tests/test_prefilter.py

"""
Validate pipeline/prefilter.py against the 50-post oracle.

Key invariant: zero false negatives — no true food post may be skipped.
Reports non-food recall (how many of the 20 non-food posts are correctly caught).

Run: python3 tests/test_prefilter.py
"""

import json
import sys
from pathlib import Path

# Allow running from repo root without installing the package
sys.path.insert(0, str(Path(__file__).parent.parent))
from pipeline.prefilter import should_skip

FIXTURE = Path(__file__).parent / "fixtures" / "test_posts.json"


def main():
    posts = json.loads(FIXTURE.read_text())

    place_posts     = [p for p in posts if p.get("ground_truth", {}).get("is_place")]
    non_place_posts = [p for p in posts if not p.get("ground_truth", {}).get("is_place")]

    # False negatives: place posts incorrectly skipped — must be zero
    false_negatives = [p for p in place_posts if should_skip(p)]

    # True positives: non-place posts correctly caught
    true_positives  = [p for p in non_place_posts if should_skip(p)]

    print(f"Oracle: {len(place_posts)} place, {len(non_place_posts)} non-place\n")

    if false_negatives:
        print(f"FAIL — {len(false_negatives)} place post(s) incorrectly pre-filtered:")
        for p in false_negatives:
            cap = (p.get("caption") or "")[:120]
            print(f"  • {cap!r}")
        print()
    else:
        print(f"PASS — zero false negatives (no place post skipped)\n")

    pct = len(true_positives) / len(non_place_posts) * 100 if non_place_posts else 0
    print(f"Non-place recall: {len(true_positives)}/{len(non_place_posts)} correctly pre-filtered ({pct:.0f}%)")
    for p in true_positives:
        cap = (p.get("caption") or "")[:100]
        print(f"  ✓ {cap!r}")

    missed = [p for p in non_place_posts if not should_skip(p)]
    if missed:
        print(f"\nNon-place posts NOT caught by pre-filter ({len(missed)}) — will reach LLM:")
        for p in missed:
            cap = (p.get("caption") or "")[:100]
            print(f"  · {cap!r}")

    sys.exit(1 if false_negatives else 0)


if __name__ == "__main__":
    main()