remove redundant error catches and improve error message

Unstructured-IO · Oct 9, 2023 · 35e82b9 · 35e82b9
1 parent d16c6d5
commit 35e82b9
Show file tree

Hide file tree

Showing 2 changed files with 5 additions and 40 deletions.
diff --git a/unstructured/partition/image.py b/unstructured/partition/image.py
@@ -2,11 +2,7 @@
 
 from unstructured.chunking.title import add_chunking_strategy
 from unstructured.documents.elements import Element, process_metadata
-from unstructured.logger import logger
 from unstructured.partition.common import exactly_one
-from unstructured.partition.lang import (
-    convert_old_ocr_languages_to_languages,
-)
 from unstructured.partition.pdf import partition_pdf_or_image
 
 
@@ -55,25 +51,6 @@ def partition_image(
     """
     exactly_one(filename=filename, file=file)
 
-    if not isinstance(languages, list):
-        raise TypeError(
-            'The language parameter must be a list of language codes as strings, ex. ["eng"]',
-        )
-
-    if ocr_languages is not None:
-        if languages != ["eng"]:
-            raise ValueError(
-                "Only one of languages and ocr_languages should be specified. "
-                "languages is preferred. ocr_languages is marked for deprecation.",
-            )
-
-        else:
-            languages = convert_old_ocr_languages_to_languages(ocr_languages)
-            logger.warning(
-                "The ocr_languages kwarg will be deprecated in a future version of unstructured. "
-                "Please use languages instead.",
-            )
-
     return partition_pdf_or_image(
         filename=filename,
         file=file,

diff --git a/unstructured/partition/pdf.py b/unstructured/partition/pdf.py
@@ -133,22 +133,6 @@ def partition_pdf(
     """
     exactly_one(filename=filename, file=file)
 
-    if ocr_languages is not None:
-        # check if languages was set to anything not the default value
-        # languages and ocr_languages were therefore both provided - raise error
-        if languages != ["eng"]:
-            raise ValueError(
-                "Only one of languages and ocr_languages should be specified. "
-                "languages is preferred. ocr_languages is marked for deprecation.",
-            )
-
-        else:
-            languages = convert_old_ocr_languages_to_languages(ocr_languages)
-            logger.warning(
-                "The ocr_languages kwarg will be deprecated in a future version of unstructured. "
-                "Please use languages instead.",
-            )
-
     return partition_pdf_or_image(
         filename=filename,
         file=file,
@@ -214,9 +198,13 @@ def partition_pdf_or_image(
     # function.
 
     if not isinstance(languages, list):
-        raise TypeError("The language parameter must be a list of language codes as strings.")
+        raise TypeError(
+            'The language parameter must be a list of language codes as strings, ex. ["eng"]',
+        )
 
     if ocr_languages is not None:
+        # check if languages was set to anything not the default value
+        # languages and ocr_languages were therefore both provided - raise error
         if languages != ["eng"]:
             raise ValueError(
                 "Only one of languages and ocr_languages should be specified. "