feat: keep input tag's class attr in table (#4064)

This change affects partition html.

Previously when there is a table in the html, we clean any tags inside
the table of their class and id attributes except for the class
attribute for `img` tags. This change also preserves the class attribute
for `input` tags inside a table. This change is reflected in a table
element's metadata.text_as_html attribute.

Author: badGarnet

CHANGELOG.md

@@ -21,6 +21,8 @@
 
 ### Enhancements
 
+- **`text_as_html` for Table element now keeps both `input` and `img` tag's `class` attribute** Previously in partition HTML any tag inside a table is stripped of its `class` attribute. Now this attribute is preserved for both `input` and `img` tag in the table element's `metadata.text_as_html`.
+
 ### Features
 - **Add language detection for PDFs** Add document and element level language detection to PDFs.
 

test_unstructured/documents/test_ontology_to_unstructured_parsing.py

@@ -37,6 +37,10 @@ def test_remove_ids_and_class_from_table():
             <td><IMG class="Signature" alt="cell 3"/></td>
             <td>cell 4</td>
         </tr>
+        <tr>
+            <td><input class="Checkbox" type="checkbox"/></td>
+            <td>Option 1</td>
+        </tr>
     </table>
     """
     soup = BeautifulSoup(html_text, "html.parser")
@@ -52,6 +56,10 @@ def test_remove_ids_and_class_from_table():
 <td><img alt="cell 3" class="Signature"/></td>
 <td>cell 4</td>
 </tr>
+<tr>
+<td><input class="Checkbox" type="checkbox"/></td>
+<td>Option 1</td>
+</tr>
 </table>
 """
     )

test_unstructured/partition/html/test_partition_v2.py

@@ -71,5 +71,7 @@ def test_attr_and_html_inside_table_cell_is_kept():
         html_parser_version="v2",
     )
 
-    assert '<input checked="" type="checkbox"/>' in table.metadata.text_as_html  # class is removed
+    assert (
+        '<input checked="" class="Checkbox" type="checkbox"/>' in table.metadata.text_as_html
+    )  # class is removed
     assert 'colspan="2"' in table.metadata.text_as_html

unstructured/documents/ontology.py

@@ -147,13 +147,16 @@ def page_number(self) -> int | None:
         return None
 
 
-def remove_ids_and_class_from_table(soup: BeautifulSoup) -> BeautifulSoup:
+def remove_ids_and_class_from_table(
+    soup: BeautifulSoup, class_attr_to_keep: list[str] = ["img", "input"]
+) -> BeautifulSoup:
     """
     Remove id and class attributes from tags inside tables,
-    except preserve class attributes for img tags.
+    except preserve class attributes for selected tags.
 
     Args:
         soup: BeautifulSoup object containing the HTML
+        class_attr_to_keep: a list of tag names whose class attr will be kept
 
     Returns:
         BeautifulSoup: Modified soup with attributes removed
@@ -162,7 +165,7 @@ def remove_ids_and_class_from_table(soup: BeautifulSoup) -> BeautifulSoup:
         if tag.name.lower() == "table":  # type: ignore
             continue  # We keep table tag
         tag.attrs.pop("id", None)  # type: ignore
-        if tag.name.lower() != "img":  # type: ignore
+        if tag.name.lower() not in class_attr_to_keep:  # type: ignore
             tag.attrs.pop("class", None)  # type: ignore
     return soup