text
stringlengths
1
93.6k
mm.soft_empty_cache()
transform = transforms.ToTensor()
tensors_list = [transform(image) for image in video]
batch_tensor = torch.stack(tensors_list, dim=0)
batch_tensor = batch_tensor.permute(0, 2, 3, 1).cpu().float()
return (batch_tensor,)
NODE_CLASS_MAPPINGS = {
"DownloadAndLoadDiffSynthExVideoSVD": DownloadAndLoadDiffSynthExVideoSVD,
"DiffSynthSampler": DiffSynthSampler,
}
NODE_DISPLAY_NAME_MAPPINGS = {
"DownloadAndLoadDiffSynthExVideoSVD": "DownloadAndLoadDiffSynthExVideoSVD",
"DiffSynthSampler": "DiffSynth Sampler",
}
# <FILESEP>
__author__ = 'Agostino Sturaro'
import os
import json
import networkx as nx
import matplotlib.pyplot as plt
try:
import Queue as Q # ver. < 3.0
except ImportError:
import queue as Q
subs_G = nx.Graph()
final_G = nx.Graph()
lines_by_id = dict()
point_to_id = dict()
com_lines_fpath = os.path.normpath('temp/datasets/ComLines.geojson')
this_dir = os.path.normpath(os.path.dirname(__file__))
os.chdir(this_dir)
if not os.path.isabs(com_lines_fpath):
com_lines_fpath = os.path.abspath(com_lines_fpath)
with open(com_lines_fpath) as com_lines_file:
# elec_lines = json.load(com_lines_file, parse_float=Decimal)
com_lines = json.load(com_lines_file)
for line in com_lines['features']:
# these ids may not start from 1 and may not be continuous, but they should be unique
# AFAIK, GeoJSON does not mandate a specific id property, so we just use our own
line_id = line['properties']['id']
if line['geometry'] is None:
print('Missing geometry for line {}'.format(line_id)) # debug
continue
if line_id in lines_by_id:
print('Duplicated line id {}, this line will be skipped!'.format(line_id)) # warning
continue
line_attrs = dict()
# remember the list of coordinates as a list of tuples (lat, long)
line_attrs['points'] = list()
for coords in line['geometry']['coordinates']:
point = tuple(coords)
line_attrs['points'].append(point)
lines_by_id[line_id] = line_attrs
for line_id in lines_by_id:
line_attrs = lines_by_id[line_id]
line_points = line_attrs['points']
# add line points as nodes to the graph
for point in line_points:
if point not in point_to_id:
point_id = len(point_to_id)
point_to_id[point] = point_id
final_G.add_node(point_id, attr_dict={'x': point[0], 'y': point[1]})
# connect consecutive line points to form arcs in the graph
for idx in range(0, len(line_points) - 1):
node = point_to_id[line_points[idx]]
other_node = point_to_id[line_points[idx + 1]]
final_G.add_edge(node, other_node)
lines_by_id[line_id] = line_attrs
print('len(lines) {}'.format(len(lines_by_id))) # debug
# throw away isolated components (this step is optional)
components = sorted(nx.connected_components(final_G), key=len, reverse=True)
for component_idx in range(1, len(components)):
print('isolated component {} = {}'.format(component_idx, components[component_idx]))
final_G.remove_nodes_from(components[component_idx])
print('node count without isolated components = {}'.format(final_G.number_of_nodes()))
# export graph in GraphML format
nx.write_graphml(final_G, 'temp/MN_com.graphml')