fix: Correct proxy configuration for Crawl4AI compatibility

- Changed proxy format from dict to string (http://user:pass@ip:port)
- Updated get_random_proxy() to return Crawl4AI-compatible format
- Replaced deprecated 'proxy' with 'proxy_config' in BrowserConfig
- Fixed test script to handle new proxy format
- Added credential hiding in proxy logging for security

Fixes AttributeError: 'dict' object has no attribute 'strip'

test_amazon_scraper.py

@@ -98,7 +98,10 @@ async def test_anti_detection():
     # Test proxy
     proxy = get_random_proxy()
     if proxy:
-        logger.info(f"✓ Proxy test: {proxy['server']}")
+        # Extract just the IP for logging (hide credentials)
+        proxy_parts = proxy.split('@')
+        proxy_server = proxy_parts[1] if len(proxy_parts) > 1 else proxy
+        logger.info(f"✓ Proxy test: {proxy_server}")
     else:
         logger.warning("⚠️ Pas de proxy configuré")