#!/usr/bin/env python3
"""Ledger — extract the committed replies/ tree into a machine-readable catalog.

Source: `git ls-tree -r -l HEAD replies/`.
For every tracked reply file we record the relative path, the on-disk size of the
blob and the git blob SHA. One row per file, tab-separated, sorted by path.

Output:  wizard-work/ledger/replies-catalog.tsv
"""

import os
import subprocess
import sys
from pathlib import Path

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


def ls_tree(repo: Path):
    """Yield (rel_path, size, blob_sha) for every tracked blob under replies/."""
    proc = subprocess.run(
        ["git", "ls-tree", "-r", "-l", "HEAD", "replies/"],
        cwd=repo,
        check=True,
        capture_output=True,
        text=True,
    )
    for line in proc.stdout.splitlines():
        # git ls-tree -l prints: mode type object size<TAB>path  (whitespace-separated)
        parts = line.split()
        if len(parts) == 5 and parts[0] == "100644" and parts[1] == "blob":
            mode, otype, obj, size, path = parts
            yield path, int(size), obj


def main():
    rows = sorted(ls_tree(REPO))
    with OUT.open("w", encoding="utf-8", newline="") as fh:
        fh.write("file\tsize_bytes\tblob_sha\n")
        for path, size, sha in rows:
            fh.write("{}\t{}\t{}\n".format(path, size, sha))
    print("wrote {} rows -> {}".format(len(rows), OUT))


if __name__ == "__main__":
    main()