out[i, len(seq), ix_EOS] = 1. # out ends with EOS return inp, out def skip(self, max_n, mix_ratio, conv_only=False): sub = 'train' if isinstance(mix_ratio, int) or isinstance(mix_ratio, float): mix_ratio = (mix_ratio,) def _read(tp, n, m): for _ in self.files[sub][tp]: if m >= n: break m += 1 if m%1e5 == 0: print('%s skipped %.2f M'%(tp, m/1e6)) return m m = dict() suffix = ['conv'] if not conv_only: suffix.append('nonc') for i in range(len(suffix)): suf = suffix[i] for tp, n in [ ('base_'+suf, max_n * (1. - mix_ratio[i])), ('bias_'+suf, max_n * mix_ratio[i]) ]: m[tp] = 0 if n < 1 or tp not in self.files[sub]: continue while m[tp] < n: m_ = _read(tp, n, m[tp]) if m_ == m[tp]: self.reset(sub, tp) m[tp] = m_ if m_ >= n: break print('conv skipped %.2f M'%((m['base_conv'] + m['bias_conv'])/1e6)) if not conv_only: print('nonc skipped %.2f M'%((m['base_nonc'] + m['bias_nonc'])/1e6)) def add_unk_noise(self, seqs): if self.noisy_vocab < 0 or len(seqs) == 0: return seqs ix_unk = self.token2index[UNK_token] ret = [] n = 0 old_n_unk = 0 new_n_unk = 0 for seq in seqs: noisy = [] n += len(seq) for ix in seq: old_n_unk += (ix == ix_unk) if np.random.random() > self.prob_keep[ix]: noisy.append(ix_unk) else: noisy.append(ix) new_n_unk += (noisy[-1] == ix_unk) ret.append(noisy) print('unk increased from %.2f to %.2f'%(old_n_unk/n, new_n_unk/n)) return ret def feed_data(self, sub, max_n, check_src=False, mix_ratio=(0.,0.), conv_only=False): if isinstance(mix_ratio, int) or isinstance(mix_ratio, float): mix_ratio = (mix_ratio,) print('loading data, check_src = %s, mix_ratio = %s'%(check_src, mix_ratio)) # load conversation data ------------- def _read_conv(tp, n, prev_ctxt, seqs): for line in self.files[sub][tp]: if len(seqs) >= n: break tt = line.strip('\n').split('\t') if len(tt) != 2: continue seq_ctxt, seq_resp = tt if check_src and (seq_ctxt == prev_ctxt): continue prev_ctxt = seq_ctxt seq_ctxt = [int(k) for k in seq_ctxt.split()] seq_resp = [int(k) for k in seq_resp.split()] seq_ctxt = seq_ctxt[-min(len(seq_ctxt), self.max_ctxt_len):] seq_resp = seq_resp[:min(len(seq_resp), self.max_resp_len)] seqs.append((seq_ctxt, seq_resp)) return seqs, prev_ctxt # get conv from different tp seqs = dict() for tp, n in [('base_conv', max_n * (1. - mix_ratio[0])), ('bias_conv', max_n * mix_ratio[0])]: seqs[tp] = [] if n < 1 or tp not in self.files[sub]: continue prev_ctxt = '' while True: m = len(seqs[tp]) seqs[tp], prev_ctxt = _read_conv(tp, n, prev_ctxt, seqs[tp]) if len(seqs[tp]) >= n: break if len(seqs[tp]) == m: self.reset(sub, tp) print('conv from %s: %i/%i'%(tp, len(seqs[tp]), n)) if 'bias_conv' in seqs and self.noisy_bias: seqs_ctxt = self.add_unk_noise([seq for seq, _ in seqs['bias_conv']]) seqs_resp = self.add_unk_noise([seq for _, seq in seqs['bias_conv']]) seqs['bias_conv'] = [(seqs_ctxt[i], seqs_resp[i]) for i in range(len(seqs['bias_conv']))] # then mix them ids = [] for tp in seqs: ids += [(tp, i) for i in range(len(seqs[tp]))] np.random.shuffle(ids) seqs_ctxt = [] seqs_resp = [] for tp, i in ids: seqs_ctxt.append(seqs[tp][i][0]) seqs_resp.append(seqs[tp][i][1]) inp_enc_ctxt = self.seqs2enc(seqs_ctxt, self.max_ctxt_len) if self.noisy_AE_src: inp_enc_resp = self.seqs2enc(self.add_unk_noise(seqs_resp), self.max_resp_len) else: inp_enc_resp = self.seqs2enc(seqs_resp, self.max_resp_len) inp_dec_resp, out_dec_resp = self.seqs2dec(seqs_resp, self.max_resp_len) n_sample_conv = len(ids) d_inp_enc = {'ctxt':inp_enc_ctxt, 'resp':inp_enc_resp} d_inp_dec = {'resp':inp_dec_resp} d_out_dec = {'resp':out_dec_resp} def get_ret(n, dd): n = BATCH_SIZE * int(n/BATCH_SIZE) ret = {'n_sample':n} for d_name in dd: d = dd[d_name] for k in d: if isinstance(d[k], list): d[k] = d[k][:n] else: d[k] = d[k][:n, :] ret[d_name] = d return ret if conv_only: return get_ret(n_sample_conv, { 'inp_enc':d_inp_enc, 'inp_dec':d_inp_dec, 'out_dec':d_out_dec, 'seqs':{'resp':seqs_resp}, }) # load non-conversation (nonc) data ------------- def _read_nonc(tp, n, seqs): for line in self.files[sub][tp]: if len(seqs) >= n: break seq = [int(k) for k in line.strip('\n').split()] seq = seq[:min(len(seq), self.max_resp_len)] seqs.append(seq) return seqs # get nonc from different tp seqs = dict() for tp, n in [('base_nonc', max_n * (1. - mix_ratio[1])), ('bias_nonc', max_n * mix_ratio[1])]: seqs[tp] = [] if n < 1 or tp not in self.files[sub]: continue while True: m = len(seqs[tp]) seqs[tp] = _read_nonc(tp, n, seqs[tp]) if len(seqs[tp]) >= n: break if len(seqs[tp]) == m: self.reset(sub, tp) print('nonc from %s: %i/%i'%(tp, len(seqs[tp]), n)) if 'bias_nonc' in seqs and self.noisy_bias: seqs['bias_nonc'] = self.add_unk_noise(seqs['bias_nonc']) seqs_nonc = seqs['base_nonc'] + seqs['bias_nonc'] np.random.shuffle(seqs_nonc) if self.noisy_AE_src: inp_enc_nonc = self.seqs2enc(self.add_unk_noise(seqs_nonc), self.max_resp_len) else: inp_enc_nonc = self.seqs2enc(seqs_nonc, self.max_resp_len) inp_dec_nonc, out_dec_nonc = self.seqs2dec(seqs_nonc, self.max_resp_len) d_inp_enc['nonc'] = inp_enc_nonc d_inp_dec['nonc'] = inp_dec_nonc d_out_dec['nonc'] = out_dec_nonc n_sample = min(n_sample_conv, len(seqs_nonc)) return get_ret(n_sample, { 'inp_enc':d_inp_enc, 'inp_dec':d_inp_dec, 'out_dec':d_out_dec, 'seqs':{'resp':seqs_resp, 'nonc':seqs_nonc}, })
9,117
26.299401
97
py