feat(ocr): Implement persistent worker pool with SQLite job queue

Major OCR infrastructure improvements:
- Add persistent SQLite-based job queue for OCR tasks
- Implement worker pool with process isolation and auto-restart
- Add OCR engine selector dropdown (Tesseract/PaddleOCR) in upload zone
- Optimize Tesseract preprocessing based on benchmark results (8x faster)
- Add recognize_cif_optimized() with multi-strategy CIF extraction
- Add Romanian CIF checksum validation
- Increase Telegram long polling timeout from 10s to 30s

Squashed commits:
- feat(ocr): Implement persistent worker pool with SQLite job queue
- feat(ocr): Add OCR engine selector dropdown to upload zone
- perf(telegram): Increase long polling timeout from 10s to 30s
- perf(ocr): Optimize Tesseract preprocessing based on benchmark results

🤖 Generated with [Claude Code](https://claude.com/claude-code)

Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>

backend/main.py

@@ -48,6 +48,7 @@ logger = logging.getLogger(__name__)
 
 # Global variables for background tasks
 telegram_bot_task = None
+ocr_job_worker_running = False
 
 
 # ============================================================================
@@ -122,15 +123,32 @@ async def init_telegram_db():
         raise
 
 
-def init_paddle_ocr_background():
-    """Initialize PaddleOCR in background thread (takes 15-20s)."""
+async def init_ocr_job_worker():
+    """Initialize OCR job worker with persistent PaddleOCR.
+
+    This replaces the old background thread approach:
+    - Starts ProcessPoolExecutor with persistent worker
+    - Pre-warms PaddleOCR (loads once, reuses for all requests)
+    - Starts job queue background task
+    """
+    global ocr_job_worker_running
+
+    logger.info("[OCR] Initializing OCR job worker...")
     try:
-        logger.info("[DATA-ENTRY] Pre-loading OCR engine (background)...")
-        from backend.modules.data_entry.services.ocr_service import ocr_service
-        ocr_service.ocr_engine._init_paddle_lazy()
-        logger.info("[DATA-ENTRY] ✅ OCR engine ready")
+        from backend.modules.data_entry.services.ocr.job_worker import start_job_worker, is_running
+
+        success = await start_job_worker()
+        ocr_job_worker_running = is_running()
+
+        if success:
+            logger.info("[OCR] ✅ Job worker started (PaddleOCR persistent)")
+        else:
+            logger.warning("[OCR] ⚠️ Job worker failed to start, falling back to sync mode")
+
     except Exception as e:
-        logger.warning(f"[DATA-ENTRY] ⚠️ OCR engine pre-load failed: {e}")
+        logger.warning(f"[OCR] ⚠️ OCR job worker init failed: {e}")
+        logger.warning("[OCR] Continuing with sync OCR mode")
+        ocr_job_worker_running = False
 
 
 async def run_telegram_bot():
@@ -178,7 +196,11 @@ async def run_telegram_bot():
         # Initialize and start
         await application.initialize()
         await application.start()
-        await application.updater.start_polling(drop_pending_updates=True)
+        await application.updater.start_polling(
+            drop_pending_updates=True,
+            poll_interval=0,        # No delay between polls
+            timeout=30              # Long poll timeout 30 seconds (reduces requests from ~6/min to ~2/min)
+        )
 
         bot_info = await application.bot.get_me()
         logger.info(f"[TELEGRAM] ✅ Bot running: @{bot_info.username}")
@@ -236,9 +258,8 @@ async def startup_event():
             init_telegram_db(),
         )
 
-        # Step 3: Start PaddleOCR initialization in background thread
-        import threading
-        threading.Thread(target=init_paddle_ocr_background, daemon=True).start()
+        # Step 3: Initialize OCR job worker (with persistent PaddleOCR)
+        await init_ocr_job_worker()
 
         # Step 4: Start Telegram bot as background task
         if settings.telegram_bot_token:
@@ -260,13 +281,24 @@ async def startup_event():
 @app.on_event("shutdown")
 async def shutdown_event():
     """Application shutdown - Cleanup resources."""
-    global telegram_bot_task
+    global telegram_bot_task, ocr_job_worker_running
 
     logger.info("=" * 80)
     logger.info("[SHUTDOWN] Stopping ROA2WEB Unified Backend...")
     logger.info("=" * 80)
 
     try:
+        # Stop OCR job worker
+        if ocr_job_worker_running:
+            logger.info("[SHUTDOWN] Stopping OCR job worker...")
+            try:
+                from backend.modules.data_entry.services.ocr.job_worker import stop_job_worker
+                await stop_job_worker()
+                ocr_job_worker_running = False
+                logger.info("[SHUTDOWN] OCR job worker stopped")
+            except Exception as e:
+                logger.error(f"[SHUTDOWN] OCR worker error: {e}")
+
         # Stop Telegram bot
         if telegram_bot_task and not telegram_bot_task.done():
             logger.info("[SHUTDOWN] Stopping Telegram bot...")
@@ -409,6 +441,26 @@ async def health_check():
     else:
         health_status["modules"]["telegram_bot"] = "disabled"
 
+    # Check OCR job worker
+    global ocr_job_worker_running
+    try:
+        from backend.modules.data_entry.services.ocr.job_worker import is_running
+        from backend.modules.data_entry.services.ocr.job_queue import job_queue
+
+        if is_running():
+            # Get queue stats
+            stats = await job_queue.get_queue_stats()
+            health_status["modules"]["ocr_worker"] = {
+                "status": "running",
+                "pending_jobs": stats.get("pending", 0),
+                "processing_jobs": stats.get("processing", 0),
+                "avg_time_seconds": stats.get("average_time_seconds", 0)
+            }
+        else:
+            health_status["modules"]["ocr_worker"] = "stopped"
+    except Exception as e:
+        health_status["modules"]["ocr_worker"] = f"error: {str(e)}"
+
     return health_status