fix(epub): forward conversion options to the HTML converter

[kevin (kevin9327)]

What this changes

Conversion options are ignored for EPUB. keep_data_uris=True keeps the image source for
HTML, DOCX, XLSX and PPTX, but not for the same image inside an EPUB chapter:

.html  ->  ![diagram](data:image/png;base64,iVBORw0KGgoAAAANSUhEUg==)
.epub  ->  ![diagram](data:image/png;base64...)

Why

EpubConverter.convert accepts **kwargs and then calls the HTML converter without them:

converted_content = self._html_converter.convert(
    f,
    StreamInfo(
        mimetype=mimetype,
        extension=extension,
        filename=filename,
    ),
)

Every other converter that renders through HtmlConverter passes the bag along:

# _docx_converter.py
return self._html_converter.convert_string(html_result, **kwargs)

# _xlsx_converter.py
self._html_converter.convert_string(html_content, **kwargs)

# _pptx_converter.py
self._html_converter.convert_string(html_table, **kwargs).markdown.strip()

keep_data_uris is the option with a visible effect today -- it is a CLI flag
(--keep-data-uris) and reaches the converter through convert_stream/convert -- but the
gap covers everything the caller passes, including the markdownify options
(heading_style, bullets, ...) and strict.

Fix

Forward **kwargs, exactly as the sibling converters do. One line.

Tests

Added to packages/markitdown/tests/test_module_misc.py, over a minimal in-memory EPUB whose
one chapter carries a data-URI image:

• keep_data_uris=True keeps the full data URI
• pinned: without the option the URI is still truncated to data:image/png;base64..., and
  the payload does not appear
• pinned: metadata, title and chapter text are unchanged

Against unmodified main:

E  AssertionError: assert '![diagram](data:image/png;base64,iVBORw0KGgoAAAANSUhEUg==)' in
   '**Title:** Example book\n\nChapter text.\n\n![diagram](data:image/png;base64...)'
1 failed, 3 passed, 65 deselected

With the fix:

4 passed, 65 deselected

How I tested

Windows 11, Python 3.12, editable install of packages/markitdown[all].

pytest tests/test_module_misc.py -k epub   ->  1 failed, 3 passed  (before)
pytest tests/test_module_misc.py -k epub   ->  4 passed            (after)
pytest tests/                              ->  18 failed, 429 passed, 4 skipped

Those 18 are unchanged by this PR: main gives 18 failed, 426 passed, 4 skipped on this
machine before any edit, and the 3 added tests account for the difference. They are the CLI
stdout-encoding, Windows file-URI and speech-transcription tests that need a UTF-8 console, a
case-sensitive path and network/ffmpeg.

black --check clean on both files.

[Copilot]

🟢 Approval recommended

The focused fix matches sibling converters and has appropriate regression coverage.

Pull request overview

Forwards conversion options from EPUB processing to the underlying HTML converter.

Changes:

• Passes **kwargs when converting EPUB chapters.
• Adds regression coverage for data-URI handling and unchanged metadata/text.

File summaries

File	Description
packages/markitdown/src/markitdown/converters/_epub_converter.py	Forwards conversion options.
packages/markitdown/tests/test_module_misc.py	Adds in-memory EPUB regression tests.

Review details

• Files reviewed: 2/2 changed files
• Comments generated: 0
• Review effort level: Balanced

---

💡 Add a code-review agent skill or configure MCP servers for context-aware, tailored reviews. Learn more in the docs.