"""Inspect real local SQLite joins over an authored synthetic event log."""
import argparse
import hashlib
import json
from pathlib import Path
from history import History, minute

ROOT = Path(__file__).resolve().parent
KEY = 'user_events_1d@v1'
CUTOFF = minute('2026-01-08T12:00:00Z')
FIRST = minute('2026-01-08T12:30:00Z')
CORRECTION = minute('2026-01-08T14:00:00Z')


def events():
    return json.loads((ROOT / 'scenario.json').read_text())['events']


def observations():
    return [dict(id=id, entity=entity, predict_at='2026-01-08T'+time+'Z')
            for id, entity, time in [('before-job','user-01','12:15:00'),
                                    ('first-read','user-01','13:00:00'),
                                    ('after-correction','user-01','15:00:00'),
                                    ('unknown-user','user-99','15:00:00')]]


def record():
    db = History()
    try:
        db.materialize(events(), CUTOFF, FIRST)
        before = db.join(observations(), KEY)
        db.materialize(events(), CUTOFF, CORRECTION, 'corrected')
        after = db.join(observations(), KEY)
        return {'mode': 'local SQLite; authored synthetic data; no model',
                'input_sha256': hashlib.sha256((ROOT/'scenario.json').read_bytes()).hexdigest(),
                'cutoff': CUTOFF, 'first_published_at': FIRST, 'corrected_published_at': CORRECTION,
                'before_backfill': before, 'after_backfill': after,
                'historical_read_unchanged': before[1] == after[1],
                'retained_vintages_user_01': db.db.execute('SELECT COUNT(*) FROM history '
                                                        'WHERE entity=?', ('user-01',)).fetchone()[0]}
    finally:
        db.close()


def main():
    p = argparse.ArgumentParser(); p.add_argument('--check', action='store_true'); p.add_argument('--save', type=Path)
    args=p.parse_args(); result=record()
    if args.check and result != json.loads((ROOT/'join-result.json').read_text()):
        raise SystemExit('joined rows differ from the bundled record')
    if args.save: args.save.write_text(json.dumps(result, indent=2)+'\n')
    print('Feature window ends at Jan 08 12:00 UTC; first publication 12:30; correction 14:00')
    for row in result['after_backfill']:
        feature=row['feature']; value='missing' if feature is None else feature['value']
        print(f"{row['id']}: predict_at={row['predict_at']}; feature={value}")
    print('13:00 historical read unchanged after correction:', result['historical_read_unchanged'])
    print('Retained user-01 vintages:', result['retained_vintages_user_01'])


if __name__=='__main__': main()