| """Second triage pass: Beam sees only the logs its first taxonomy did not match, plus that taxonomy, and proposes |
| additional families (or corrected regexes). Code re-measures coverage over all logs; nothing here is self-reported. |
| usage: triage_refine.py [census] [triage.json] [--reuse]""" |
| import json, re, sys |
|
|
| import beam |
| |
| TAIL = 600 |
| ANSI = re.compile(r"\x1b\[[0-9;]*[A-Za-z]|\[A") |
| PROGRESS = re.compile(r"[^\n|]*(?:it/s|MB/s|kB/s|%\|)[^\n|]*") |
|
|
|
|
| def clean(err): |
| e = ANSI.sub("", err or "") |
| e = PROGRESS.sub("", e) |
| e = re.sub(r"[ \t]+", " ", e) |
| e = re.sub(r"\n\s*\n+", "\n", e) |
| return e.strip()[-TAIL:] |
|
|
| CENSUS = sys.argv[1] if len(sys.argv) > 1 else "/data/census.json" |
| T1 = sys.argv[2] if len(sys.argv) > 2 else "/data/triage.json" |
| RAW = T1.replace(".json", "_refine_raw.json") |
|
|
| rows = [r for r in json.load(open(CENSUS)) if r["stage"] in ("RUNTIME_ERROR", "BUILD_ERROR", "CONFIG_ERROR") and r["err"] |
| and r["err"] != "FETCHFAIL"] |
| logs = {r["id"]: clean(r["err"]) for r in rows} |
| likes = {r["id"]: r["likes"] for r in rows} |
| t1 = json.load(open(T1)) |
| missed = [sid for sid in logs if sid not in t1["assignment"]] |
| corpus = "\n".join(f"<<{sid}>>\n{logs[sid]}" for sid in missed) |
| known = json.dumps([{k: f.get(k) for k in ("name", "regex", "fixable_in_code")} for f in t1["families"]], indent=0) |
|
|
| prompt = f"""You earlier grouped the error logs of broken Hugging Face Spaces into root-cause families, each with a |
| Python regex. We ran your regexes over all {len(logs)} logs. They matched {len(logs) - len(missed)}; the {len(missed)} logs |
| below matched none of them. |
| |
| Your first families: |
| {known} |
| |
| Propose ADDITIONAL families that cover the unmatched logs (or a corrected regex for an existing family whose |
| regex was too narrow; reuse its exact name then). Same fields as before: "name", "regex" (Python, applied with |
| re.search and re.S to each cleaned log), "fixable_in_code", "typical_fix", "example_ids". Prefer exact exception |
| text over vague words. Return ONLY JSON: {{"families": [...]}} |
| |
| UNMATCHED LOGS: |
| {corpus}""" |
|
|
| if "--reuse" in sys.argv: |
| r = json.load(open(RAW)) |
| else: |
| print(f"{len(missed)} unmatched logs, prompt ~{len(prompt) // 4} tokens", flush=True) |
| r = beam.chat([{"role": "user", "content": prompt}], effort="high", max_tokens=32000, tag="triage-refine") |
| json.dump(r, open(RAW, "w")) |
| text = r["content"] |
| m = re.search(r"```(?:json)?\s*(\{.*?\})\s*```", text, re.S) |
| new = json.JSONDecoder().raw_decode(m.group(1) if m else text[text.find("{"):])[0]["families"] |
|
|
| |
| fam = {f["name"]: dict(f, round=1) for f in t1["families"]} |
| for f in new: |
| if f.get("name") in fam: |
| fam[f["name"]].update({k: f[k] for k in ("regex", "typical_fix") if f.get(k)}, round="1+2") |
| else: |
| fam[f.get("name", f"family {len(fam)}")] = dict(f, round=2) |
| fam = list(fam.values()) |
|
|
| assigned = {} |
| for f in fam: |
| try: |
| rx = re.compile(f.get("regex") or "(?!)", re.S) |
| except re.error as e: |
| f["regex_error"], f["matched"], f["first_match_only"] = str(e), 0, 0 |
| continue |
| hits = [sid for sid, log in logs.items() if rx.search(log)] |
| f["matched"], f["first_match_only"] = len(hits), 0 |
| for sid in hits: |
| if sid not in assigned: |
| assigned[sid] = f["name"] |
| f["first_match_only"] += 1 |
| for f in fam: |
| f["likes"] = sum(likes[s] for s, n in assigned.items() if n == f["name"]) |
| report = {"n_logs": len(logs), "covered": len(assigned), "coverage": round(len(assigned) / len(logs), 3), |
| "coverage_round1": t1["coverage"], |
| "fixable_in_code_spaces": sum(f["first_match_only"] for f in fam if f.get("fixable_in_code")), |
| "fixable_in_code_likes": sum(f["likes"] for f in fam if f.get("fixable_in_code")), |
| "usage": [t1["usage"], r["usage"]], "secs": [t1["secs"], r["secs"]], "families": fam, "assignment": assigned} |
| json.dump(report, open(T1, "w"), indent=1) |
| print(json.dumps({k: v for k, v in report.items() if k not in ("families", "assignment")}, indent=1)) |
| for f in sorted(fam, key=lambda f: -f.get("first_match_only", 0))[:45]: |
| print(f"{f.get('first_match_only', 0):5d} r{f.get('round')} {str(f.get('fixable_in_code')):5s} {f['name']}") |
|
|
