refactor: fix ocr_languages parameter type (#375)

This PR adds support for both `list[str]` and `str` input formats for
`ocr_languages` parameter (e.g. ["eng", "deu"] or "eng+deu")

### Testing
CI should pass.

Author: christinestraub

CHANGELOG.md

@@ -1,3 +1,6 @@
+## 0.0.65-dev0
+* Add support for both `list[str]` and `str` input formats for `ocr_languages` parameter
+
 ## 0.0.64
 * Bump Pydantic to 2.5.x and remove it from explicit dependencies list (will be managed by fastapi)
 * Introduce Form params description in the code, which will form openapi and swagger documentation

prepline_general/api/app.py

@@ -12,7 +12,7 @@
 app = FastAPI(
     title="Unstructured Pipeline API",
     summary="Partition documents with the Unstructured library",
-    version="0.0.64",
+    version="0.0.65",
     docs_url="/general/docs",
     openapi_url="/general/openapi.json",
     servers=[

prepline_general/api/general.py

@@ -688,7 +688,7 @@ def return_content_type(filename: str):
 
 
 @router.get("/general/v0/general", include_in_schema=False)
-@router.get("/general/v0.0.64/general", include_in_schema=False)
+@router.get("/general/v0.0.65/general", include_in_schema=False)
 async def handle_invalid_get_request():
     raise HTTPException(
         status_code=status.HTTP_405_METHOD_NOT_ALLOWED, detail="Only POST requests are supported."
@@ -703,7 +703,7 @@ async def handle_invalid_get_request():
     description="Description",
     operation_id="partition_parameters",
 )
-@router.post("/general/v0.0.64/general", include_in_schema=False)
+@router.post("/general/v0.0.65/general", include_in_schema=False)
 def general_partition(
     request: Request,
     # cannot use annotated type here because of a bug described here:

prepline_general/api/models/form_params.py

@@ -61,7 +61,7 @@ def as_form(
                 description="The languages present in the document, for use in partitioning and/or OCR",
                 example="[eng]",
             ),
-            # BeforeValidator(SmartValueParser[List[str]]().value_or_first_element),
+            BeforeValidator(SmartValueParser[List[str]]().value_or_first_element),
         ] = [],
         skip_infer_table_types: Annotated[
             List[str],

prepline_general/api/utils.py

@@ -39,7 +39,15 @@ def _return_cast_first_element(values: list[E], origin_class: type) -> E | None:
 
 
 def is_convertible_to_list(s: str) -> Tuple[bool, Union[List, str]]:
-    """Determines if a given string is convertible to a list by attempting to parse it as JSON."""
+    """
+    Determines if a given string is convertible to a list.
+
+    This function first tries to parse the string as JSON. If the parsed JSON is a list, it returns
+    True along with the list. If parsing as JSON fails, it then checks if the string can be split
+    into a list using predefined delimiters ("," or "+"). If so, it returns True and the resulting list.
+    If neither condition is met, it returns False and a message indicating the string cannot
+    be converted to a list.
+    """
 
     try:
         result = json.loads(s)
@@ -48,7 +56,14 @@ def is_convertible_to_list(s: str) -> Tuple[bool, Union[List, str]]:
         else:
             return False, "Input is valid JSON but not a list."  # Valid JSON but not a list
     except json.JSONDecodeError:
-        return False, "Input is not valid JSON."  # Invalid JSON
+        pass  # proceed to check using delimiters if JSON parsing fails
+
+    delimiters = ["+", ","]
+    for delimiter in delimiters:
+        if delimiter in delimiters:
+            return True, s.split(delimiter)
+
+    return False, "Input is not valid JSON."  # Invalid JSON
 
 
 class SmartValueParser(Generic[T]):

preprocessing-pipeline-family.yaml

@@ -1,2 +1,2 @@
 name: general
-version: 0.0.64
+version: 0.0.65

test_general/api/test_app.py

@@ -140,7 +140,8 @@ def test_metadata_fields_removed():
         assert "detection_class_prob" not in response_without_coords[i]["metadata"]
 
 
-def test_ocr_languages_param():  # will eventually be depricated
+@pytest.mark.parametrize("ocr_languages", [["eng", "kor"], ["eng+kor"]])
+def test_ocr_languages_param(ocr_languages):  # will eventually be deprecated
     """
     Verify that we get the corresponding languages from the response with ocr_languages
     """
@@ -149,7 +150,7 @@ def test_ocr_languages_param():  # will eventually be depricated
     response = client.post(
         MAIN_API_ROUTE,
         files=[("files", (str(test_file), open(test_file, "rb")))],
-        data={"strategy": "ocr_only", "ocr_languages": ["eng", "kor"]},
+        data={"strategy": "ocr_only", "ocr_languages": ocr_languages},
     )
 
     assert response.status_code == 200