feat(chronology): rewrite classifier to use git-history evidence + 7-status enum + Needs Review section

scripts/audit/generate_chronology.py

@@ -1,12 +1,13 @@
 #!/usr/bin/env python3
-"""Generate chronology draft for Manual Slop conductor tracks.
+"""Generate chronology for Manual Slop conductor tracks.
 
 Walks conductor/tracks/ and conductor/archive/, extracts per-track data
-(date, ID, status, summary, commit range), and emits a draft to stdout.
+(date, ID, status, summary, commit range), and emits a markdown chronology
+to stdout.
 
-The script is READ-ONLY on the source folders. It writes to stdout only.
-The human cross-check (FR6 of the chronology_20260619 track) is the authority;
-this script is a starting point, not the canonical source.
+The v2 classifier uses git-history evidence (work-commit count + report
+overrides) instead of stale metadata.json.status. Returns
+(status, confidence, reason) per row.
 
 Usage:
  uv run python scripts/audit/generate_chronology.py --draft
@@ -28,6 +29,20 @@ _SENTENCE_END_RE = re.compile(r"\.\s")
 _GIT_TIMEOUT = 30
 _DEFAULT_ROOT = "conductor/"
 
+_METADATA_FIELD_PREFIXES = (
+ "**Priority:**",
+ "**Date:**",
+ "**Initialized:**",
+ "**Track:**",
+ "**Track ID:**",
+ "**Parent umbrella:**",
+ "**Status:**",
+ "**Confidence:**",
+)
+
+_WORK_COMMIT_PREFIXES = ("feat:", "fix:", "refactor:", "perf:", "test:", "docs(report):")
+_METADATA_COMMIT_PREFIXES = ("conductor(plan):", "conductor(state):", "conductor(track):", "docs(spec):", "docs(plan):")
+
 
 def extract_slug_date(folder_name: str) -> Optional[str]:
  m = _SLUG_DATE_RE.search(folder_name)
@@ -65,8 +80,8 @@ def extract_summary(folder_path: Path) -> str:
   try:
    data = json.loads(md_path.read_text(encoding="utf-8"))
    desc = str(data.get("description", "")).strip()
-   if desc:
-    return desc
+   if desc and not desc.startswith(_METADATA_FIELD_PREFIXES):
+    return _truncate_to_25_words(_first_sentence(desc))
   except (json.JSONDecodeError, OSError):
    pass
  for fname in ("spec.md", "plan.md"):
@@ -80,14 +95,14 @@ def extract_summary(folder_path: Path) -> str:
   for line in text.splitlines():
    stripped = line.strip()
    if not stripped:
-     continue
+    continue
    if stripped.startswith("#"):
-     continue
+    continue
    if stripped.startswith(">"):
-     continue
+    continue
    bare = stripped.lstrip(">").strip()
-   if bare.startswith("**Status:**") or bare.startswith("**Track ID:**") or bare.startswith("**Track:**"):
-     continue
+   if bare.startswith(_METADATA_FIELD_PREFIXES):
+    continue
    return _truncate_to_25_words(_first_sentence(bare))
  return "Imported from archive (no spec)"
 
@@ -133,47 +148,6 @@ def _repo_root(start: Path) -> Path:
  return start.parent
 
 
-def _git_log(folder_relpath: str, *args: str) -> str:
- try:
-  result = subprocess.run(
-   ["git", "log", *args, "--", folder_relpath],
-   capture_output=True,
-   text=True,
-   timeout=_GIT_TIMEOUT,
-   check=False,
-  )
-  if result.returncode != 0:
-   return ""
-  return result.stdout
- except (subprocess.SubprocessError, OSError):
-  return ""
-
-
-def _git_first_line(folder_relpath: str, *args: str) -> str:
- out = _git_log(folder_relpath, *args)
- stripped = out.strip()
- if not stripped:
-  return ""
- return stripped.splitlines()[0]
-
-
-def _repo_root(start: Path) -> Path:
- try:
-  result = subprocess.run(
-   ["git", "rev-parse", "--show-toplevel"],
-   capture_output=True,
-   text=True,
-   timeout=10,
-   check=False,
-   cwd=str(start),
-  )
-  if result.returncode == 0 and result.stdout.strip():
-   return Path(result.stdout.strip())
- except (subprocess.SubprocessError, OSError):
-  pass
- return start.parent
-
-
 def _parse_state_phase(state_path: Path) -> str:
  if not state_path.is_file():
   return "no-state-toml"
@@ -182,11 +156,27 @@ def _parse_state_phase(state_path: Path) -> str:
    if line.startswith("current_phase"):
     v = line.split("=", 1)[1].strip().split("#")[0].strip().strip('"')
     return v
- except (subprocess.SubprocessError, OSError, Exception):
+ except (OSError, Exception):
   pass
  return "?"
 
