feat: map <input> tags by type + add coverage (#4068)

Implements type-aware classification of `<input>` elements in
`extract_tag_and_ontology_class_from_tag` (checkbox → `Checkbox`, radio
→ `RadioButton`, else → `FormFieldValue`) and updates/extends the
HTML-to-ontology test suite to validate the new behaviour.

Author: MaksOpp

CHANGELOG.md

@@ -1,8 +1,10 @@
-## 0.18.11-dev0
+## 0.18.11-dev1
 
 ### Enhancements
 
 ### Features
+- **Type-aware `<input>` mapping in HTML transformations** Bare `<input>` elements are now classified by their `type` attribute (checkbox → Checkbox, radio → RadioButton, others → FormFieldValue).
+
 
 ### Fixes
 - **Recognize '|' as a delimiter** csv parser will now recognize '|' as a delimiter in addition to ',' and ';'.

test_unstructured/partition/html/test_html_to_ontology_parsing.py

@@ -1,6 +1,17 @@
+from typing import Optional, Type
+
+import pytest
 from bs4 import BeautifulSoup
 
-from unstructured.documents.ontology import Form, FormFieldValue, Image, OntologyElement, Page
+from unstructured.documents.ontology import (
+    Checkbox,
+    Form,
+    FormFieldValue,
+    Image,
+    OntologyElement,
+    Page,
+    RadioButton,
+)
 from unstructured.partition.html.html_utils import indent_html
 from unstructured.partition.html.transformations import RECURSION_LIMIT, parse_html_to_ontology
 
@@ -330,7 +341,7 @@ def test_when_unknown_element_keyword_only_attributes_are_preserved_during_mappi
     <div class="Page">
         <form class="Form">
             <label class="FormField" for="option1">
-                <input class="Checkbox" type="radio" name="option1" value="2" checked />
+                <input class="RadioButton" type="radio" name="option1" value="2" checked />
                 <span class="UncategorizedText">
                     Option 1 (Checked)
                 </span>
@@ -713,3 +724,30 @@ def test_uncategorizedtest_has_image_and_no_text():
     element = ontology.children[0].children[0]
     assert type(element) is Image
     assert element.css_class_name == "Image"
+
+
+@pytest.mark.parametrize(
+    ("input_type", "expected_class"),
+    [
+        ("checkbox", Checkbox),
+        ("radio", RadioButton),
+        ("text", FormFieldValue),  # explicit non-specialised type
+        (None, FormFieldValue),  # missing type attribute
+    ],
+)
+def test_input_tag_type_is_mapped_to_correct_ontology_class(
+    input_type: Optional[str], expected_class: Type[OntologyElement]
+) -> None:
+    """Ensure bare <input> tags are classified based on their *type* attribute."""
+
+    type_attr = f' type="{input_type}"' if input_type is not None else ""
+    html_snippet = f'<div class="Page"><input{type_attr} name="field" /></div>'
+
+    page = parse_html_to_ontology(html_snippet)
+    assert len(page.children) == 1
+    element = page.children[0]
+
+    # Validate chosen ontology class and preserved HTML semantics
+    assert isinstance(element, expected_class)
+    assert element.html_tag_name == "input"
+    assert element.css_class_name == expected_class.__name__

unstructured/__version__.py

@@ -1 +1 @@
-__version__ = "0.18.11-dev0"  # pragma: no cover
+__version__ = "0.18.11-dev1"  # pragma: no cover

unstructured/partition/html/transformations.py

@@ -423,11 +423,24 @@ def extract_tag_and_ontology_class_from_tag(
         element_class = CSS_CLASS_TO_ELEMENT_TYPE_MAP.get(soup.attrs["class"][0])
         html_tag = element_class().allowed_tags[0]
 
-    # Scenario 3: CSS class incorrect, but HTML tag correct and exclusive in ontology
+    # Scenario 3: <input> elements, handled explicitly based on their 'type' attribute
+    if not element_class and soup.name == "input":
+        input_type = (str(soup.get("type")) or "").lower()
+        if input_type == "checkbox":
+            element_class = ontology.Checkbox
+        elif input_type == "radio":
+            element_class = ontology.RadioButton
+        else:
+            # Any other input (including missing type or text/number/etc.) is considered
+            # a generic form field value.
+            element_class = ontology.FormFieldValue
+        html_tag = "input"
+
+    # Scenario 4: CSS class incorrect, but HTML tag correct and exclusive in ontology
     if not element_class and soup.name in HTML_TAG_TO_DEFAULT_ELEMENT_TYPE_MAP:
         html_tag, element_class = soup.name, HTML_TAG_TO_DEFAULT_ELEMENT_TYPE_MAP[soup.name]
 
-    # Scenario 4: CSS class incorrect, HTML tag incorrect
+    # Scenario 5: CSS class incorrect, HTML tag incorrect
     # Fallback to default UncategorizedText
     if not element_class:
         # TODO (Pluto): Sometimes we could infer that from parent type and soup.name
@@ -436,7 +449,7 @@ def extract_tag_and_ontology_class_from_tag(
         html_tag = "span"
         element_class = ontology.UncategorizedText
 
-    # Scenario 5: UncategorizedText has image and no text
+    # Scenario 6: UncategorizedText has image and no text
     # Typically, this happens with a span or div tag with an image inside
     if element_class == ontology.UncategorizedText and soup.find("img") and not soup.text.strip():
         element_class = ontology.Image