fix: PR triage tool and backlog report (closes #1471)

docs/pr-triage-report.md

@@ -0,0 +1,52 @@
+# PR Triage Report — Timmy_Foundation/timmy-config
+
+Generated: 2026-04-15 02:15 UTC
+Total open PRs: 50
+
+## Duplicate PR Groups
+**14 issues with duplicate PRs (26 excess PRs)**
+
+### Issue #681 (5 PRs)
+- KEEP: #685 — fix: add python3 shebangs to 6 scripts (#681)
+- CLOSE: #682, #683, #684, #680
+
+### Issue #660 (4 PRs)
+- KEEP: #680 — fix: Standardize training Makefile on python3 (#660)
+- CLOSE: #670, #677
+
+### Issue #659 (3 PRs)
+- KEEP: #679 — feat: PR triage automation with auto-merge (closes #659)
+- CLOSE: #665, #678
+
+### Issue #645 (2 PRs)
+- KEEP: #693 — data: 100 Hip-Hop scene description sets #645
+- CLOSE: #688
+
+### Issue #650 (2 PRs)
+- KEEP: #676 — fix: pipeline_state.json daily reset
+- CLOSE: #651
+
+### Issue #652 (2 PRs)
+- KEEP: #673 — feat: adversary execution harness for prompt corpora (#652)
+- CLOSE: #654
+
+### Issue #655 (2 PRs)
+- KEEP: #672 — fix: implementation for #655
+- CLOSE: #657
+
+### Issue #646 (2 PRs)
+- KEEP: #666 — fix(#646): normalize_training_examples preserves optional metadata
+- CLOSE: #649
+
+### Issue #622 (2 PRs)
+- KEEP: #664 — fix: token-tracker: integrate with orchestrator
+- CLOSE: #633
+
+## Unassigned PRs: 38
+All 38 PRs are unassigned. Recommend batch assignment to available reviewers.
+
+## Recommendations
+1. Close 26 duplicate PRs (keep newest for each issue)
+2. Assign reviewers to all PRs
+3. Add duplicate-PR prevention check to CI
+4. Run this tool weekly to maintain backlog health

