#!/usr/bin/env python3
# verify_census: re-derive the extension census from repo_files.tsv and diff it
# against the committed ext-census.tsv. Exit 0 iff identical.
import collections
import os
import sys

REPO = "/root/troupe-standup"
CATALOG = "/root/.hermes/wizard-outputs/ledger/ext-census.tsv"

def ext_of(b):
    if b.startswith(".") or "." not in b:
        return "(none)"
    return b.rsplit(".", 1)[1].lower()

def rederived():
    cnt = collections.Counter()
    with open(os.path.join(REPO, "repo_files.tsv")) as fh:
        for line in fh:
            line = line.rstrip("\n")
            if not line.strip() or line.startswith("rank\tpath"):
                continue
            _, _, p = line.partition("\t")
            cnt[ext_of(os.path.basename(p.strip()))] += 1
    return cnt

def read_catalog():
    out = {}
    with open(CATALOG) as fh:
        next(fh)  # header
        for line in fh:
            parts = line.rstrip("\n").split("\t")
            if len(parts) == 3:
                out[parts[1]] = int(parts[2])
    return out

def main():
    want = rederived()
    got = read_catalog()
    keys = sorted(set(want) | set(got), key=lambda k: (-want.get(k, 0), k))
    bad = 0
    for k in keys:
        w = want.get(k, 0)
        g = got.get(k, 0)
        if w != g:
            bad += 1
            print("MISMATCH " + k + ": rederived=" + str(w) + " catalog=" + str(g))
    if want != got:
        print("FAIL: catalog differs from re-derivation (" + str(bad) + " mismatched rows)")
        return 1
    print("PASS: " + str(len(keys)) + " extension rows; " + str(sum(want.values())) + " files match the re-derived census")
    return 0

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