Skip to content

chore: allows evaluator to run on existing predictions #734

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
jemeza-codegen merged 2 commits into from
Mar 4, 2025
Merged

chore: allows evaluator to run on existing predictions #734

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
2 commits
Select commit Hold shift + click to select a range
a020d30
chore: allows evaluator to run on existing predictions
jemeza-codegen Mar 4, 2025
73c514c
chore: minor fix
jemeza-codegen Mar 4, 2025
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
6 changes: 3 additions & 3 deletions codegen-examples/examples/swebench_agent_run/run_eval.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -159,14 +159,14 @@ async def run_eval(
"verified": SWEBenchDataset.VERIFIED,
}
dataset_enum = dataset_dict[dataset]
print(repo)

examples = get_swe_bench_examples(dataset=dataset_enum, length=length, instance_id=instance_id, repo=repo)
print(f"Examples:\n{'\n'.join([f'{e.instance_id} - {e.repo} - {e.base_commit}' for e in examples])}")

try:
if use_existing_preds is None:
print(f"Repo: {repo}")
print(f"Examples:\n{'\n'.join([f'{e.instance_id} - {e.repo} - {e.base_commit}' for e in examples])}")
print(f"Processing {len(examples)} examples...")

# Create output directory if it doesn't exist
predictions_dir.mkdir(exist_ok=True, parents=True)

Expand Down
34 changes: 17 additions & 17 deletions src/codegen/extensions/swebench/report.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -113,6 +113,8 @@ def generate_report(predictions_dir: Path, logs_dir: Path, dataset: SWEBenchData
print(f"Directory does not exist: {predictions_dir}")
return 1

predictions_jsonl = predictions_dir / "all_preds.jsonl"
existing_preds = predictions_jsonl.exists()
prediction_files = list(predictions_dir.glob("*.json"))
print(f"Found {len(prediction_files)} prediction files")

Expand All @@ -126,29 +128,27 @@ def generate_report(predictions_dir: Path, logs_dir: Path, dataset: SWEBenchData
except json.JSONDecodeError:
print(f"Error reading JSON from {file_path}")
continue
if not existing_preds:
if not predictions:
print("No valid predictions found")
return 1

print(f"Successfully loaded {len(predictions)} predictions")
print(f"Successfully loaded {len(predictions)} predictions")

if predictions:
# Create predictions JSONL file
predictions_jsonl = preds_to_jsonl(predictions, predictions_dir)
print(f"\nCreated predictions JSONL: {predictions_jsonl}")

# Setup log directory
log_dir = logs_dir / "results"
log_dir.mkdir(exist_ok=True, parents=True)
print(f"Using log directory: {log_dir}")
# Setup log directory
log_dir = logs_dir / "results"
log_dir.mkdir(exist_ok=True, parents=True)
print(f"Using log directory: {log_dir}")

# Run evaluations
run_evals(predictions_jsonl, logs_dir, dataset, run_id)
# Run evaluations
run_evals(predictions_jsonl, logs_dir, dataset, run_id)

# Get and display report
report = get_report(predictions_jsonl, logs_dir)
# Get and display report
report = get_report(predictions_jsonl, logs_dir)

# Update prediction JSONs with results
predictions = update_pred_json(predictions, report, predictions_dir)
else:
print("No valid predictions found")
return 1
# Update prediction JSONs with results
predictions = update_pred_json(predictions, report, predictions_dir)

return 0
Loading