rfctr: prepare for fix to raises on file-like-object with name not a …

…path to a file (#2617)

**Summary**
Improve typing and other mechanical refactoring in preparation for fix
to issue 2308.

CHANGELOG.md

@@ -1,4 +1,4 @@
-## 0.12.6-dev5
+## 0.12.6-dev6
 
 ### Enhancements
 

test_unstructured/partition/test_common.py

@@ -5,7 +5,8 @@
 from PIL import Image
 from unstructured_inference.inference import layout
 from unstructured_inference.inference.elements import TextRegion
-from unstructured_inference.inference.layout import DocumentLayout, LayoutElement, PageLayout
+from unstructured_inference.inference.layout import DocumentLayout, PageLayout
+from unstructured_inference.inference.layoutelement import LayoutElement
 
 from unstructured.documents.coordinates import PixelSpace
 from unstructured.documents.elements import (
@@ -28,7 +29,7 @@
 
 
 class MockPageLayout(layout.PageLayout):
-    def __init__(self, number: int, image: Image):
+    def __init__(self, number: int, image: Image.Image):
         self.number = number
         self.image = image
 

typings/pptx/table.pyi

@@ -1 +1,11 @@
-class Table: ...
+class Table:
+    @property
+    def rows(self) -> tuple[_Row]: ...
+
+class _Row:
+    @property
+    def cells(self) -> tuple[_Cell]: ...
+
+class _Cell:
+    @property
+    def text(self) -> str: ...

unstructured/__version__.py

@@ -1 +1 @@
-__version__ = "0.12.6-dev5"  # pragma: no cover
+__version__ = "0.12.6-dev6"  # pragma: no cover

unstructured/documents/elements.py

@@ -899,7 +899,7 @@ class Footer(Text):
     category = "Footer"
 
 
-TYPE_TO_TEXT_ELEMENT_MAP: Dict[str, Any] = {
+TYPE_TO_TEXT_ELEMENT_MAP: Dict[str, type[Text]] = {
     ElementType.TITLE: Title,
     ElementType.SECTION_HEADER: Title,
     ElementType.HEADLINE: Title,

unstructured/file_utils/filetype.py

@@ -15,9 +15,9 @@
 from unstructured.file_utils.encoding import detect_file_encoding, format_encoding_str
 from unstructured.nlp.patterns import LIST_OF_DICTS_PATTERN
 from unstructured.partition.common import (
-    _add_element_metadata,
-    _remove_element_metadata,
+    add_element_metadata,
     exactly_one,
+    remove_element_metadata,
     set_element_hierarchy,
 )
 
@@ -602,16 +602,11 @@ def wrapper(*args: _P.args, **kwargs: _P.kwargs) -> List[Element]:
                 # NOTE(robinson) - Attached files have already run through this logic
                 # in their own partitioning function
                 if element.metadata.attached_to_filename is None:
-                    _add_element_metadata(
-                        element,
-                        **metadata_kwargs,  # type: ignore
-                    )
+                    add_element_metadata(element, **metadata_kwargs)
 
             return elements
         else:
-            return _remove_element_metadata(
-                elements,
-            )
+            return remove_element_metadata(elements)
 
     return wrapper
 
@@ -639,16 +634,11 @@ def wrapper(*args: _P.args, **kwargs: _P.kwargs) -> List[Element]:
                     # NOTE(robinson) - Attached files have already run through this logic
                     # in their own partitioning function
                     if element.metadata.attached_to_filename is None:
-                        _add_element_metadata(
-                            element,
-                            filetype=FILETYPE_TO_MIMETYPE[filetype],
-                        )
+                        add_element_metadata(element, filetype=FILETYPE_TO_MIMETYPE[filetype])
 
                 return elements
             else:
-                return _remove_element_metadata(
-                    elements,
-                )
+                return remove_element_metadata(elements)
 
         return wrapper