Fix layout parsing (#3754)

Unstructured-IO · Oct 25, 2024 · 5a91f0c · 5a91f0c
1 parent 2417f8e
commit 5a91f0c
Show file tree

Hide file tree

Showing 4 changed files with 44 additions and 9 deletions.
diff --git a/CHANGELOG.md b/CHANGELOG.md
@@ -1,4 +1,4 @@
-## 0.16.3-dev1
+## 0.16.3-dev2
 
 ### Enhancements
 
@@ -7,6 +7,7 @@
 ### Fixes
 
 * **V2 elements without first parent ID can be parsed**
+* **Fix missing elements when layout element parsed in V2 ontology**
 
 
 ## 0.16.2

diff --git a/test_unstructured/partition/html/test_html_to_ontology_parsing.py b/test_unstructured/partition/html/test_html_to_ontology_parsing.py
@@ -59,7 +59,7 @@ def test_when_class_is_missing_it_can_be_inferred_from_type():
     expected_html = _wrap_with_body(
         """
     <div class="Page">
-        <aside class='Sidebar'>Some text</aside>
+        <aside class='Sidebar'><p class='Paragraph'>Some text</p></aside>
     </div>
     """
     )
@@ -87,7 +87,7 @@ def test_when_class_is_wrong_tag_name_is_overwritten():
     expected_html = _wrap_with_body(
         """
     <div class="Page">
-        <aside class='Sidebar'>Some text</aside>
+        <aside class='Sidebar'><p class='Paragraph'>Some text</p></aside>
     </div>
     """
     )
@@ -535,6 +535,8 @@ def test_malformed_html():
     # language=HTML
     expected_html = """
     <body class="Document">
+
+      <p class="Paragraph">
      Unclosed comment
      <div class="">
       <p>
@@ -554,6 +556,7 @@ def test_malformed_html():
      <p>
       Paragraph with invalid characters: � � �
      </p>
+     </p>
     </body>
     """
 
@@ -563,3 +566,31 @@ def test_malformed_html():
     parsed_ontology = indent_html(remove_all_ids(ontology.to_html()))
 
     assert parsed_ontology == expected_html
+
+
+def test_text_is_wrapped_inside_layout_element():
+    # language=HTML
+    base_html = _wrap_with_body(
+        """
+    <div class="Page">
+        Text
+    </div>
+    """
+    )
+    base_html = indent_html(base_html)
+
+    # language=HTML
+    expected_html = _wrap_with_body(
+        """
+    <div class="Page">
+        <p class='Paragraph'>Text</p>
+    </div>
+    """
+    )
+
+    expected_html = indent_html(expected_html)
+
+    ontology: OntologyElement = parse_html_to_ontology(base_html)
+    parsed_ontology = indent_html(remove_all_ids(ontology.to_html()))
+
+    assert parsed_ontology == expected_html
diff --git a/unstructured/__version__.py b/unstructured/__version__.py
@@ -1 +1 @@
-__version__ = "0.16.3-dev1"  # pragma: no cover
+__version__ = "0.16.3-dev2"  # pragma: no cover
diff --git a/unstructured/partition/html/transformations.py b/unstructured/partition/html/transformations.py
@@ -79,15 +79,17 @@ def ontology_to_unstructured_elements(
                     ),
                 )
             ]
-
+        childreen = []
         for child in ontology_element.children:
-            elements_to_return += ontology_to_unstructured_elements(
+            childreen += ontology_to_unstructured_elements(
                 child,
                 parent_id=ontology_element.id,
                 page_number=page_number,
                 depth=0 if isinstance(ontology_element, Document) else depth + 1,
                 filename=filename,
             )
+
+        elements_to_return += childreen
     else:
         unstructured_element_class_name = ONTOLOGY_CLASS_NAME_TO_UNSTRUCTURED_ELEMENT_TYPE_NAME[
             ontology_element.__class__.__name__
@@ -98,7 +100,6 @@ def ontology_to_unstructured_elements(
             BeautifulSoup(html_code_of_ontology_element, "html.parser").get_text().strip()
         )
         # TODO value attribute from form input should be added to the text
-
         unstructured_element = element_class(
             text=element_text,
             element_id=ontology_element.id,
@@ -255,8 +256,10 @@ def parse_html_to_ontology_element(soup: Tag) -> OntologyElement | None:
             additional_attributes=escaped_attrs,
         )
 
-    has_children = (ontology_class != UncategorizedText) and any(
-        isinstance(content, Tag) for content in soup.contents
+    has_children = (
+        (ontology_class != UncategorizedText)
+        and any(isinstance(content, Tag) for content in soup.contents)
+        or ontology_class().elementType == ElementTypeEnum.layout
     )
 
     if has_children:
Original file line number	Diff line number	Diff line change
		@@ -1 +1 @@
		__version__ = "0.16.3-dev1" # pragma: no cover
		__version__ = "0.16.3-dev2" # pragma: no cover