feat: add DocumentData type (#4031)

In scenarios where there is a large amount of data that represents the
document rather than individual elements in the document, it may be
preferable to specify this in a single location rather than duplicating
the data across all elements (as we do for smaller metadata like
filename or filetype)

This PR adds DocumentData element type which can be used to uniquely
capture this data.

Author: ryannikolaidis

CHANGELOG.md

@@ -1,8 +1,9 @@
-## 0.18.1-dev0
+## 0.18.1
 
 ### Enhancements
 
 ### Features
+- **Add DocumentData element type** This is helpful in scenarios where there is large data that does not make sense to represent across each element in the document.
 
 ### Fixes
 - The `encoding` property of the `_CsvPartitioningContext` is now properly used.

unstructured/__version__.py

@@ -1 +1 @@
-__version__ = "0.18.1-dev0"  # pragma: no cover
+__version__ = "0.18.1"  # pragma: no cover

unstructured/documents/elements.py

@@ -640,6 +640,7 @@ class ElementType:
     PAGE_NUMBER = "PageNumber"
     CODE_SNIPPET = "CodeSnippet"
     FORM_KEYS_VALUES = "FormKeysValues"
+    DOCUMENT_DATA = "DocumentData"
 
     @classmethod
     def to_dict(cls):
@@ -975,6 +976,14 @@ class FormKeysValues(Text):
     category = "FormKeysValues"
 
 
+class DocumentData(Text):
+    """An element for capturing document-level data,
+    particularly for large data that does not make sense to
+    represent across each element in the document."""
+
+    category = "DocumentData"
+
+
 TYPE_TO_TEXT_ELEMENT_MAP: dict[str, type[Text]] = {
     ElementType.TITLE: Title,
     ElementType.SECTION_HEADER: Title,
@@ -1013,6 +1022,7 @@ class FormKeysValues(Text):
     ElementType.CODE_SNIPPET: CodeSnippet,
     ElementType.PAGE_NUMBER: PageNumber,
     ElementType.FORM_KEYS_VALUES: FormKeysValues,
+    ElementType.DOCUMENT_DATA: DocumentData,
 }
 
 

unstructured/partition/html/convert.py

@@ -216,6 +216,7 @@ def _inject_html_element_attrs(self, element_html: Tag) -> None:
     ElementType.PAGE_NUMBER: ElementHtml,
     ElementType.CODE_SNIPPET: ElementHtml,
     ElementType.FORM_KEYS_VALUES: ElementHtml,
+    ElementType.DOCUMENT_DATA: ElementHtml,
 }