feat(docx): add strategy parameter to DOC and ODT (#3042)

**Summary**
Because DOCX now supports the `strategy` argument to control aspects of
image extraction, `partition_doc()` and `partition_odt()` will need to
support it to because they delegate partitioning to `partition_docx()`.
This will allow image extraction to work the same way for those two
additional document-types.

CHANGELOG.md

@@ -1,4 +1,4 @@
-## 0.13.8-dev15
+## 0.13.8-dev16
 
 ### Enhancements
 

test_unstructured/partition/docx/test_doc.py

@@ -5,14 +5,17 @@ from __future__ import annotations
 import os
 import pathlib
 import tempfile
+from typing import Any
 
 import pytest
 from pytest_mock import MockFixture
 
 from test_unstructured.unit_utils import (
     CaptureFixture,
+    FixtureRequest,
     assert_round_trips_through_JSON,
     example_doc_path,
+    function_mock,
 )
 from unstructured.chunking.basic import chunk_elements
 from unstructured.documents.elements import (
@@ -243,6 +246,24 @@ def test_partition_doc_respects_detect_language_per_element_arg():
 # -- miscellaneous -------------------------------------------------------------------------------
 
 
+@pytest.mark.parametrize(
+    ("kwargs", "expected_value"),
+    [({}, None), ({"strategy": None}, None), ({"strategy": "hi_res"}, "hi_res")],
+)
+def test_partition_doc_forwards_strategy_arg_to_partition_docx(
+    request: FixtureRequest, kwargs: dict[str, Any], expected_value: str | None
+):
+    partition_docx_ = function_mock(request, "unstructured.partition.doc.partition_docx")
+
+    partition_doc(example_doc_path("simple.doc"), **kwargs)
+
+    call_kwargs = partition_docx_.call_args.kwargs
+    # -- `strategy` keyword-argument appeared in the call --
+    assert "strategy" in call_kwargs
+    # -- `strategy` argument was passed with the expected value --
+    assert call_kwargs["strategy"] == expected_value
+
+
 def test_partition_doc_grabs_emphasized_texts():
     expected_emphasized_text_contents = ["bold", "italic", "bold-italic", "bold-italic"]
     expected_emphasized_text_tags = ["b", "i", "b", "i"]

test_unstructured/partition/odt/test_odt.py

@@ -8,7 +8,12 @@ from typing import Any
 import pytest
 from pytest_mock import MockFixture
 
-from test_unstructured.unit_utils import assert_round_trips_through_JSON, example_doc_path
+from test_unstructured.unit_utils import (
+    FixtureRequest,
+    assert_round_trips_through_JSON,
+    example_doc_path,
+    function_mock,
+)
 from unstructured.chunking.basic import chunk_elements
 from unstructured.documents.elements import CompositeElement, Table, TableChunk, Title
 from unstructured.partition.docx import partition_docx
@@ -220,6 +225,24 @@ def test_partition_odt_respects_detect_language_per_element_arg():
 # -- miscellaneous -------------------------------------------------------------------------------
 
 
+@pytest.mark.parametrize(
+    ("kwargs", "expected_value"),
+    [({}, None), ({"strategy": None}, None), ({"strategy": "hi_res"}, "hi_res")],
+)
+def test_partition_odt_forwards_strategy_arg_to_partition_docx(
+    request: FixtureRequest, kwargs: dict[str, Any], expected_value: str | None
+):
+    partition_docx_ = function_mock(request, "unstructured.partition.odt.partition_docx")
+
+    partition_odt(example_doc_path("simple.odt"), **kwargs)
+
+    call_kwargs = partition_docx_.call_args.kwargs
+    # -- `strategy` keyword-argument appeared in the call --
+    assert "strategy" in call_kwargs
+    # -- `strategy` argument was passed with the expected value --
+    assert call_kwargs["strategy"] == expected_value
+
+
 def test_partition_odt_round_trips_through_json():
     """Elements produced can be serialized then deserialized without loss."""
     assert_round_trips_through_JSON(partition_odt(example_doc_path("simple.odt")))

test_unstructured/unit_utils.py

@@ -134,7 +134,7 @@ def cls_attr_mock(
 
 def function_mock(
     request: FixtureRequest, q_function_name: str, autospec: bool = True, **kwargs: Any
-):
+) -> Mock:
     """Return mock patching function with qualified name `q_function_name`.
 
     Patch is reversed after calling test returns.

unstructured/__version__.py

@@ -1 +1 @@
-__version__ = "0.13.8-dev15"  # pragma: no cover
+__version__ = "0.13.8-dev16"  # pragma: no cover

unstructured/partition/doc.py

@@ -32,6 +32,7 @@ def partition_doc(
     detect_language_per_element: bool = False,
     date_from_file_object: bool = False,
     starting_page_number: int = 1,
+    strategy: Optional[str] = None,
     **kwargs: Any,
 ) -> list[Element]:
     """Partitions Microsoft Word Documents in .doc format into its document elements.
@@ -112,6 +113,7 @@ def partition_doc(
             metadata_filename=metadata_filename,
             metadata_last_modified=metadata_last_modified or last_modified,
             starting_page_number=starting_page_number,
+            strategy=strategy,
         )
 
     # -- Remove temporary document.docx path from metadata when necessary. Note `metadata_filename`

unstructured/partition/odt.py

@@ -30,6 +30,7 @@ def partition_odt(
     metadata_filename: Optional[str] = None,
     metadata_last_modified: Optional[str] = None,
     starting_page_number: int = 1,
+    strategy: Optional[str] = None,
     **kwargs: Any,
 ) -> list[Element]:
     """Partitions Open Office Documents in .odt format into its document elements.
@@ -76,6 +77,7 @@ def partition_odt(
             metadata_filename=metadata_filename,
             metadata_last_modified=metadata_last_modified or last_modification_date,
             starting_page_number=starting_page_number,
+            strategy=strategy,
         )
 
     return elements