 
+def _parse_state_status(state_path: Path) -> str:
+ if not state_path.is_file():
+  return ""
+ try:
+  text = state_path.read_text(encoding="utf-8")
+ except OSError:
+  return ""
+ for line in text.splitlines():
+  stripped = line.strip()
+  if stripped.startswith("status") and "=" in stripped:
+   parts = stripped.split("=", 1)
+   if len(parts) == 2:
+    return parts[1].strip().strip('"').strip("'").split("#")[0].strip()
+ return ""
+
+
 def _last_commit_date(folder_relpath: str) -> str:
  try:
   result = subprocess.run(
@@ -198,38 +188,75 @@ def _last_commit_date(folder_relpath: str) -> str:
   return "never"
 
 
-def _classify_status(folder_link: str, current: str, track_id: str) -> str:
- """Per-row manual review classification (FR6 hard gate).
+def _count_work_commits(folder_relpath: str) -> int:
+ log: str = _git_log(folder_relpath, "--oneline")
+ count: int = 0
+ for line in log.splitlines():
+  msg: str = line.split(" ", 1)[1] if " " in line else ""
+  if msg.startswith(_WORK_COMMIT_PREFIXES) and not msg.startswith(_METADATA_COMMIT_PREFIXES):
+   count += 1
+ return count
 
- Logic (per user directive 2026-06-20):
- - PLACEHOLDER tracks: keep as is
- - archive/ folder: default to Completed (the work was done and archived; metadata status may be stale)
- - tracks/ folder + state_phase=complete OR chrono in {completed, complete, shipped}: Completed
- - tracks/ folder + everything else: keep original chrono status (in flight)
- - Abandoned is reserved for explicit user marking; the script does NOT auto-mark.
 
- Note: "Completed" (not "Shipped") is the canonical term per user directive 2026-06-20.
- This is a side-project, not a shipped product.
+def _has_report_matching(reports_dir: Path, track_id: str, prefix: str) -> bool:
+ if not reports_dir.is_dir():
+  return False
+ for f in reports_dir.iterdir():
+  if f.is_file() and f.name.startswith(prefix) and track_id in f.name:
+   return True
+ return False
+
+
+def classify_status(
+ folder_link: str,
+ current: str,
+ track_id: str,
+ repo_root: Path,
+ reports_dir: Path,
+ has_abort_report: bool = False,
+ state_status: str = "",
+) -> tuple[str, str, str]:
+ """Git-history evidence classifier returning (status, confidence, reason).
+
+ Evidence priority:
+ 1. Override signals (highest): TRACK_COMPLETION/TRACK_ABORTED reports, state.toml superseded
+ 2. Git commit evidence (medium): work-commit count
+ 3. Directory location (low): archive/ vs tracks/
+ 4. Fallback: Needs Review
  """
  if "PLACEHOLDER" in track_id:
-  return current
+  return ("Special", "high", "placeholder track")
  if "contingency" in current.lower():
-  return current
+  return ("Special", "high", "contingency track")
+ # 1. Override signals
+ if state_status == "superseded":
+  return ("Superseded", "high", "state.toml status=superseded")
+ if has_abort_report or _has_report_matching(reports_dir, track_id, "TRACK_ABORTED_"):
+  return ("Abandoned", "high", "abort report found")
+ if _has_report_matching(reports_dir, track_id, "TRACK_COMPLETION_"):
+  return ("Completed", "high", "completion report found")
+ # 2. Git commit evidence
  is_archive = folder_link.startswith("conductor/archive/")
  is_tracks = folder_link.startswith("conductor/tracks/")
+ work_commits: int = _count_work_commits(folder_link)
+ if work_commits >= 3:
+  return ("Completed", "medium", f"{work_commits} work commits")
+ if 1 <= work_commits <= 2 and is_tracks:
+  return ("In Progress", "medium", f"{work_commits} work commits in tracks/")
+ if work_commits == 0 and is_tracks:
+  return ("Active", "medium", "0 work commits in tracks/ (spec/plan only)")
+ # 3. Directory location
  if is_archive:
-  return "Completed"
- folder = Path(folder_link)
- state_phase = _parse_state_phase(folder / "state.toml") if is_tracks else "?"
- chrono_lower = current.lower()
- is_completed = chrono_lower in {"completed", "complete", "shipped"} or state_phase in {"complete", '"complete"'}
- if is_tracks and is_completed:
-  return "Completed"
- return current
+  if work_commits == 0:
+   return ("Abandoned", "low", "archived with 0 commits")
+  return ("Completed", "low", "archived but no completion report")
+ # 4. Fallback
+ return ("Needs Review", "none", "classifier inconclusive")
 
