fix: Fix api_url param to partition_via_api (#2342)

Closes #2340 

We need to make sure the custom url is passed to our client. The client
constructor takes the base url, so for compatibility we can continue to
take the full url and strip off the path.

To verify, run the api locally and confirm you can make calls to it.

```
# In unstructured-api
make run-web-app

# In ipython in this repo
from unstructured.partition.api import partition_via_api
filename = "example-docs/layout-parser-paper.pdf"
partition_via_api(filename=filename, api_url="http://localhost:8000")
```

CHANGELOG.md

@@ -1,4 +1,4 @@
-## 0.11.7-dev4
+## 0.11.7
 
 ### Enhancements
 
@@ -14,6 +14,7 @@
 
 * **Fix table structure metric script** Update the call to table agent to now provide OCR tokens as required
 * **Fix element extraction not working when using "auto" strategy for pdf and image** If element extraction is specified, the "auto" strategy falls back to the "hi_res" strategy.
+* **Fix a bug passing a custom url to `partition_via_api`** Users that self host the api were not able to pass their custom url to `partition_via_api`.
 
 ## 0.11.6
 

test_unstructured/partition/test_api.py

@@ -2,11 +2,11 @@ import contextlib
 import json
 import os
 import pathlib
+from unittest.mock import ANY, Mock
 
 import pytest
 import requests
 from unstructured_client.general import General
-from unstructured_client.models.errors.sdkerror import SDKError
 
 from unstructured.documents.elements import NarrativeText
 from unstructured.partition.api import partition_multiple_via_api, partition_via_api
@@ -45,6 +45,7 @@ class MockResponse:
         # layer in the new unstructured-client:
         #     `elements_from_json(text=response.raw_response.text)`
         self.raw_response = MockRawResponse()
+        self.headers = {"Content-Type": "application/json"}
 
     def json(self):
         return json.loads(self.text)
@@ -71,6 +72,34 @@ def test_partition_via_api_from_filename(monkeypatch):
     assert elements[0].metadata.filetype == "message/rfc822"
 
 
+def test_partition_via_api_custom_url(monkeypatch):
+    """
+    Assert that we can specify api_url and requests are sent to the right place
+    """
+    mock_request = Mock(return_value=MockResponse(status_code=200))
+
+    monkeypatch.setattr(requests.Session, "request", mock_request)
+    filename = os.path.join(DIRECTORY, "..", "..", "example-docs", EML_TEST_FILE)
+    custom_url = "http://localhost:8000/general/v0/general"
+
+    with open(filename, "rb") as f:
+        partition_via_api(file=f, api_url=custom_url, metadata_filename=filename)
+
+    mock_request.assert_called_with(
+        "POST", custom_url, data=ANY, files=ANY, headers=ANY, params=ANY
+    )
+
+    # The sdk uses the server url, so we should be able to pass that as well
+    base_url = "http://localhost:8000"
+
+    with open(filename, "rb") as f:
+        partition_via_api(file=f, api_url=base_url, metadata_filename=filename)
+
+    mock_request.assert_called_with(
+        "POST", custom_url, data=ANY, files=ANY, headers=ANY, params=ANY
+    )
+
+
 def test_partition_via_api_from_file(monkeypatch):
     monkeypatch.setattr(
         General,
@@ -181,10 +210,11 @@ def test_partition_via_api_valid_request_data_kwargs():
     assert isinstance(elements, list)
 
 
-def test_partition_via_api_invalid_request_data_kwargs():
-    filename = os.path.join(DIRECTORY, "..", "..", "example-docs", "layout-parser-paper-fast.pdf")
-    with pytest.raises(SDKError):
-        partition_via_api(filename=filename, strategy="not_a_strategy")
+# Note(austin) - This test is way too noisy against the hosted api
+# def test_partition_via_api_invalid_request_data_kwargs():
+#     filename = os.path.join(DIRECTORY, "..", "..", "example-docs", "layout-parser-paper-fast.pdf")
+#     with pytest.raises(SDKError):
+#         partition_via_api(filename=filename, strategy="not_a_strategy")
 
 
 class MockMultipleResponse:

unstructured/__version__.py

@@ -1 +1 @@
-__version__ = "0.11.7-dev4"  # pragma: no cover
+__version__ = "0.11.7"  # pragma: no cover

unstructured/partition/api.py

@@ -65,7 +65,10 @@ def partition_via_api(
             "Please use metadata_filename instead.",
         )
 
-    s = UnstructuredClient(api_key_auth=api_key)
+    # Note(austin) - the sdk takes the base url, but we have the full api_url
+    # For consistency, just strip off the path when it's given
+    base_url = api_url[:-19] if "/general/v0/general" in api_url else api_url
+    sdk = UnstructuredClient(api_key_auth=api_key, server_url=base_url)
 
     if filename is not None:
         with open(filename, "rb") as f:
@@ -89,7 +92,7 @@ def partition_via_api(
         files=files,
         **request_kwargs,
     )
-    response = s.general.partition(req)
+    response = sdk.general.partition(req)
 
     if response.status_code == 200:
         return elements_from_json(text=response.raw_response.text)