fix: Use configured server_url for split page dummy request (#192)

Fixes #191. Due to a limitation in our templated codebase, we need our
split page logic to return a request that's just going to give a 200
response. Then, we jump back into the AfterSuccessHook and await all of
the pdf splits. Hopefully we can clean this up soon to avoid having an
extra call at all, but for now let's make sure the GET /general/docs is
done on the correct base url.

Author: awalker4

CHANGELOG.md

@@ -1,3 +1,12 @@
+## 0.26.1
+
+### Enhancements
+
+### Features
+
+### Fixes
+* Use the configured server_url for our split page "dummy" request
+
 ## 0.26.0
 
 ### Enhancements

_test_unstructured_client/integration/test_decorators.py

@@ -304,12 +304,14 @@ async def mock_send(_, request: httpx.Request, **kwargs):
 
         We want to make sure both code paths are retried.
         """
-        # Assert that the SDK issues our no-op request
+        # Assert that the SDK issues our dummy request
         # returned by the BeforeRequestHook
         nonlocal mock_endpoint_called
-        if request.url.host == "no-op" or "docs" in request.url.path:
+        if request.url.host == "localhost" and "docs" in request.url.path:
             mock_endpoint_called = True
             return Response(200, request=request)
+        elif "docs" in request.url.path:
+            assert False, "The server URL was not set in the dummy request"
 
         request_body = request.read()
 

gen.yaml

@@ -10,7 +10,7 @@ generation:
   auth:
     oAuth2ClientCredentialsEnabled: false
 python:
-  version: 0.26.0
+  version: 0.26.1
   additionalDependencies:
     dev:
       deepdiff: '>=6.0'

src/unstructured_client/_hooks/custom/split_pdf_hook.py

@@ -105,6 +105,7 @@ class SplitPdfHook(SDKInitHook, BeforeRequestHook, AfterSuccessHook, AfterErrorH
 
     def __init__(self) -> None:
         self.client: Optional[HttpClient] = None
+        self.base_url: Optional[str] = None
         self.async_client: Optional[AsyncHttpClient] = None
         self.coroutines_to_execute: dict[
             str, list[Coroutine[Any, Any, httpx.Response]]
@@ -156,6 +157,9 @@ def handle_request(self, request: httpx.Request) -> httpx.Response:
         #         # Otherwise, pass the request to the default transport
         #         return await self.base_transport.handle_async_request(request)
 
+        # Instead, save the base url so we can use it for our dummy request
+        self.base_url = base_url
+
         # Explicit cast to httpx.Client to avoid a typing error
         httpx_client = cast(httpx.Client, client)
         # async_httpx_client = cast(httpx.AsyncClient, async_client)
@@ -346,7 +350,7 @@ async def call_api_partial(page):
         # dummy_request = httpx.Request("GET",  "http://no-op")
         return httpx.Request(
             "GET",
-            "https://api.unstructuredapp.io/general/docs",
+            f"{self.base_url}/general/docs",
             headers={"operation_id": operation_id},
         )