"""Derive reviewed numerical summaries from measured visits, excluding discovery."""
import json
from collections import Counter
from pathlib import Path

root = Path(__file__).parent / 'evidence/2026-10-13/fresh-sessions'
rows = []
for f in sorted(root.glob('*.json')):
    doc = json.loads(f.read_text())
    if not isinstance(doc, dict) or 'result' not in doc:
        continue
    r = doc['result']
    if r['condition'] == 'discovery':
        continue
    phases = {}
    for phase in ('initial', 'choice', 'comparison', 'reload'):
        requests = [e for e in doc['events'] if e['kind'] == 'request' and e['phase'] == phase]
        responses = [e for e in doc['events'] if e['kind'] == 'response' and e['phase'] == phase]
        phases[phase] = {
            'requestCount': len(requests),
            'requestHosts': dict(sorted(Counter(e.get('host', '') for e in requests).items())),
            'responseStatuses': dict(Counter(str(e.get('status')) for e in responses)),
            'failures': sum(e['kind'] == 'failure' and e['phase'] == phase for e in doc['events']),
        }
    cookies = {}
    for checkpoint in ('before', 'afterAction', 'after', 'afterReload'):
        if checkpoint in r:
            entries = r[checkpoint]['cookies']
            cookies[checkpoint] = {'count': len(entries), 'names': sorted(c['name'] for c in entries)}
    rows.append({
        'file': f.name, 'geo': r['geo'], 'provider': r['provider'], 'condition': r['condition'],
        'status': r['status'], 'startedAt': r['startedAt'], 'elapsedMs': r['elapsedMs'],
        'attempt': r.get('attempt', 'primary'), 'extraRenderWait': r.get('extraRenderWait', False),
        'browserVersion': r['browserVersion'], 'headless': r['headless'],
        'cookiesAtStartCount': len(r.get('cookiesAtStart', [])),
        'action': r.get('action'), 'cookies': cookies, 'network': phases,
        'observationWindows': r.get('phases', []),
    })
out = {'runCount': len(rows), 'statusCounts': dict(Counter(r['status'] for r in rows)), 'visits': rows}
(root / 'measured-summary.json').write_text(json.dumps(out, indent=2))
for r in rows:
    before = r['cookies'].get('before', {}).get('count')
    after = r['cookies'].get('after', {}).get('count')
    print(r['geo'], r['provider'], r['condition'], r['status'], f'cookies={before}->{after}',
          'comparison_hosts=' + ','.join(r['network']['comparison']['requestHosts']))