fix(doc-intel): include HTML in the default DocumentIntelligenceConverter file types

The constructor documents file_types as "Defaults to all supported file
types", but the default list has never contained
DocumentIntelligenceFileType.HTML, even though the enum lists it under
"# No OCR" and _get_mime_type_prefixes / _get_file_extensions /
_analysis_features all handle it.

So accepts() rejects .html, text/html and application/xhtml+xml unless
the caller passes file_types explicitly. The HTML mappings added in #1352
are unreachable through the default constructor, and MarkItDown never
passes file_types when it builds the converter from docintel_endpoint.

Author: Anai-Guo

packages/markitdown/src/markitdown/converters/_doc_intel_converter.py

@@ -140,6 +140,7 @@ def __init__(
             DocumentIntelligenceFileType.DOCX,
             DocumentIntelligenceFileType.PPTX,
             DocumentIntelligenceFileType.XLSX,
+            DocumentIntelligenceFileType.HTML,
             DocumentIntelligenceFileType.PDF,
             DocumentIntelligenceFileType.JPEG,
             DocumentIntelligenceFileType.PNG,
@@ -251,4 +252,4 @@ def convert(
 
         # remove comments from the markdown content generated by Doc Intelligence and append to markdown string
         markdown_text = re.sub(r"<!--.*?-->", "", result.content, flags=re.DOTALL)
-        return DocumentConverterResult(markdown=markdown_text)
+        return DocumentConverterResult(markdown=markdown_text)

packages/markitdown/tests/test_docintel_html.py

@@ -1,3 +1,4 @@
+import inspect
 import io
 from markitdown.converters._doc_intel_converter import (
     DocumentIntelligenceConverter,
@@ -6,6 +7,11 @@
 from markitdown._stream_info import StreamInfo
 
 
+def _default_file_types():
+    params = inspect.signature(DocumentIntelligenceConverter.__init__).parameters
+    return params["file_types"].default
+
+
 def _make_converter(file_types):
     conv = DocumentIntelligenceConverter.__new__(DocumentIntelligenceConverter)
     conv._file_types = file_types
@@ -24,3 +30,19 @@ def test_docintel_accepts_html_mimetype():
     assert conv.accepts(io.BytesIO(b""), stream_info)
     stream_info = StreamInfo(mimetype="application/xhtml+xml", extension=None)
     assert conv.accepts(io.BytesIO(b""), stream_info)
+
+
+def test_docintel_default_file_types_cover_every_supported_type():
+    # The file_types docstring promises "Defaults to all supported file types",
+    # so the default must stay in sync with the enum.
+    assert set(_default_file_types()) == set(DocumentIntelligenceFileType)
+
+
+def test_docintel_default_accepts_html():
+    conv = _make_converter(_default_file_types())
+    for stream_info in (
+        StreamInfo(mimetype=None, extension=".html"),
+        StreamInfo(mimetype="text/html", extension=None),
+        StreamInfo(mimetype="application/xhtml+xml", extension=None),
+    ):
+        assert conv.accepts(io.BytesIO(b""), stream_info)