#!/usr/bin/env python3
"""Verify the fresh CoinGecko top-20 extraction artifact."""
import json
import sys

JSON_PATH = "/root/.hermes/wizard-outputs/ledger/crypto-top20-fresh.json"
MD_PATH = "/root/.hermes/wizard-outputs/ledger/crypto-top20-fresh.md"

errors = []


def fail(msg):
    errors.append(msg)


def main():
    try:
        with open(JSON_PATH) as f:
            rows = json.load(f)
    except Exception as e:
        fail(f"cannot read {JSON_PATH}: {e}")
        rows = None

    if rows is not None:
        if not isinstance(rows, list) or len(rows) != 20:
            fail(f"expected 20 rows, got {len(rows) if isinstance(rows, list) else type(rows)}")
        ranks = [r.get("rank") for r in rows]
        if ranks != list(range(1, 21)):
            fail(f"ranks not 1..20 contiguous: {ranks}")
        if not all(isinstance(r.get("market_cap_usd"), int) and r["market_cap_usd"] > 0 for r in rows):
            fail("market_cap_usd not positive int for all rows")
        if any(not r.get("name") or not r.get("symbol") for r in rows):
            fail("missing name/symbol in some row")

    try:
        with open(MD_PATH) as f:
            md = f.read()
    except Exception as e:
        fail(f"cannot read {MD_PATH}: {e}")
        md = ""

    if md:
        lines = [l for l in md.splitlines() if l.strip()]
        if not lines or not lines[0].startswith("# "):
            fail("md first line not a title")
        data_rows = [l for l in lines[4:] if l.startswith("|") and "---" not in l]
        if len(data_rows) != 20:
            fail(f"md data rows != 20: {len(data_rows)}")

    if errors:
        print("FAIL")
        for e in errors:
            print(" -", e)
        sys.exit(1)
    print("OK: 20 rows, ranks 1..20 contiguous, all market caps positive; md has 20 rows")


if __name__ == "__main__":
    main()