fix: simplify test and address broken test

Signed-off-by: Yusik Kim <[email protected]>
DS4SD · Feb 19, 2025 · a241ebc · a241ebc
1 parent 69350aa
commit a241ebc
Show file tree

Hide file tree

Showing 3 changed files with 26 additions and 173 deletions.
diff --git a/docling_core/types/doc/document.py b/docling_core/types/doc/document.py
@@ -2778,7 +2778,7 @@ def _image_fallback(item: TextItem):
 
                 # Building a math equation in MathML format
                 # ref https://www.w3.org/TR/wai-aria-1.1/#math
-                elif formula_to_mathml and img_fallback is not None:
+                elif formula_to_mathml:
                     try:
                         mathml_element = latex2mathml.converter.convert_to_element(
                             math_formula, display="block"
@@ -2794,9 +2794,13 @@ def _image_fallback(item: TextItem):
                             "Malformed formula cannot be rendered. "
                             f"Error {err.__class__.__name__}, formula={math_formula}"
                         )
-                        if image_mode == ImageRefMode.EMBEDDED and len(item.prov) > 0:
+                        if (
+                            image_mode == ImageRefMode.EMBEDDED
+                            and len(item.prov) > 0
+                            and img_fallback is not None
+                        ):
                             text = img_fallback
-                        else:
+                        elif len(math_formula) > 0:
                             text = f"<pre>{math_formula}</pre>"
 
                 elif math_formula != "":

diff --git a/test/test_docling_doc.py b/test/test_docling_doc.py
@@ -775,6 +775,25 @@ def test_formula_mathml():
     assert doc_html == gt_html
 
 
+def test_formula_with_missing_fallback():
+    doc = DoclingDocument(name="Dummy")
+    bbox = BoundingBox.from_tuple((1, 2, 3, 4), origin=CoordOrigin.BOTTOMLEFT)
+    prov = ProvenanceItem(page_no=1, bbox=bbox, charspan=(0, 2))
+    doc.add_text(label=DocItemLabel.FORMULA, text="", orig="(II.24) 2 Imar", prov=prov)
+
+    actual = doc.export_to_html(
+        formula_to_mathml=True, html_head="", image_mode=ImageRefMode.EMBEDDED
+    )
+
+    expected = """<!DOCTYPE html>
+<html lang="en">
+
+<div class="formula-not-decoded">Formula not decoded</div>
+</html>"""
+
+    assert actual == expected
+
+
 def test_docitem_get_image():
     # Prepare the document
     doc = DoclingDocument(name="Dummy")

diff --git a/test/test_export_to_html.py b/test/test_export_to_html.py