#!/usr/bin/env python3
"""
Verify that every kommune has its full set of screenshots in Azure
storage. Reads sales/kommune/*-meta.json to know which kommuner exist
and which ones have a real eiendomFokusBilder URL (those need the
aerial-photo shot).

Output:
  /tmp/kommune-screenshot-audit.txt   per-kommune missing-file list
  /tmp/kommune-screenshot-bad-slugs.txt   slug list ready for re-shoot

Usage (after sourcing .env so AZURE_STORAGE_CONNECTION_STRING is set):
  python3 scripts/kommune/verify-screenshots.py
"""
from __future__ import annotations

import json
import os
import sys
from pathlib import Path

REPO = Path(__file__).resolve().parent.parent.parent
KOMMUNE_DIR = REPO / 'sales' / 'kommune'

# Every kommune is expected to have these. eiendom-fokus-bilder.png is
# captured only when the picker found a real cadastre, so it's checked
# conditionally below.
BASE_SHOTS = [
    'kommune.png',
    'kommuneplaner.png',
    'trafikk.png',
    'trafikkulykker.png',
    'natur.png',
    'matching.png',
    'matching-zoom.png',
    'eiendom.png',
    'eiendom-fokus.png',
    'eiendom-3d.png',
    'eiendom-leietagere.png',
    'eiendom-leietagere-tabell.png',
    'vapen.svg',
]

CONDITIONAL_SHOTS = ['eiendom-fokus-bilder.png']  # only when meta has real URL

def has_real_aerial_url(meta: dict) -> bool:
    fb = meta.get('eiendomFokusBilder') or ''
    match = meta.get('matching') or ''
    return bool(fb) and fb != match

def main() -> int:
    conn = os.environ.get('AZURE_STORAGE_CONNECTION_STRING')
    if not conn:
        print('ERROR: AZURE_STORAGE_CONNECTION_STRING not set; source .env first', file=sys.stderr)
        return 2

    try:
        from azure.storage.blob import BlobServiceClient
    except ImportError:
        print('ERROR: azure-storage-blob not installed.', file=sys.stderr)
        print('  pip install -r scripts/spellcheck/requirements.txt', file=sys.stderr)
        return 2

    svc = BlobServiceClient.from_connection_string(conn)
    container = svc.get_container_client('$web')

    print('Listing all kommune/img/* blobs...')
    have: dict[str, set[str]] = {}
    for blob in container.list_blobs(name_starts_with='kommune/img/'):
        # blob.name e.g. "kommune/img/bardu/eiendom-fokus.png"
        parts = blob.name.split('/')
        if len(parts) < 4:
            continue
        slug = parts[2]
        fname = parts[3]
        have.setdefault(slug, set()).add(fname)
    print(f'  found shots for {len(have)} slugs')

    metas = sorted(KOMMUNE_DIR.glob('*-meta.json'))
    audit_lines: list[str] = []
    bad_slugs: list[str] = []
    missing_total = 0

    for meta_path in metas:
        slug = meta_path.name.removesuffix('-meta.json')
        try:
            meta = json.loads(meta_path.read_text())
        except Exception:
            continue

        expected = list(BASE_SHOTS)
        if has_real_aerial_url(meta):
            expected.extend(CONDITIONAL_SHOTS)

        present = have.get(slug, set())
        missing = [s for s in expected if s not in present]
        if missing:
            audit_lines.append(f'{slug}: missing {", ".join(missing)}')
            bad_slugs.append(slug)
            missing_total += len(missing)

    audit_path = Path('/tmp/kommune-screenshot-audit.txt')
    bad_path = Path('/tmp/kommune-screenshot-bad-slugs.txt')
    audit_path.write_text('\n'.join(audit_lines) + ('\n' if audit_lines else ''))
    bad_path.write_text(','.join(bad_slugs) + ('\n' if bad_slugs else ''))

    print()
    print(f'Total kommuner audited: {len(metas)}')
    print(f'Kommuner with at least one missing shot: {len(bad_slugs)}')
    print(f'Missing files (cumulative): {missing_total}')
    print(f'  → {audit_path}')
    print(f'  → {bad_path}  (slug list for re-shoot)')

    if not bad_slugs:
        return 0

    print()
    print('Top 20 missing files:')
    for line in audit_lines[:20]:
        print(f'  {line}')
    if len(audit_lines) > 20:
        print(f'  ... +{len(audit_lines) - 20} more (see audit file)')
    return 0

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