Log all requests/responses (debug level) (#37)

Author: elacuesta

scrapy_playwright/handler.py

@@ -11,6 +11,7 @@
     Page,
     PlaywrightContextManager,
     Request as PlaywrightRequest,
+    Response as PlaywrightResponse,
     Route,
 )
 from scrapy import Spider, signals
@@ -35,6 +36,26 @@
 logger = logging.getLogger("scrapy-playwright")
 
 
+def _make_request_logger(context_name: str) -> Callable:
+    def _log_request(request: PlaywrightRequest) -> None:
+        logger.debug(
+            f"[Context={context_name}] Request: <{request.method.upper()} {request.url}> "
+            f"(resource type: {request.resource_type}, referrer: {request.headers.get('referer')})"
+        )
+
+    return _log_request
+
+
+def _make_response_logger(context_name: str) -> Callable:
+    def _log_request(response: PlaywrightResponse) -> None:
+        logger.debug(
+            f"[Context={context_name}] Response: <{response.status} {response.url}> "
+            f"(referrer: {response.headers.get('referer')})"
+        )
+
+    return _log_request
+
+
 class ScrapyPlaywrightDownloadHandler(HTTPDownloadHandler):
     def __init__(self, crawler: Crawler) -> None:
         super().__init__(settings=crawler.settings, crawler=crawler)
@@ -107,6 +128,8 @@ async def _create_page(self, request: Request) -> Page:
             context = await self._create_browser_context(context_name, context_kwargs)
             self.contexts[context_name] = context
         page = await context.new_page()
+        page.on("request", _make_request_logger(context_name))
+        page.on("response", _make_response_logger(context_name))
         self.stats.inc_value("playwright/page_count")
         if self.default_navigation_timeout:
             page.set_default_navigation_timeout(self.default_navigation_timeout)