chore: throw UnstructuredIngestError correctly in plugins (#62)

### Summary

saw error for pinecone
```
500: Error in uploader - [UnstructuredIngestError] http error: (400) .... Vector dimension 1536 does not match the dimension of the index 1024
```
I believe this should be a 400


so I need to raise `UnstructuredIngestError` here

unfortunately its defined twice:
*
https://github.com/Unstructured-IO/unstructured-ingest/blob/main/unstructured_ingest/errors_v2.py#L6
*
https://github.com/Unstructured-IO/unstructured-ingest/blob/main/unstructured_ingest/error.py#L6

Author: yuming-long

test/api/test_api.py

@@ -322,3 +322,151 @@ def test_streaming_exception_with_valid_status_code():
     # Should use the exception's status_code
     assert invoke_response.status_code == 422
     assert "ExceptionWithValidStatusCode" in invoke_response.status_code_text
+
+
+@pytest.mark.parametrize(
+    "file_data", mock_file_data, ids=[type(fd).__name__ for fd in mock_file_data]
+)
+def test_unstructured_ingest_error_with_status_code(file_data):
+    """Test that UnstructuredIngestError with status_code is handled correctly."""
+    from test.assets.exception_status_code import (
+        function_raises_unstructured_ingest_error_with_status_code as test_fn,
+    )
+
+    client = TestClient(wrap_in_fastapi(func=test_fn, plugin_id="mock_plugin"))
+
+    post_body = {"file_data": file_data.model_dump()}
+    resp = client.post("/invoke", json=post_body)
+    resp_content = resp.json()
+    invoke_response = InvokeResponse.model_validate(resp_content)
+
+    # Should use the UnstructuredIngestError's status_code
+    assert invoke_response.status_code == 400
+    assert invoke_response.status_code_text == "Test UnstructuredIngestError with status_code"
+
+
+@pytest.mark.parametrize(
+    "file_data", mock_file_data, ids=[type(fd).__name__ for fd in mock_file_data]
+)
+def test_unstructured_ingest_error_without_status_code(file_data):
+    """Test that UnstructuredIngestError without status_code defaults to 500."""
+    from test.assets.exception_status_code import (
+        function_raises_unstructured_ingest_error_without_status_code as test_fn,
+    )
+
+    client = TestClient(wrap_in_fastapi(func=test_fn, plugin_id="mock_plugin"))
+
+    post_body = {"file_data": file_data.model_dump()}
+    resp = client.post("/invoke", json=post_body)
+    resp_content = resp.json()
+    invoke_response = InvokeResponse.model_validate(resp_content)
+
+    # Should default to 500 when UnstructuredIngestError has no status_code
+    assert invoke_response.status_code == 500
+    assert invoke_response.status_code_text == "Test UnstructuredIngestError without status_code"
+
+
+@pytest.mark.parametrize(
+    "file_data", mock_file_data, ids=[type(fd).__name__ for fd in mock_file_data]
+)
+def test_unstructured_ingest_error_with_none_status_code(file_data):
+    """Test that UnstructuredIngestError with None status_code defaults to 500."""
+    from test.assets.exception_status_code import (
+        function_raises_unstructured_ingest_error_with_none_status_code as test_fn,
+    )
+
+    client = TestClient(wrap_in_fastapi(func=test_fn, plugin_id="mock_plugin"))
+
+    post_body = {"file_data": file_data.model_dump()}
+    resp = client.post("/invoke", json=post_body)
+    resp_content = resp.json()
+    invoke_response = InvokeResponse.model_validate(resp_content)
+
+    # Should default to 500 when UnstructuredIngestError status_code is None
+    assert invoke_response.status_code == 500
+    assert invoke_response.status_code_text == "Test UnstructuredIngestError with None status_code"
+
+
+@pytest.mark.parametrize(
+    "file_data", mock_file_data, ids=[type(fd).__name__ for fd in mock_file_data]
+)
+def test_async_unstructured_ingest_error(file_data):
+    """Test that async functions with UnstructuredIngestError are handled correctly."""
+    from test.assets.exception_status_code import (
+        async_function_raises_unstructured_ingest_error as test_fn,
+    )
+
+    client = TestClient(wrap_in_fastapi(func=test_fn, plugin_id="mock_plugin"))
+
+    post_body = {"file_data": file_data.model_dump()}
+    resp = client.post("/invoke", json=post_body)
+    resp_content = resp.json()
+    invoke_response = InvokeResponse.model_validate(resp_content)
+
+    # Should use the UnstructuredIngestError's status_code
+    assert invoke_response.status_code == 503
+    assert invoke_response.status_code_text == "Async test UnstructuredIngestError"
+
+
+def test_streaming_unstructured_ingest_error():
+    """Test that async generator functions with UnstructuredIngestError are handled correctly."""
+    from test.assets.exception_status_code import (
+        async_gen_function_raises_unstructured_ingest_error as test_fn,
+    )
+
+    client = TestClient(wrap_in_fastapi(func=test_fn, plugin_id="mock_plugin"))
+
+    post_body = {"file_data": mock_file_data[0].model_dump()}
+    resp = client.post("/invoke", json=post_body)
+
+    # For streaming responses, we get NDJSON
+    assert resp.status_code == 200
+    assert resp.headers["content-type"] == "application/x-ndjson"
+
+    # Parse the streaming response - should be a single error response
+    lines = resp.content.decode().strip().split("\n")
+    assert len(lines) == 1  # Only error response since no items were yielded
+
+    # Parse the error response
+    import json
+
+    error_response = json.loads(lines[0])
+    invoke_response = InvokeResponse.model_validate(error_response)
+
+    # Should use the UnstructuredIngestError's status_code
+    assert invoke_response.status_code == 502
+    assert "Async gen test UnstructuredIngestError" in invoke_response.status_code_text
+
+
+def test_streaming_unstructured_ingest_error_with_none_status_code():
+    """Test that async generator functions with UnstructuredIngestError
+    with None status_code are handled correctly."""
+    from test.assets.exception_status_code import (
+        async_gen_function_raises_unstructured_ingest_error_with_none_status_code as test_fn,
+    )
+
+    client = TestClient(wrap_in_fastapi(func=test_fn, plugin_id="mock_plugin"))
+
+    post_body = {"file_data": mock_file_data[0].model_dump()}
+    resp = client.post("/invoke", json=post_body)
+
+    # For streaming responses, we get NDJSON
+    assert resp.status_code == 200
+    assert resp.headers["content-type"] == "application/x-ndjson"
+
+    # Parse the streaming response - should be a single error response
+    lines = resp.content.decode().strip().split("\n")
+    assert len(lines) == 1  # Only error response since no items were yielded
+
+    # Parse the error response
+    import json
+
+    error_response = json.loads(lines[0])
+    invoke_response = InvokeResponse.model_validate(error_response)
+
+    # Should default to 500 when UnstructuredIngestError status_code is None
+    assert invoke_response.status_code == 500
+    assert (
+        "Async gen test UnstructuredIngestError with None status_code"
+        in invoke_response.status_code_text
+    )

