#!/usr/bin/env python3
"""Verify the finished GitHub `python` org repos table (evidence test).

Input:  /root/.hermes/wizard-outputs/ledger/python_org_repos_top50.csv
Output: PASS line + exit 0, or FAIL + exit 1.

Checks:
  A. Schema/handshake of the stored CSV (exact header, 50 rows, unique
     repo names, ranks 1..50, integer counts >= 0, stars non-increasing).
  B. Stored top-50 equals a fresh fetch of the live GitHub API
     (independent source provenance, not the extractor's own output).
"""
import csv
import json
import sys
import urllib.request

CSV_PATH = "/root/.hermes/wizard-outputs/ledger/python_org_repos_top50.csv"
ORG = "python"
PER_PAGE = 100
N_TOP = 50
EXPECTED_HEADER = ["rank", "repo", "owner", "description", "language",
                   "stars", "forks", "open_issues", "updated_at"]
HEADERS = {
    "Accept": "application/vnd.github+json",
    "X-GitHub-Api-Version": "2022-11-28",
    "User-Agent": "ledger-verify-python-org-repos/1.0",
}


def fetch_repos():
    page, all_repos = 1, []
    while True:
        url = "https://api.github.com/orgs/%s/repos?per_page=%d&page=%d" % (ORG, PER_PAGE, page)
        req = urllib.request.Request(url, headers=HEADERS)
        with urllib.request.urlopen(req, timeout=30) as r:
            all_repos.extend(json.loads(r.read()))
            link = r.getheader("Link") or ""
        if 'rel="next"' not in link:
            break
        page += 1
    return all_repos


def main():
    try:
        with open(CSV_PATH, newline="", encoding="utf-8") as f:
            reader = csv.DictReader(f)
            header = reader.fieldnames
            rows = list(reader)
    except Exception as e:
        print("FAIL A: cannot read CSV: %s" % e)
        sys.exit(1)

    if header != EXPECTED_HEADER:
        print("FAIL A: header mismatch: %r" % (header,))
        sys.exit(1)
    if len(rows) != N_TOP:
        print("FAIL A: expected %d rows, got %d" % (N_TOP, len(rows)))
        sys.exit(1)

    names = [r["repo"] for r in rows]
    if len(set(names)) != N_TOP:
        print("FAIL A: duplicate repo names")
        sys.exit(1)
    if [int(r["rank"]) for r in rows] != list(range(1, N_TOP + 1)):
        print("FAIL A: ranks not contiguous 1..%d" % N_TOP)
        sys.exit(1)
    for r in rows:
        for c in ("stars", "forks", "open_issues"):
            try:
                v = int(r[c])
            except (TypeError, ValueError):
                print("FAIL A: %s not an integer: %r" % (c, r[c]))
                sys.exit(1)
            if v < 0:
                print("FAIL A: %s < 0: %s" % (c, v))
                sys.exit(1)
    stars = [int(r["stars"]) for r in rows]
    if stars != sorted(stars, reverse=True):
        print("FAIL A: stars not sorted non-increasing")
        sys.exit(1)

    # Source provenance: recompute the top-50 from a fresh live fetch.
    live = fetch_repos()
    live_sorted = sorted(live, key=lambda r: (-int(r.get("stargazers_count") or 0),
                                              r["full_name"] or ""))
    live_top = live_sorted[:N_TOP]
    stored_top = [(r["repo"], int(r["stars"]), r["updated_at"]) for r in rows]
    live_top_t = [(r["name"], int(r["stargazers_count"] or 0), r["updated_at"])
                  for r in live_top]
    if stored_top != live_top_t:
        print("FAIL B: stored table differs from a fresh live API fetch")
        sys.exit(1)

    print("OK: %d rows verified against a fresh live GitHub API fetch (%d total repos fetched)"
          % (len(rows), len(live)))
    sys.exit(0)


if __name__ == "__main__":
    main()