Spaces:
Sleeping
Sleeping
File size: 7,936 Bytes
c5292d8 | 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 | """Main evaluation script for running all checks."""
import asyncio
import json
import shutil
from datetime import datetime
from pathlib import Path
from instructor.checks.dynamic_checks import DynamicChecker
from instructor.checks.llm_checks import LLMChecker
from instructor.checks.static_checks import StaticChecker
from instructor.database import Database
from shared.config import settings
from shared.logger import setup_logger
logger = setup_logger(__name__)
class Evaluator:
"""Main evaluator for running all checks on submissions."""
def __init__(self) -> None:
"""Initialize evaluator."""
self.db = Database()
self.llm_checker = LLMChecker()
self.temp_dir = Path("./temp_evaluations")
self.temp_dir.mkdir(parents=True, exist_ok=True)
def get_pending_repos(self) -> list[dict]:
"""Get repos that need evaluation.
Returns:
List of repo dictionaries
"""
all_repos = self.db.get_repos()
pending = []
for repo in all_repos:
repo_dict = repo.to_dict()
# Check if already evaluated
existing_results = self.db.get_results(
email=repo_dict["email"], task=repo_dict["task"]
)
# Filter to this round
round_results = [
r for r in existing_results if r.to_dict()["round"] == repo_dict["round"]
]
if not round_results:
pending.append(repo_dict)
logger.info(f"Found {len(pending)} pending repos for evaluation")
return pending
async def evaluate_repo(self, repo: dict) -> None:
"""Evaluate a single repository.
Args:
repo: Repository submission data
"""
logger.info(f"Evaluating {repo['email']}/{repo['task']}, round {repo['round']}")
# Get task details
session = self.db.get_session()
try:
task = (
session.query(self.db.Task)
.filter_by(
email=repo["email"],
task=repo["task"],
round=repo["round"],
)
.first()
)
if not task:
logger.error(f"No task found for {repo['task']}")
return
task_dict = task.to_dict()
checks_list = task_dict["checks"]
finally:
session.close()
# Clone repo
clone_dir = self.temp_dir / f"{repo['task']}_{repo['round']}"
static_checker = StaticChecker(repo["repo_url"], repo["commit_sha"], clone_dir)
try:
static_checker.clone_repo()
# Run static checks
logger.info("Running static checks...")
static_results = await self._run_static_checks(
static_checker, task_dict["timestamp"]
)
# Run LLM checks
logger.info("Running LLM checks...")
llm_results = self._run_llm_checks(clone_dir)
# Run dynamic checks
logger.info("Running dynamic checks...")
dynamic_results = await self._run_dynamic_checks(repo["pages_url"], checks_list)
# Save all results
all_results = static_results + llm_results + dynamic_results
for result in all_results:
self.db.add_result(
{
"timestamp": datetime.utcnow(),
"email": repo["email"],
"task": repo["task"],
"round": repo["round"],
"repo_url": repo["repo_url"],
"commit_sha": repo["commit_sha"],
"pages_url": repo["pages_url"],
"check": result["check"],
"score": result["score"],
"reason": result["reason"],
"logs": result.get("logs", ""),
}
)
logger.info(
f"Completed evaluation for {repo['email']}/{repo['task']}: "
f"{len(all_results)} checks"
)
except Exception as e:
logger.error(f"Error evaluating {repo['task']}: {e}", exc_info=True)
# Save error result
self.db.add_result(
{
"timestamp": datetime.utcnow(),
"email": repo["email"],
"task": repo["task"],
"round": repo["round"],
"repo_url": repo["repo_url"],
"commit_sha": repo["commit_sha"],
"pages_url": repo["pages_url"],
"check": "Evaluation",
"score": 0.0,
"reason": f"Evaluation failed: {e}",
"logs": str(e),
}
)
finally:
# Cleanup clone
if clone_dir.exists():
shutil.rmtree(clone_dir)
async def _run_static_checks(
self, checker: StaticChecker, task_timestamp: str
) -> list[dict]:
"""Run static checks.
Args:
checker: Static checker instance
task_timestamp: When task was sent
Returns:
List of check results
"""
results = []
# Parse timestamp
if isinstance(task_timestamp, str):
task_time = datetime.fromisoformat(task_timestamp.replace("Z", "+00:00"))
else:
task_time = task_timestamp
# Repo creation time
result = checker.check_created_after_task(task_time)
results.append({"check": "Repo created after task", **result})
# MIT License
result = checker.check_mit_license()
results.append({"check": "MIT LICENSE exists", **result})
# README exists and basic quality
result = checker.check_readme()
results.append({"check": "README.md basic quality", **result})
# No secrets
result = checker.check_no_secrets()
results.append({"check": "No secrets in code", **result})
return results
def _run_llm_checks(self, code_dir: Path) -> list[dict]:
"""Run LLM-based checks.
Args:
code_dir: Directory containing code
Returns:
List of check results
"""
results = []
# README quality
readme_path = code_dir / "README.md"
if readme_path.exists():
result = self.llm_checker.check_readme_quality(readme_path)
results.append({"check": "LLM: README.md quality", **result})
# Code quality
result = self.llm_checker.check_code_quality(code_dir)
results.append({"check": "LLM: Code quality", **result})
return results
async def _run_dynamic_checks(
self, pages_url: str, checks: list[str]
) -> list[dict]:
"""Run dynamic Playwright checks.
Args:
pages_url: GitHub Pages URL
checks: List of checks to run
Returns:
List of check results
"""
dynamic_checker = DynamicChecker(pages_url, checks)
results = await dynamic_checker.run_checks()
return results
async def run(self) -> None:
"""Run evaluation on all pending repos."""
logger.info("Starting evaluation process")
pending_repos = self.get_pending_repos()
if not pending_repos:
logger.info("No pending repos to evaluate")
return
# Evaluate each repo
for repo in pending_repos:
await self.evaluate_repo(repo)
logger.info("Evaluation process complete")
async def main():
"""Main entry point."""
evaluator = Evaluator()
await evaluator.run()
if __name__ == "__main__":
asyncio.run(main())
|