test/assets/exception_status_code.py

@@ -1,6 +1,7 @@
 """Test assets for testing exception handling with various status_code scenarios."""
 
 from fastapi import HTTPException
+from unstructured_ingest.error import UnstructuredIngestError
 
 
 class ExceptionWithNoneStatusCode(Exception):
@@ -51,6 +52,25 @@ def function_raises_generic_exception():
     raise ValueError("Generic error")
 
 
+def function_raises_unstructured_ingest_error_with_status_code():
+    """Function that raises UnstructuredIngestError with status_code."""
+    error = UnstructuredIngestError("Test UnstructuredIngestError with status_code")
+    error.status_code = 400
+    raise error
+
+
+def function_raises_unstructured_ingest_error_without_status_code():
+    """Function that raises UnstructuredIngestError without status_code."""
+    raise UnstructuredIngestError("Test UnstructuredIngestError without status_code")
+
+
+def function_raises_unstructured_ingest_error_with_none_status_code():
+    """Function that raises UnstructuredIngestError with None status_code."""
+    error = UnstructuredIngestError("Test UnstructuredIngestError with None status_code")
+    error.status_code = None
+    raise error
+
+
 # Async versions for streaming response tests
 async def async_function_raises_exception_with_none_status_code():
     """Async function that raises an exception with status_code=None."""
@@ -67,6 +87,13 @@ async def async_function_raises_exception_without_status_code():
     raise ExceptionWithoutStatusCode("Async test exception without status_code")
 
 
+async def async_function_raises_unstructured_ingest_error():
+    """Async function that raises UnstructuredIngestError."""
+    error = UnstructuredIngestError("Async test UnstructuredIngestError")
+    error.status_code = 503
+    raise error
+
+
 # Async generator versions for streaming response error tests
 async def async_gen_function_raises_exception_with_none_status_code():
     """Async generator that raises an exception with status_code=None."""
@@ -90,3 +117,23 @@ async def async_gen_function_raises_exception_without_status_code():
     if False:  # This ensures the function is detected as a generator but never yields
         yield None
     raise ExceptionWithoutStatusCode("Async gen test exception without status_code")
+
+
+async def async_gen_function_raises_unstructured_ingest_error():
+    """Async generator that raises UnstructuredIngestError."""
+    # Don't yield anything, just raise the exception
+    if False:  # This ensures the function is detected as a generator but never yields
+        yield None
+    error = UnstructuredIngestError("Async gen test UnstructuredIngestError")
+    error.status_code = 502
+    raise error
+
+
+async def async_gen_function_raises_unstructured_ingest_error_with_none_status_code():
+    """Async generator that raises UnstructuredIngestError with None status_code."""
+    # Don't yield anything, just raise the exception
+    if False:  # This ensures the function is detected as a generator but never yields
+        yield None
+    error = UnstructuredIngestError("Async gen test UnstructuredIngestError with None status_code")
+    error.status_code = None
+    raise error

unstructured_platform_plugins/__version__.py

@@ -1 +1 @@
-__version__ = "0.0.40"  # pragma: no cover
+__version__ = "0.0.41"  # pragma: no cover

unstructured_platform_plugins/etl_uvicorn/api_generator.py

@@ -12,6 +12,7 @@
 from pydantic import BaseModel, Field, create_model
 from starlette.responses import RedirectResponse
 from unstructured_ingest.data_types.file_data import BatchFileData, FileData, file_data_from_dict
+from unstructured_ingest.error import UnstructuredIngestError
 from uvicorn.config import LOG_LEVELS
 from uvicorn.importer import import_from_string
 
@@ -223,6 +224,19 @@ async def _stream_response():
                 else exc.detail,
                 file_data=request_dict.get("file_data", None),
             )
+        except UnstructuredIngestError as exc:
+            logger.error(
+                f"UnstructuredIngestError: {str(exc)} (status_code={exc.status_code})",
+                exc_info=True,
+            )
+            return InvokeResponse(
+                usage=usage,
+                message_channels=message_channels,
+                filedata_meta=filedata_meta_model.model_validate(filedata_meta.model_dump()),
+                status_code=exc.status_code or status.HTTP_500_INTERNAL_SERVER_ERROR,
+                status_code_text=str(exc),
+                file_data=request_dict.get("file_data", None),
+            )
         except Exception as invoke_error:
             logger.error(f"failed to invoke plugin: {invoke_error}", exc_info=True)
             return InvokeResponse(