#!/usr/bin/env python3
"""Independent ground-truth cross-check for the ledger extension census.

Walks `wizard-work/ledger/` two ways:
  (A) os.walk
  (B) `find -type f` via subprocess
and prints per-extension counts for each, plus the files that differ.
Exit 0 = both methods agree exactly.
"""
import os
import subprocess
import sys
from collections import Counter

ROOT = "wizard-work/ledger"


def ext_of(name):
    if "." in name:
        return "." + name.rsplit(".", 1)[1].lower()
    return "(none)"


def via_oswalk():
    c = Counter()
    files = []
    for dp, dn, fns in os.walk(ROOT):
        for f in fns:
            files.append(os.path.join(dp, f))
            c[ext_of(f)] += 1
    return c, files


def via_find():
    out = subprocess.run(
        ["find", ROOT, "-type", "f"],
        capture_output=True, text=True, check=True,
    )
    c = Counter()
    files = []
    for line in out.stdout.splitlines():
        if not line:
            continue
        files.append(line)
        c[ext_of(os.path.basename(line))] += 1
    return c, files


def main():
    a_c, a_files = via_oswalk()
    b_c, b_files = via_find()
    print("os.walk total=%d find total=%d" % (len(a_files), len(b_files)))
    all_exts = sorted(set(a_c) | set(b_c))
    for ext in sorted(all_exts, key=lambda e: (-a_c[e], e)):
        print("%-10s oswalk=%d find=%d" % (ext, a_c.get(ext, 0), b_c.get(ext, 0)))
    only_a = set(a_files) - set(b_files)
    only_b = set(b_files) - set(a_files)
    if only_a:
        print("ONLY os.walk (%d):" % len(only_a))
        for p in sorted(only_a):
            print("  " + p)
    if only_b:
        print("ONLY find (%d):" % len(only_b))
        for p in sorted(only_b):
            print("  " + p)
    if not only_a and not only_b:
        print("AGREE: no file differences")
    return 0 if (not only_a and not only_b) else 1


if __name__ == "__main__":
    sys.exit(main())