WeKnora deployment source with the reviewed wiki identity, summary and recovery fixes

scripts/verify_business_etl.py

@@ -0,0 +1,73 @@
+#!/usr/bin/env python3
+"""Exercise historical business imports in PostgreSQL and roll back every case."""
+import argparse
+import json
+from pathlib import Path
+
+from apply_weknora_business import build_plan, target_snapshot, PRESERVED_TABLES
+from apply_weknora_identities import sql_run
+from verify_asset_etl import unchanged_check
+from verify_identity_etl import rejected_case
+from weknora_business_etl import LEDGER, TARGET_DATABASE, TARGET_TENANT, business_sql
+from weknora_category_etl import literal, save_json
+from weknora_etl import consistent_snapshot
+
+
+def main():
+    parser = argparse.ArgumentParser(description=__doc__)
+    parser.add_argument('--sqlite', type=Path, required=True)
+    parser.add_argument('--kb-mapping', type=Path, required=True)
+    parser.add_argument('--artifact-root', type=Path, required=True)
+    parser.add_argument('--out', type=Path, required=True)
+    args = parser.parse_args()
+    if args.out.exists():
+        parser.error('Report exists')
+    with consistent_snapshot(args.sqlite) as (db, info):
+        plan = build_plan(db, args.kb_mapping, args.artifact_root)
+        plan['source_snapshot_sha256'] = info['sha256']
+        query = business_sql(plan)
+    first = plan['resources'][0]['target']
+    rid = literal(first['id'])
+    owner = literal(next(uid for uid in plan['identity_mapping'].values() if uid != first['owner_id']))
+    scope = f"target_id={rid} AND target_tenant_id={TARGET_TENANT}"
+    cases = [
+        ('wrong_resource_tenant_rejected', f'UPDATE workbench_resources SET tenant_id={TARGET_TENANT + 1} WHERE id={rid};', 'Mapped business resource removed'),
+        ('wrong_resource_owner_rejected', f'UPDATE workbench_resources SET owner_id={owner} WHERE id={rid};', 'Mapped business resource removed'),
+        ('deleted_resource_not_recreated', f'DELETE FROM workbench_resources WHERE id={rid};', 'Mapped business resource removed'),
+        ('ledger_target_conflict_rejected', f"UPDATE {LEDGER} SET target_id='rollback-only' WHERE {scope};", 'Business durable mapping conflict'),
+        ('ledger_source_conflict_rejected', f"UPDATE {LEDGER} SET source_snapshot='{{}}' WHERE {scope};", 'Business durable mapping conflict'),
+        ('ledger_projection_conflict_rejected', f"UPDATE {LEDGER} SET native_snapshot='{{}}' WHERE {scope};", 'Business durable mapping conflict'),
+        ('id_without_provenance_rejected', f'DELETE FROM {LEDGER} WHERE {scope};', 'Native business ID or owner/name collision'),
+        ('inactive_tenant_rejected', f"UPDATE tenants SET status='suspended' WHERE id={TARGET_TENANT};", 'Business migration requires active target tenant'),
+    ]
+    before = target_snapshot()
+    ledger_before = sql_run(f"SELECT to_regclass('public.{LEDGER}') IS NOT NULL;")
+    sql = "BEGIN ISOLATION LEVEL SERIALIZABLE; SET LOCAL statement_timeout='45s'; SET LOCAL lock_timeout='5s';\n"
+    preserved = [(table, f'_preserved_{i}') for i, table in enumerate(PRESERVED_TABLES)]
+    for table, name in preserved:
+        sql += f'CREATE TEMP TABLE {name} ON COMMIT DROP AS SELECT * FROM {table};\n'
+    sql += query
+    pairs = [('workbench_resources', '_resource_repeat'), (LEDGER, '_ledger_repeat')]
+    for table, name in pairs:
+        sql += f'CREATE TEMP TABLE {name} ON COMMIT DROP AS SELECT * FROM {table};\n'
+    sql += query + unchanged_check('exact_repeat_noop', pairs + preserved)
+    for name, mutation, expected in cases:
+        sql += rejected_case(name, mutation, expected, query)
+    sql += f"UPDATE workbench_resources SET name='Rollback native edit',status='paused',version=99,data=jsonb_set(data,'{{description}}','\"Local edit\"'::jsonb) WHERE id={rid};\n"
+    sql += 'CREATE TEMP TABLE _native_edit ON COMMIT DROP AS SELECT * FROM workbench_resources;\n'
+    sql += query + unchanged_check('native_edits_preserved', [('workbench_resources', '_native_edit'), pairs[1], *preserved])
+    sql += 'ROLLBACK;'
+    checks = [json.loads(line) for line in sql_run(sql).splitlines() if line.startswith('{')]
+    checks.extend([
+        {'check': 'all_cases_executed', 'passed': len(checks) == len(cases) + 2},
+        {'check': 'all_regressions_rolled_back', 'passed': before == target_snapshot() and ledger_before == sql_run(f"SELECT to_regclass('public.{LEDGER}') IS NOT NULL;")},
+        {'check': 'source_unchanged', 'passed': not info['source_mutated']},
+    ])
+    result = {'database': TARGET_DATABASE, 'committed': False, 'passed': all(c['passed'] for c in checks), 'checks': checks}
+    save_json(args.out, result)
+    print(json.dumps(result))
+    return 0 if result['passed'] else 1
+
+
+if __name__ == '__main__':
+    raise SystemExit(main())