feat: Improve Search & Comparateur with persistent browser ScraperService pattern

- Created improved_search_service.py using persistent browser pattern
- Persistent browser connection with auto-reconnect capability
- Proper popup/cookie handling across all sites
- Price extraction with validation (reject unrealistic prices)
- Stock status detection
- Concurrent scraping with semaphore limits (2 sites, 2 products)
- Reuses browser context for better session management
- Screenshot capture for product images
- Updated search router to use improved service
- Added initialization/shutdown in main.py lifespan

Benefits:
✓ More reliable scraping with persistent connections
✓ Better anti-detection (consistent sessions)
✓ Improved price accuracy with validation
✓ Faster performance (reuses browser contexts)
✓ Auto-recovery from connection failures

app/main.py

@@ -18,6 +18,7 @@ from app.services.scheduler_service import scheduled_refresh, scheduler
 from app.services import auth_service, search_service, seed_enseignes
 from app.services.scheduler import start_scheduler as start_catalog_scheduler, stop_scheduler as stop_catalog_scheduler
 from app.services.amazon_scraper_service import amazon_scraper_service
+from app.services.improved_search_service import improved_search_service
 
 # Configure logging
 logging.basicConfig(
@@ -145,12 +146,17 @@ async def lifespan(app: FastAPI):
     logger.info("Initializing Amazon scraper service...")
     await amazon_scraper_service.initialize()
 
+    # Initialize Improved Search Service (persistent browser for Search/Comparatif)
+    logger.info("Initializing Improved Search Service...")
+    await improved_search_service.initialize()
+
     logger.info("Application started")
     yield
     logger.info("Shutting down schedulers and services...")
     scheduler.shutdown(wait=True)
     stop_catalog_scheduler()
     await amazon_scraper_service.shutdown()
+    await improved_search_service.shutdown()
     logger.info("Application shutdown complete")