Merge 97f7a29fec into a30945312a

api/core/rag/extractor/pdf_extractor.py

@@ -55,14 +55,18 @@ class PdfExtractor(BaseExtractor):
         import pypdfium2  # type: ignore
 
         with blob.as_bytes_io() as file_path:
-            pdf_reader = pypdfium2.PdfDocument(file_path, autoclose=True)
             try:
+                pdf_reader = pypdfium2.PdfDocument(file_path, autoclose=True)
                 for page_number, page in enumerate(pdf_reader):
-                    text_page = page.get_textpage()
+                    try:
-                    content = text_page.get_text_range()
+                        text_page = page.get_textpage()
-                    text_page.close()
+                        try:
-                    page.close()
+                            content = text_page.get_text_range()
-                    metadata = {"source": blob.source, "page": page_number}
+                            metadata = {"source": blob.source, "page": page_number}
-                    yield Document(page_content=content, metadata=metadata)
+                            yield Document(page_content=content, metadata=metadata)
+                        finally:
+                            text_page.close()
+                    finally:
+                        page.close()
             finally:
                 pdf_reader.close()

api/core/workflow/nodes/http_request/executor.py

@@ -266,21 +266,26 @@ class Executor:
         return headers
 
     def _validate_and_parse_response(self, response: httpx.Response) -> Response:
-        executor_response = Response(response)
+        try:
+            executor_response = Response(response)
 
-        threshold_size = (
+            threshold_size = (
-            dify_config.HTTP_REQUEST_NODE_MAX_BINARY_SIZE
+                dify_config.HTTP_REQUEST_NODE_MAX_BINARY_SIZE
-            if executor_response.is_file
+                if executor_response.is_file
-            else dify_config.HTTP_REQUEST_NODE_MAX_TEXT_SIZE
+                else dify_config.HTTP_REQUEST_NODE_MAX_TEXT_SIZE
-        )
-        if executor_response.size > threshold_size:
-            raise ResponseSizeError(
-                f"{'File' if executor_response.is_file else 'Text'} size is too large,"
-                f" max size is {threshold_size / 1024 / 1024:.2f} MB,"
-                f" but current size is {executor_response.readable_size}."
             )
+            if executor_response.size > threshold_size:
+                raise ResponseSizeError(
+                    f"{'File' if executor_response.is_file else 'Text'} size is too large,"
+                    f" max size is {threshold_size / 1024 / 1024:.2f} MB,"
+                    f" but current size is {executor_response.readable_size}."
+                )
 
-        return executor_response
+            return executor_response
+        except Exception as e:
+            # Ensure response is closed when an exception occurs
+            response.close()
+            raise e
 
     def _do_http_request(self, headers: dict[str, Any]) -> httpx.Response:
         """
@@ -316,21 +321,29 @@ class Executor:
             "follow_redirects": True,
             "max_retries": self.max_retries,
         }
-        # request_args = {k: v for k, v in request_args.items() if v is not None}
+
-        try:
+        # Use with statement to ensure proper resource cleanup
-            response = getattr(ssrf_proxy, self.method.lower())(**request_args)
+        with httpx.Client() as client:
-        except (ssrf_proxy.MaxRetriesExceededError, httpx.RequestError) as e:
+            try:
-            raise HttpRequestNodeError(str(e))
+                response = getattr(client, self.method.lower())(**request_args)
-        # FIXME: fix type ignore, this maybe httpx type issue
+                # Create a new Response object and copy required data
-        return response  # type: ignore
+                # This allows safe closure of the original response
+                copied_response = response.copy()
+                response.close()
+                return copied_response
+            except (ssrf_proxy.MaxRetriesExceededError, httpx.RequestError) as e:
+                raise HttpRequestNodeError(str(e))
 
     def invoke(self) -> Response:
-        # assemble headers
+        response = None
-        headers = self._assembling_headers()
+        try:
-        # do http request
+            headers = self._assembling_headers()
-        response = self._do_http_request(headers)
+            response = self._do_http_request(headers)
-        # validate response
+            return self._validate_and_parse_response(response)
-        return self._validate_and_parse_response(response)
+        except Exception as e:
+            if response is not None:
+                response.close()
+            raise e
 
     def to_log(self):
         url_parts = urlparse(self.url)