fix(docx): properly escape special markdown characters at start of lines

packages/markitdown/src/markitdown/converters/_markdownify.py

@@ -122,5 +122,13 @@ def convert_input(
             return "[x] " if el.has_attr("checked") else "[ ] "
         return ""
 
+    def process_text(self, text: str) -> str:
+        text = super().process_text(text)  # type: ignore
+        # Escape markdown block characters at the beginning of the text node or line
+        # to prevent them from being parsed as headings, lists, blockquotes, etc.
+        # markdownify handles some inline escaping, but we need to ensure block starters are escaped.
+        text = re.sub(r"(?m)^([#>+\-=|])", r"\\\1", text)
+        return text
+
     def convert_soup(self, soup: Any) -> str:
         return super().convert_soup(soup)  # type: ignore

packages/markitdown/tests/test_docx_escaping.py

@@ -0,0 +1,12 @@
+import pytest
+from markitdown.converters._html_converter import HtmlConverter
+
+def test_markdown_special_character_escaping():
+    html_content = "<p># Hello World</p><p>> Quote</p><p>- List</p>"
+    converter = HtmlConverter()
+    result = converter.convert_string(html_content)
+
+    # Verify that the special characters at the start of paragraphs are escaped
+    assert r"\# Hello World" in result.markdown
+    assert r"\> Quote" in result.markdown
+    assert r"\- List" in result.markdown