# Score a run file against human.jsonl or generated.jsonl. # A run file is JSONL, one line per question: {"qid": "h0000", "ranking": ["", ...]} (best first). # Rank only modules from the question's own book (corpus.jsonl). Questions missing from the run count as wrong. # python score.py human.jsonl run.jsonl import json, sys def load(p): return [json.loads(l) for l in open(p, encoding="utf-8")] def score(qs, runs): rank = {r["qid"]: r["ranking"] for r in runs} hit = lambda q, k: q["gold"] in rank.get(q["qid"], [])[:k] out = {} groups = {"all": qs} if "lo_tagged" in qs[0]: groups |= {"LO-tagged": [q for q in qs if q["lo_tagged"]], "untagged": [q for q in qs if not q["lo_tagged"]]} for name, g in groups.items(): out[name] = {"n": len(g), "R@1": round(100 * sum(hit(q, 1) for q in g) / len(g), 1), "R@5": round(100 * sum(hit(q, 5) for q in g) / len(g), 1)} return out if __name__ == "__main__": for k, v in score(load(sys.argv[1]), load(sys.argv[2])).items(): print(f"{k:10} n={v['n']:5} R@1 {v['R@1']:5.1f} R@5 {v['R@5']:5.1f}")