Root cause of the 2GB prod import.db: the sync_run_orders audit junction
recorded every order on every run; under the 1-minute scheduler ~98% of
21.7M rows were no-op ALREADY_IMPORTED re-observations. NSSM stdout/stderr
also grew unbounded (rotation never applied to the live service).
Changes:
- sqlite_service: skip ALREADY_IMPORTED rows in sync_run_orders (write-side
guard, _SKIP_JUNCTION_STATUSES); add prune_sync_history(retention_days)
with incremental_vacuum.
- maintenance_service (new): cleanup_old_logs + run_daily_maintenance.
- scheduler_service: start_maintenance_job (daily CronTrigger).
- main.py: RotatingFileHandler (sync_comenzi_current.log, 10MB x5) instead
of a new timestamped file per start; schedule daily maintenance + one-shot
catch-up at startup.
- scripts/db_maintenance.py (new): one-shot prune + VACUUM + log cleanup,
plain sqlite3, invoked by deploy.ps1 while the service is stopped.
- deploy.ps1: stop -> run db_maintenance.py -> (re)apply NSSM AppRotate*
idempotently -> start, so rotation reaches pre-existing services.
Retention defaults: 7 days history, 7 days logs.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
After a power loss the app started before Oracle was ready; init_oracle() failed
once, the pool stayed None forever (no retry), and every sync silently failed
("Oracle pool not initialized") while still hammering the GoMag API each minute,
and order-detail 500'd.
- database.ensure_oracle_pool(force): thread-safe (re)create of the pool, called
at the start of every sync cycle → self-heals within one cycle once Oracle is
back (incl. after an Oracle service restart). init_oracle_client made idempotent
so re-init can't fall back to thin mode.
- database.oracle_status() exposed; main.py startup is non-fatal via ensure pool.
- run_sync ensures the pool before the GoMag download; on failure it records a
clear run status instead of crashing and skips the wasted API calls.
- /api/sync/health reports oracle_ready/last_error; dashboard health pill shows
"Oracle indisponibil" (top priority). Recovery via the existing Start Sync button.
- order_detail degrades gracefully (200 without CODMAT + notice) instead of 500.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
GoMag vs ROA price comparison generated too many false positives
(kits, volume discounts, special prices). Removes comparison columns,
dots, badges, catalog sync endpoints, and ~950 lines of dead code.
Keeps WRITE path (sync_prices_from_order) for kit pricing.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Root cause: GET /api/sync/schedule returned interval_minutes=null when
scheduler was stopped, causing dropdown to stay on first HTML option
(1 min). Setting .value programmatically could trigger onchange, sending
a second PUT with interval=1 right after the user's intended interval.
- GET schedule falls back to DB/default (10 min) when scheduler is off
- Add _schedulerLoading flag to block onchange during loadSchedulerStatus
- Default interval 10 min everywhere (was 5 in backend)
- Cache bust dashboard.js v=33
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
- Add price_match column to SQLite, cached on order detail view
- Background backfill on startup checks all unchecked imported orders
- Extract _enrich_items_with_codmat() helper to deduplicate SKU enrichment
- Attention card now shows same nefacturate count as filter pill
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Replace Flask admin with FastAPI app (api/app/) featuring:
- Dashboard with stat cards, sync control, and history
- Mappings CRUD for ARTICOLE_TERTI with CSV import/export
- Article autocomplete from NOM_ARTICOLE
- SKU pre-validation before import
- Sync orchestration: read JSONs -> validate -> import -> log to SQLite
- APScheduler for periodic sync from UI
- File logging to logs/sync_comenzi_YYYYMMDD_HHMMSS.log
- Oracle pool None guard (503 vs 500 on unavailable)
Test suite:
- test_app_basic.py: 30 tests (imports + routes) without Oracle
- test_integration.py: 9 integration tests with Oracle
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
