perf(screenshot): optimize memory usage and add URL collection fallback logic

- Add iterator(chunk_size=50) to ScreenshotSnapshot query to prevent BinaryField data caching and reduce memory consumption
- Implement fallback logic in URL collection: WebSite → HostPortMapping → Default URL with priority handling
- Update _collect_urls_from_provider to return tuple with data source information for better logging and debugging
- Add detailed logging to track which data source was used during URL collection
- Improve code documentation with clear return type hints and fallback priority explanation
- Prevents memory spikes when processing large screenshot datasets with binary image data

backend/apps/asset/services/screenshot_service.py

@@ -146,7 +146,9 @@ class ScreenshotService:
         """
         from apps.asset.models import Screenshot, ScreenshotSnapshot
         
-        snapshots = ScreenshotSnapshot.objects.filter(scan_id=scan_id)
+        # 使用 iterator() 避免 QuerySet 缓存大量 BinaryField 数据导致内存飙升
+        # chunk_size=50: 每次只加载 50 条记录，处理完后释放内存
+        snapshots = ScreenshotSnapshot.objects.filter(scan_id=scan_id).iterator(chunk_size=50)
         count = 0
         
         for snapshot in snapshots:

backend/apps/scan/flows/screenshot_flow.py

@@ -31,10 +31,35 @@ def _parse_screenshot_config(enabled_tools: dict) -> dict:
     }
 
 
-def _collect_urls_from_provider(provider: TargetProvider) -> list[str]:
-    """从 Provider 收集网站 URL"""
+def _collect_urls_from_provider(provider: TargetProvider) -> tuple[list[str], str]:
+    """
+    从 Provider 收集网站 URL（带回退逻辑）
+    
+    优先级：WebSite → HostPortMapping → Default URL
+    
+    Returns:
+        tuple: (urls, source)
+            - urls: URL 列表
+            - source: 数据来源 ('website' | 'host_port' | 'default')
+    """
     logger.info("从 Provider 获取网站 URL - Provider: %s", type(provider).__name__)
-    return list(provider.iter_websites())
+
+    # 优先从 WebSite 获取
+    urls = list(provider.iter_websites())
+    if urls:
+        logger.info("使用 WebSite 数据源 - 数量: %d", len(urls))
+        return urls, "website"
+
+    # 回退到 HostPortMapping
+    urls = list(provider.iter_host_port_urls())
+    if urls:
+        logger.info("WebSite 为空，回退到 HostPortMapping - 数量: %d", len(urls))
+        return urls, "host_port"
+
+    # 最终回退到默认 URL
+    urls = list(provider.iter_default_urls())
+    logger.info("HostPortMapping 为空，回退到默认 URL - 数量: %d", len(urls))
+    return urls, "default"
 
 
 def _build_empty_result(scan_id: int, target_name: str) -> dict:
@@ -96,9 +121,9 @@ def screenshot_flow(
         concurrency = config['concurrency']
         logger.info("截图配置 - 并发: %d", concurrency)
 
-        # Step 2: 从 Provider 收集 URL 列表
-        urls = _collect_urls_from_provider(provider)
-        logger.info("URL 收集完成 - 数量: %d", len(urls))
+        # Step 2: 从 Provider 收集 URL 列表（带回退逻辑）
+        urls, source = _collect_urls_from_provider(provider)
+        logger.info("URL 收集完成 - 来源: %s, 数量: %d", source, len(urls))
 
         if not urls:
             logger.warning("没有可截图的 URL，跳过截图任务")