 
 def walk_track_folders(root: Path) -> list[dict]:
  repo_root: Path = _repo_root(root)
+ reports_dir: Path = repo_root / "docs" / "reports"
  rows: list[dict] = []
  for parent_dir, default_status in (
   (root / "tracks", "Active"),
@@ -252,16 +279,22 @@ def walk_track_folders(root: Path) -> list[dict]:
     first_commit = _git_first_line(folder_relpath, "--reverse", "--format=%aI")
     date = first_commit[:10] if first_commit else ""
    metadata_path = folder / "metadata.json"
-   status: str = default_status
+   meta_status: str = ""
    if metadata_path.is_file():
     try:
      data = json.loads(metadata_path.read_text(encoding="utf-8"))
      meta_status = str(data.get("status", "")).strip()
-     if meta_status:
-      status = meta_status
     except (json.JSONDecodeError, OSError):
      pass
-   status = _classify_status(folder_relpath, status, track_id)
+   state_status: str = _parse_state_status(folder / "state.toml")
+   status, confidence, reason = classify_status(
+    folder_link=folder_relpath,
+    current=meta_status or default_status,
+    track_id=track_id,
+    repo_root=repo_root,
+    reports_dir=reports_dir,
+    state_status=state_status,
+   )
    summary: str = extract_summary(folder)
    init_sha: str = _git_first_line(folder_relpath, "--reverse", "--format=%h")
    end_sha: str = _git_first_line(folder_relpath, "-1", "--format=%h")
@@ -279,6 +312,8 @@ def walk_track_folders(root: Path) -> list[dict]:
     "date": date,
     "track_id": track_id,
     "status": status,
+    "confidence": confidence,
+    "reason": reason,
     "summary": summary,
     "init_sha": init_sha,
     "end_sha": end_sha,
@@ -291,16 +326,25 @@ def walk_track_folders(root: Path) -> list[dict]:
 
 
 def format_markdown(rows: list[dict]) -> str:
- lines: list[str] = [
-  "| Date | ID | Status | Summary | Folder | Range |",
-  "| --- | --- | --- | --- | --- | --- |",
- ]
- for row in rows:
-  range_str: str = f"`{row['init_sha']}..{row['end_sha']}` ({row['commit_count']})"
+ from datetime import date as today_date
+ lines: list[str] = []
+ lines.append(f"<!-- Generated {today_date.today().isoformat()} | {len(rows)} rows -->")
+ lines.append("")
+ lines.append("| Date | ID | Status | Summary | Folder | Range |")
+ lines.append("| --- | --- | --- | --- | --- | --- |")
+ for r in rows:
+  range_str: str = f"`{r['init_sha']}..{r['end_sha']}` ({r['commit_count']})" if r["init_sha"] else "n/a"
   lines.append(
-   f"| {row['date']} | `{row['track_id']}` | {row['status']} | "
-   f"{_md_escape(row['summary'])} | `{row['folder_link']}` | {range_str} |"
+   f"| {r['date']} | `{r['track_id']}` | {r['status']} | "
+   f"{_md_escape(r['summary'])} | `{r['folder_link']}` | {range_str} |"
   )
+ needs_review = [r for r in rows if r["status"] == "Needs Review"]
+ if needs_review:
+  lines.append("")
+  lines.append("## Needs Review")
+  lines.append("")
+  for r in needs_review:
+   lines.append(f"- `{r['track_id']}` (`{r['folder_link']}`): {r['reason']}")
  return "\n".join(lines) + "\n"
 