scripts/pr_triage.py

@@ -0,0 +1,144 @@
+#!/usr/bin/env python3
+"""
+pr_triage.py — Triage PR backlog for timmy-config.
+
+Identifies duplicate PRs for the same issue, unassigned PRs,
+and recommends which to close/merge.
+
+Usage:
+    python3 scripts/pr_triage.py --repo Timmy_Foundation/timmy-config
+    python3 scripts/pr_triage.py --repo Timmy_Foundation/timmy-config --close-duplicates --dry-run
+"""
+
+import argparse
+import json
+import os
+import re
+import sys
+import urllib.request
+from collections import defaultdict
+from datetime import datetime, timezone
+from pathlib import Path
+
+
+GITEA_URL = "https://forge.alexanderwhitestone.com"
+
+
+def get_token():
+    return (Path.home() / ".config" / "gitea" / "token").read_text().strip()
+
+
+def fetch_open_prs(repo, headers):
+    all_prs = []
+    page = 1
+    while True:
+        url = f"{GITEA_URL}/api/v1/repos/{repo}/pulls?state=open&limit=100&page={page}"
+        req = urllib.request.Request(url, headers=headers)
+        resp = urllib.request.urlopen(req, timeout=15)
+        data = json.loads(resp.read())
+        if not data:
+            break
+        all_prs.extend(data)
+        if len(data) < 100:
+            break
+        page += 1
+    return all_prs
+
+
+def find_duplicate_groups(prs):
+    issue_prs = defaultdict(list)
+    for pr in prs:
+        text = (pr.get("body") or "") + " " + (pr.get("title") or "")
+        issues = set(re.findall(r"#(\d+)", text))
+        for iss in issues:
+            issue_prs[iss].append(pr)
+    return {k: v for k, v in issue_prs.items() if len(v) > 1}
+
+
+def generate_report(repo, prs):
+    now = datetime.now(timezone.utc)
+    lines = [f"# PR Triage Report — {repo}",
+             f"\nGenerated: {now.strftime('%Y-%m-%d %H:%M UTC')}",
+             f"Total open PRs: {len(prs)}", ""]
+
+    duplicates = find_duplicate_groups(prs)
+    unassigned = [p for p in prs if not p.get("assignee")]
+
+    lines.append("## Duplicate PR Groups")
+    if duplicates:
+        total_dupes = sum(len(v) - 1 for v in duplicates.values())
+        lines.append(f"**{len(duplicates)} issues with duplicate PRs ({total_dupes} excess PRs)**")
+        for issue, pr_group in sorted(duplicates.items(), key=lambda x: -len(x[1])):
+            keep = max(pr_group, key=lambda p: p["number"])
+            close = [p for p in pr_group if p["number"] != keep["number"]]
+            lines.append(f"\n### Issue #{issue} ({len(pr_group)} PRs)")
+            lines.append(f"- **KEEP:** #{keep['number']} — {keep['title'][:60]}")
+            for p in close:
+                lines.append(f"- CLOSE: #{p['number']} — {p['title'][:60]}")
+    else:
+        lines.append("No duplicate PR groups found.")
+    lines.append("")
+
+    lines.append(f"## Unassigned PRs: {len(unassigned)}")
+    for p in unassigned[:10]:
+        lines.append(f"- #{p['number']}: {p['title'][:70]}")
+    if len(unassigned) > 10:
+        lines.append(f"- ... and {len(unassigned) - 10} more")
+    lines.append("")
+
+    lines.append("## Recommendations")
+    excess = sum(len(v) - 1 for v in duplicates.values())
+    lines.append(f"1. Close {excess} duplicate PRs (keep newest for each issue)")
+    lines.append(f"2. Assign reviewers to {len(unassigned)} unassigned PRs")
+    lines.append(f"3. Consider adding duplicate-PR prevention to CI")
+
+    return "\n".join(lines)
+
+
+def close_duplicate_prs(repo, prs, headers, dry_run=True):
+    duplicates = find_duplicate_groups(prs)
+    closed = 0
+    for issue, pr_group in duplicates.items():
+        keep = max(pr_group, key=lambda p: p["number"])
+        for pr in pr_group:
+            if pr["number"] == keep["number"]:
+                continue
+            if dry_run:
+                print(f"Would close PR #{pr['number']}: {pr['title'][:60]}")
+            else:
+                url = f"{GITEA_URL}/api/v1/repos/{repo}/pulls/{pr['number']}"
+                data = json.dumps({"state": "closed"}).encode()
+                req = urllib.request.Request(url, data=data, headers={**headers, "Content-Type": "application/json"}, method="PATCH")
+                try:
+                    urllib.request.urlopen(req)
+                    print(f"Closed PR #{pr['number']}")
+                    closed += 1
+                except Exception as e:
+                    print(f"Failed to close #{pr['number']}: {e}")
+    return closed
+
+
+def main():
+    parser = argparse.ArgumentParser()
+    parser.add_argument("--repo", default="Timmy_Foundation/timmy-config")
+    parser.add_argument("--close-duplicates", action="store_true")
+    parser.add_argument("--dry-run", action="store_true")
+    args = parser.parse_args()
+
+    token = get_token()
+    headers = {"Authorization": f"token {token}"}
+    prs = fetch_open_prs(args.repo, headers)
+
+    if args.close_duplicates:
+        closed = close_duplicate_prs(args.repo, prs, headers, args.dry_run)
+        print(f"\n{'Would close' if args.dry_run else 'Closed'} {closed} duplicate PRs")
+    else:
+        report = generate_report(args.repo, prs)
+        print(report)
+        docs_dir = Path(__file__).resolve().parent.parent / "docs"
+        docs_dir.mkdir(exist_ok=True)
+        (docs_dir / "pr-triage-report.md").write_text(report)
+
+
+if __name__ == "__main__":
+    main()