#!/usr/bin/env python3
"""Build the wizards roster extract from the live checkins/ directory.

Ground truth: every checkin file named checkins/2026-10-2N-<wizard>.md. The CSV
is the machine-readable table; a peer CHECK script re-derives identical bytes.
"""
import hashlib
import re
import csv
from pathlib import Path

ROOT = Path("/root/troupe-standup")
CHECKINS = ROOT / "checkins"
OUT = ROOT / "wizard-work" / "ledger"


def parse(path):
    name = path.stem                      # 2026-10-25-volt (11-char stem)
    date = name[:10]                      # 2026-10-25
    wizard = name[11:]                    # volt
    text = path.read_text(encoding="utf-8")
    header = text.splitlines()[0] if text else ""
    title = header.lstrip("#").strip() or wizard

    counts = {f: len(re.findall(r"^\s*\*\*(" + f + r")\s*:\s*$", text, re.M | re.I))
              for f in ("did", "blocked", "next", "ask")}
    counts["total_lines"] = len(text.splitlines())
    return {"wizard": wizard, "date": date, "title": title,
            **counts, "body_sha256": hashlib.sha256(text.encode()).hexdigest()}


def main():
    rows = []
    for p in sorted(CHECKINS.glob("2026-10-2*.md")):
        rows.append(parse(p))
    rows.sort(key=lambda r: (r["date"], r["wizard"]))

    OUT.mkdir(parents=True, exist_ok=True)
    csv_path = OUT / "wizards-roster-2026-10-25.csv"
    with csv_path.open("w", encoding="utf-8", newline="") as fh:
        w = csv.writer(fh)
        w.writerow(["wizard", "date", "title", "did", "blocked", "next", "ask",
                    "total_lines", "body_sha256"])
        for r in rows:
            w.writerow([r["wizard"], r["date"], r["title"], r["did"],
                        r["blocked"], r["next"], r["ask"], r["total_lines"],
                        r["body_sha256"]])
    print("wrote %s with %d rows" % (csv_path, len(rows)))


if __name__ == "__main__":
    main()