 
@@ -311,12 +355,12 @@ def main() -> None:
   except (OSError, ValueError):
    pass
  parser = argparse.ArgumentParser(
-  description="Generate chronology draft for Manual Slop conductor tracks.",
+  description="Generate chronology for Manual Slop conductor tracks.",
  )
  parser.add_argument(
   "--draft",
   action="store_true",
-  help="Emit markdown draft table to stdout.",
+  help="Emit markdown table to stdout.",
  )
  parser.add_argument(
   "--root",
@@ -335,4 +379,4 @@ def main() -> None:
 
 
 if __name__ == "__main__":
- main()
+ main()

tests/test_generate_chronology.py

@@ -149,7 +149,7 @@ def test_classify_status_work_commits_completed(tmp_path: Path) -> None:
  reports_dir = tmp_path / "docs/reports"
  reports_dir.mkdir(parents=True)
  with patch("scripts.audit.generate_chronology._git_log") as mock_log:
-  mock_log.return_value = "feat: add thing\nfix: fix thing\nrefactor: refactor thing\n"
+  mock_log.return_value = "abc1234 feat: add thing\ndef5678 fix: fix thing\nghi9012 refactor: refactor thing\n"
   result = classify_status(
    folder_link="conductor/tracks/my_track_20260701",
    current="active",
@@ -162,12 +162,12 @@ def test_classify_status_work_commits_completed(tmp_path: Path) -> None:
  assert "work commits" in result[2]
 
 
-def test_classify_status_metadata_commits_not_counted_as_work(tmp_path: Path) -> None:
+def test_classify_status_metadata_commits_not_countd_as_work(tmp_path: Path) -> None:
  """conductor(plan): commits don't count as work commits."""
  reports_dir = tmp_path / "docs/reports"
  reports_dir.mkdir(parents=True)
  with patch("scripts.audit.generate_chronology._git_log") as mock_log:
-  mock_log.return_value = "conductor(plan): mark task\nconductor(state): update\nconductor(track): init\n"
+  mock_log.return_value = "abc1234 conductor(plan): mark task\ndef5678 conductor(state): update\nghi9012 conductor(track): init\n"
   result = classify_status(
    folder_link="conductor/tracks/my_track_20260701",
    current="active",
@@ -183,7 +183,7 @@ def test_classify_status_1_2_work_commits_in_progress(tmp_path: Path) -> None:
  reports_dir = tmp_path / "docs/reports"
  reports_dir.mkdir(parents=True)
  with patch("scripts.audit.generate_chronology._git_log") as mock_log:
-  mock_log.return_value = "feat: add thing\nfix: fix thing\n"
+  mock_log.return_value = "abc1234 feat: add thing\ndef5678 fix: fix thing\n"
   result = classify_status(
    folder_link="conductor/tracks/my_track_20260701",
    current="active",
@@ -199,7 +199,7 @@ def test_classify_status_archive_no_override_completed_low(tmp_path: Path) -> No
  reports_dir = tmp_path / "docs/reports"
  reports_dir.mkdir(parents=True)
  with patch("scripts.audit.generate_chronology._git_log") as mock_log:
-  mock_log.return_value = "feat: thing\nfix: thing\nrefactor: thing\n"
+  mock_log.return_value = "abc1234 feat: thing\ndef5678 fix: thing\nghi9012 refactor: thing\n"
   result = classify_status(
    folder_link="conductor/archive/my_track_20260701",
    current="active",
@@ -212,13 +212,13 @@ def test_classify_status_archive_no_override_completed_low(tmp_path: Path) -> No
 
 
 def test_classify_status_fallback_needs_review(tmp_path: Path) -> None:
- """Inconclusive -> Needs Review."""
+ """Inconclusive -> Needs Review (path is neither tracks/ nor archive/)."""
  reports_dir = tmp_path / "docs/reports"
  reports_dir.mkdir(parents=True)
  with patch("scripts.audit.generate_chronology._git_log") as mock_log:
   mock_log.return_value = ""
   result = classify_status(
-   folder_link="conductor/tracks/my_track",
+   folder_link="some/other/path/my_track",
    current="",
    track_id="my_track",
    repo_root=tmp_path,