hf-papers / scripts /run_hf_hub_prompt_variant.py
evalstate's picture
evalstate HF Staff
sync: promote hf_hub_community prompt v3 + add prompt/coverage harness
bba4fab verified
#!/usr/bin/env python3
from __future__ import annotations
import argparse
import subprocess
from pathlib import Path
ROOT = Path(__file__).resolve().parents[1]
CH_SCRIPT = ROOT / 'scripts' / 'score_hf_hub_community_challenges.py'
COV_SCRIPT = ROOT / 'scripts' / 'score_hf_hub_community_coverage.py'
def run(cmd: list[str]) -> None:
print('\n[run]', ' '.join(cmd))
subprocess.run(cmd, check=True)
def main() -> None:
ap = argparse.ArgumentParser(description='Run one hf_hub_community prompt variant on challenge + coverage packs')
ap.add_argument('--variant-id', required=True, help='Variant ID for output path, e.g. v3')
ap.add_argument('--cards-dir', type=Path, required=True)
ap.add_argument('--model', default='gpt-oss')
ap.add_argument('--agent', default='hf_hub_community')
ap.add_argument('--timeout', type=int, default=240)
ap.add_argument('--out-root', type=Path, default=ROOT / 'docs' / 'hf_hub_prompt_ab')
args = ap.parse_args()
model_stem = args.model.replace('/', '_')
out_dir = args.out_root / args.variant_id / model_stem
out_dir.mkdir(parents=True, exist_ok=True)
run([
'python', str(CH_SCRIPT),
'--model', args.model,
'--agent', args.agent,
'--agent-cards', str(args.cards_dir),
'--timeout', str(args.timeout),
'--raw-results-dir', str(out_dir / 'raw' / 'challenges'),
'--json-out', str(out_dir / 'challenges.json'),
'--md-out', str(out_dir / 'challenges.md'),
])
run([
'python', str(COV_SCRIPT),
'--model', args.model,
'--agent', args.agent,
'--agent-cards', str(args.cards_dir),
'--timeout', str(args.timeout),
'--raw-results-dir', str(out_dir / 'raw' / 'coverage'),
'--json-out', str(out_dir / 'coverage.json'),
'--md-out', str(out_dir / 'coverage.md'),
])
print('\nWrote variant outputs under:')
print(out_dir)
if __name__ == '__main__':
main()