fix(auto): partition() passes strategy to PPTX,DOCX (#3273)

**Summary**
Remedy gap where `strategy` argument passed to `partition()` was not
forwarded to `partition_pptx()` or `partition_docx()`.

Author: scanny

CHANGELOG.md

@@ -1,11 +1,13 @@
-## 0.14.8-dev1
+## 0.14.8-dev2
 
 ### Enhancements
 
 ### Features
 
 ### Fixes
 
+* **`partition()` now forwards `strategy` arg to `partition_docx()` and `partition_pptx()`.** A `strategy` argument passed to `partition()` (or the default value "auto" assigned by `partition()`) is now forwarded to `partition_docx()` and `partition_pptx()` when those filetypes are detected.
+
 ## 0.14.7
 
 ### Enhancements

test_unstructured/partition/test_auto.py

@@ -1,3 +1,5 @@
+# pyright: reportPrivateUsage=false
+
 from __future__ import annotations
 
 import json
@@ -6,6 +8,7 @@
 import tempfile
 import warnings
 from importlib import import_module
+from typing import Iterator
 from unittest.mock import Mock, patch
 
 import docx
@@ -20,10 +23,12 @@
     EXPECTED_TEXT_XLSX,
     EXPECTED_TITLE,
 )
+from test_unstructured.unit_utils import ANY, FixtureRequest, example_doc_path, method_mock
 from unstructured.chunking.title import chunk_by_title
 from unstructured.cleaners.core import clean_extra_whitespace
 from unstructured.documents.elements import (
     Address,
+    Element,
     ElementMetadata,
     ListItem,
     NarrativeText,
@@ -173,6 +178,34 @@ def test_auto_partition_doc_with_file(mock_docx_document, expected_docx_elements
     assert elements == expected_docx_elements
 
 
+@pytest.mark.parametrize(
+    "strategy",
+    [
+        PartitionStrategy.AUTO,
+        PartitionStrategy.FAST,
+        PartitionStrategy.HI_RES,
+        PartitionStrategy.OCR_ONLY,
+    ],
+)
+def test_partition_forwards_strategy_arg_to_partition_docx(request: FixtureRequest, strategy: str):
+    from unstructured.partition.docx import _DocxPartitioner
+
+    def fake_iter_document_elements(self: _DocxPartitioner) -> Iterator[Element]:
+        yield Text(f"strategy=={self._opts.strategy}")
+
+    _iter_elements_ = method_mock(
+        request,
+        _DocxPartitioner,
+        "_iter_document_elements",
+        side_effect=fake_iter_document_elements,
+    )
+
+    (element,) = partition(example_doc_path("simple.docx"), strategy=strategy)
+
+    _iter_elements_.assert_called_once_with(ANY)
+    assert element.text == f"strategy=={strategy}"
+
+
 @pytest.mark.parametrize(
     ("pass_metadata_filename", "content_type"),
     [(False, None), (False, "text/html"), (True, "text/html"), (True, None)],
@@ -556,6 +589,34 @@ def test_auto_partition_pptx_from_filename():
     assert elements[0].metadata.file_directory == os.path.split(filename)[0]
 
 
+@pytest.mark.parametrize(
+    "strategy",
+    [
+        PartitionStrategy.AUTO,
+        PartitionStrategy.FAST,
+        PartitionStrategy.HI_RES,
+        PartitionStrategy.OCR_ONLY,
+    ],
+)
+def test_partition_forwards_strategy_arg_to_partition_pptx(request: FixtureRequest, strategy: str):
+    from unstructured.partition.pptx import _PptxPartitioner
+
+    def fake_iter_presentation_elements(self: _PptxPartitioner) -> Iterator[Element]:
+        yield Text(f"strategy=={self._opts.strategy}")
+
+    _iter_elements_ = method_mock(
+        request,
+        _PptxPartitioner,
+        "_iter_presentation_elements",
+        side_effect=fake_iter_presentation_elements,
+    )
+
+    (element,) = partition(example_doc_path("simple.pptx"), strategy=strategy)
+
+    _iter_elements_.assert_called_once_with(ANY)
+    assert element.text == f"strategy=={strategy}"
+
+
 @pytest.mark.skipif(is_in_docker, reason="Skipping this test in Docker container")
 def test_auto_partition_ppt_from_filename():
     filename = os.path.join(EXAMPLE_DOCS_DIRECTORY, "fake-power-point.ppt")

unstructured/__version__.py

@@ -1 +1 @@
-__version__ = "0.14.8-dev1"  # pragma: no cover
+__version__ = "0.14.8-dev2"  # pragma: no cover

unstructured/partition/auto.py

@@ -327,6 +327,7 @@ def partition(
             languages=languages,
             detect_language_per_element=detect_language_per_element,
             starting_page_number=starting_page_number,
+            strategy=strategy,
             **kwargs,
         )
     elif filetype == FileType.ODT:
@@ -499,6 +500,7 @@ def partition(
             languages=languages,
             detect_language_per_element=detect_language_per_element,
             starting_page_number=starting_page_number,
+            strategy=strategy,
             **kwargs,
         )
     elif filetype == FileType.JSON: