Feat/pass down strategy to partition ppt as well (#3274)

Following the same pattern of
#3273 and pass down
`strategy` parameter to `partition_ppt` as well.

Author: badGarnet

CHANGELOG.md

@@ -1,12 +1,12 @@
-## 0.14.8-dev2
+## 0.14.8-dev3
 
 ### 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.
+* **`partition()` now forwards `strategy` arg to `partition_docx()`, `partition_ppt()`, and `partition_pptx()`.** A `strategy` argument passed to `partition()` (or the default value "auto" assigned by `partition()`) is now forwarded to `partition_docx()`, `partition_ppt()`, and `partition_pptx()` when those filetypes are detected.
 
 ## 0.14.7
 

test_unstructured/partition/test_auto.py

@@ -617,6 +617,34 @@ def fake_iter_presentation_elements(self: _PptxPartitioner) -> Iterator[Element]
     assert element.text == f"strategy=={strategy}"
 
 
+@pytest.mark.parametrize(
+    "strategy",
+    [
+        PartitionStrategy.AUTO,
+        PartitionStrategy.FAST,
+        PartitionStrategy.HI_RES,
+        PartitionStrategy.OCR_ONLY,
+    ],
+)
+def test_partition_forwards_strategy_arg_to_partition_ppt(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("fake-power-point.ppt"), 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-dev2"  # pragma: no cover
+__version__ = "0.14.8-dev3"  # pragma: no cover

unstructured/partition/auto.py

@@ -488,6 +488,7 @@ def partition(
             infer_table_structure=infer_table_structure,
             languages=languages,
             detect_language_per_element=detect_language_per_element,
+            strategy=strategy,
             **kwargs,
         )
     elif filetype == FileType.PPTX:

unstructured/partition/ppt.py

@@ -14,6 +14,7 @@
     get_last_modified_date_from_file,
 )
 from unstructured.partition.pptx import partition_pptx
+from unstructured.partition.utils.constants import PartitionStrategy
 
 
 @process_metadata()
@@ -33,6 +34,7 @@ def partition_ppt(
     detect_language_per_element: bool = False,
     date_from_file_object: bool = False,
     starting_page_number: int = 1,
+    strategy: str = PartitionStrategy.FAST,
     **kwargs: Any,
 ) -> list[Element]:
     """Partitions Microsoft PowerPoint Documents in .ppt format into their document elements.
@@ -113,6 +115,7 @@ def partition_ppt(
             metadata_filename=metadata_filename,
             metadata_last_modified=metadata_last_modified or last_modification_date,
             starting_page_number=starting_page_number,
+            strategy=strategy,
         )
 
     # remove tmp.name from filename if parsing file