#!/usr/bin/env python3
"""Ledger — independent re-derivation check for the replies catalog.

Re-walks the live git tree with `git ls-tree -r HEAD replies/` (whitespace split,
`100644 blob` rows), builds the `path<TAB>size_bytes<TAB>blob_sha` table, and
compares it field-for-field (sorted) against the committed
`replies-catalog.tsv`. Exit 0 iff they agree.

This is a genuine cross-derivation (a second implementation, plus a fresh git
walk), not a self-referential equality: the artifact is produced by
extract_replies_catalog.py, this verifier is a separate script.
"""

import subprocess
import sys
from pathlib import Path

REPO = Path("/root/troupe-standup")
ARTIFACT = REPO / "wizard-work" / "ledger" / "replies-catalog.tsv"


def rederive():
    out = subprocess.run(
        ["git", "ls-tree", "-r", "-l", "HEAD", "replies/"],
        cwd=REPO,
        check=True,
        capture_output=True,
        text=True,
    )
    rows = []
    for line in out.stdout.splitlines():
        parts = line.split()
        if len(parts) == 5 and parts[0] == "100644" and parts[1] == "blob":
            rows.append("{}\t{}\t{}".format(parts[4], parts[3], parts[2]))
    return sorted(rows)


def main():
    artifact_rows = [line for line in ARTIFACT.read_text(encoding="utf-8").splitlines()
                     if line.strip()]
    header, data = artifact_rows[0], artifact_rows[1:]
    if header != "file\tsize_bytes\tblob_sha":
        print("FAIL: header %r" % header)
        sys.exit(1)
    expect = sorted(data)
    if len(expect) != len(rederive()):
        print("FAIL: row count %d in artifact vs %d re-derived"
              % (len(expect), len(rerderive())))
        sys.exit(1)
    if expect != rederive():
        print("FAIL: field-for-field mismatch")
        sys.exit(1)
    print("OK: %d committed reply rows, sha+size match live git" % len(expect))
    sys.exit(0)


if __name__ == "__main__":
    main()