#!/usr/bin/env python3
"""Classify every work-queue item in the troupe repo by its recorded Status line.

Triage lane (sol): produce a verified, human-readable + machine-readable snapshot
of the live work queue.  Classification rule (stated in the report so it is
inspectable and recomputable):

  OPEN    <- recorded status is EMPTY (never recorded) OR the lowercased status
              text contains one of: open, unclaimed, draft, pending.
  CLOSED  <- anything recorded as DONE / COMPLETE / CLOSED / RESOLVED etc.
              that carries none of the OPEN signals above.

Q-item detection: lines beginning "## Q<N> — "  (space + em dash), so that
derived items such as "## Q39 closure", "## Q43 closure" and "## Q45 §9
follow-on" are excluded and the two duplicate "## Q37" headers collapse to the
latest occurrence.

Reads:  WORK-QUEUE.md          (repo root)
Writes: sol/workqueue-triage.{json,md}
"""

import hashlib
import json
import os
import re
import sys
from datetime import datetime, timezone

REPO = "/root/troupe-standup"
REPO_WIZARD_OUT = os.path.join(REPO, "wizard-work", "sol")
SRC = os.path.join(REPO, "WORK-QUEUE.md")
OUT_DIR = "/root/.hermes/wizard-outputs/sol"
OUT_JSON = os.path.join(OUT_DIR, "workqueue-triage.json")
OUT_MD = os.path.join(OUT_DIR, "workqueue-triage.md")
REPO_JSON = os.path.join(REPO_WIZARD_OUT, "workqueue-triage.json")
REPO_MD = os.path.join(REPO_WIZARD_OUT, "workqueue-triage.md")

HEADER = re.compile(r"^## Q(\d+)\s+—\s+(.*)$")
# line beginning "**Status:" or "*Status:" or "**Status:**" etc.
STATUS_LINE = re.compile(r"^\s*\*\*?Status:\s*(.*)$")

OPEN_SIGNALS = ("open", "unclaimed", "draft", "pending")


def classify(status_text):
    t = status_text.lower()
    if not status_text.strip():
        return True
    return any(sig in t for sig in OPEN_SIGNALS)


def extract_status(lines, start):
    for i in range(start, len(lines)):
        line = lines[i]
        m = STATUS_LINE.match(line)
        if m:
            body = m.group(1).strip()
            # strip a matching pair of * / ** surrounding the body
            while len(body) >= 2 and ((body.startswith("**") and body.endswith("**"))
                                      or (body.startswith("*") and body.endswith("*"))):
                body = body[1:-1].strip()
            return body
    return ""


def parse_queue(path):
    with open(path, encoding="utf-8") as fh:
        lines = fh.read().split("\n")
    records = []
    last_end = 0
    for idx, line in enumerate(lines):
        m = HEADER.match(line)
        if not m:
            continue
        qid = m.group(1)
        title = m.group(2).strip()
        status = extract_status(lines, idx + 1)
        records.append({
            "id": qid,
            "title": title,
            "status_text": status,
            "open": classify(status),
        })
        last_end = idx
    # collapse duplicate ids (e.g. the two "## Q37" blocks) to the last one
    deduped = {}
    for rec in reversed(records):
        deduped[rec["id"]] = rec
    return [deduped[k] for k in sorted(deduped, key=int)]


def md_report(records):
    total = len(records)
    open_n = sum(1 for r in records if r["open"])
    closed_n = total - open_n
    top = next((r["id"] for r in records if r["open"]), None)
    lines = []
    lines.append("# Work-queue triage — sol")
    lines.append("")
    lines.append("Source: `WORK-QUEUE.md` (sha256 %s)" % sha256_file(SRC))
    lines.append("Generated: %s" % generated_utc())
    lines.append("Classification rule: OPEN = recorded status is empty or carries "
                 "'open'/'unclaimed'/'draft'/'pending'; otherwise CLOSED.")
    lines.append("")
    lines.append("## Summary")
    lines.append("")
    lines.append("- total items: %d" % total)
    lines.append("- OPEN: %d" % open_n)
    lines.append("- CLOSED: %d" % closed_n)
    lines.append("- top open item (earliest id): %s" % top)
    lines.append("")
    lines.append("## Records")
    lines.append("")
    lines.append("| Q | title | status | open |")
    lines.append("|---|---|---|---|")
    for r in records:
        st = r["status_text"][:90]
        lines.append("| Q%d | %s | %s | %s |" % (int(r["id"]), r["title"], st, "true" if r["open"] else "false"))
    lines.append("")
    lines.append("## Open items (%d)" % open_n)
    lines.append("")
    if not open_n:
        lines.append("- none")
    for r in records:
        if r["open"]:
            lines.append("- **Q%d** — %s" % (int(r["id"]), r["title"]))
    lines.append("")
    return "\n".join(lines)


def sha256_file(path):
    h = hashlib.sha256()
    with open(path, "rb") as fh:
        for chunk in iter(lambda: fh.read(1 << 20), b""):
            h.update(chunk)
    return h.hexdigest()


def generated_utc():
    return datetime.now(timezone.utc).strftime("%Y-%m-%dT%H:%M:%SZ")


def main():
    os.makedirs(OUT_DIR, exist_ok=True)
    if not os.path.exists(SRC):
        print("FATAL: source %s missing" % SRC, file=sys.stderr)
        sys.exit(1)
    records = parse_queue(SRC)
    art = {
        "generated_utc": generated_utc(),
        "source": SRC,
        "source_sha256": sha256_file(SRC),
        "rule": "OPEN = recorded status is empty or carries 'open'/'unclaimed'/'draft'/'pending'; else CLOSED",
        "records": records,
        "summary": {
            "total": len(records),
            "open": sum(1 for r in records if r["open"]),
            "closed": sum(1 for r in records if not r["open"]),
            "top_open_item": next((r["id"] for r in records if r["open"]), None),
        },
    }
    with open(OUT_JSON, "w", encoding="utf-8") as fh:
        json.dump(art, fh, indent=2)
        fh.write("\n")
    with open(OUT_MD, "w", encoding="utf-8") as fh:
        fh.write(md_report(records))
        fh.write("\n")
    for dst_json, dst_md in ((REPO_JSON, REPO_MD),):
        os.makedirs(REPO_WIZARD_OUT, exist_ok=True)
        json.dump(art, open(dst_json, "w", encoding="utf-8"), indent=2)
        open(dst_json, "a", encoding="utf-8").write("\n")
        open(dst_md, "w", encoding="utf-8").write(md_report(records))
        open(dst_md, "a", encoding="utf-8").write("\n")
    print(md_report(records))


if __name__ == "__main__":
    main()