feat: add 6 new converters, fix 7 bugs, add document stats utility

[AKIB473]

Summary

This PR adds significant new functionality to MarkItDown: 6 new file format converters, 7 bug fixes across existing converters, and a new document stats utility.

---

🐛 Bug Fixes (7)

1. _ipynb_converter.py — Cell outputs now rendered

Cell outputs (stdout, execute_result, display_data, errors/tracebacks) were completely ignored. Now rendered as plain text or fenced code blocks.

2. _xlsx_converter.py — Empty sheets handled gracefully

Empty sheets previously produced a broken heading with no table. Now outputs _No data_ instead.

3. _zip_converter.py — Conversion failures surfaced as warnings

FileConversionException was silently swallowed. Now outputs > ⚠️ Could not convert file: {name} — {reason} so users know what was skipped.

4. _pptx_converter.py — Chart table separator fixed

Separator used |---|---| format which breaks some Markdown renderers. Changed to | --- | --- | (consistent with rest of codebase).

5. _audio_converter.py — Added .ogg, .flac, .aac support

These common audio formats were silently rejected. Added to both ACCEPTED_FILE_EXTENSIONS and ACCEPTED_MIME_TYPE_PREFIXES.

6. _wikipedia_converter.py — Infoboxes stripped by default

Wikipedia infoboxes, navboxes, and wikitables cluttered LLM-oriented output. Added strip_wikipedia_infoboxes=True kwarg (default: True) that removes .infobox, .wikitable, .navbox, .metadata elements.

7. _rss_converter.py — Item/entry links now extracted

RSS <link> text nodes and Atom <link href="..."> attributes were never included in output. Now rendered as **Link:** [url](url) per item/entry.

---

✨ New Converters (6)

TomlConverter — .toml files

Converts pyproject.toml, Cargo.toml, config.toml etc. to structured Markdown with section headers and key-value lists.

SitemapConverter — XML sitemaps

Converts sitemap.xml (urlset and sitemapindex) to Markdown link lists with lastmod/priority metadata.

EnvConverter — .env / dotenv files

Converts .env files to a Markdown table. Values are masked by default to prevent secrets leaking into LLM context. Pass show_values=True to reveal.

YamlConverter — .yaml / .yml files

Converts YAML files to structured Markdown — dicts as ## Section headers, lists as bullet points, scalars as - **key:** value.

RequirementsConverter — requirements.txt / Pipfile

Converts Python dependency files to a Markdown table with columns: Package | Version Constraint | Notes (inline # comments extracted as notes).

HtmlConverter enhancement — OG/meta extraction

Added optional include_html_metadata=True kwarg to extract Open Graph tags, author, description, keywords as a YAML front-matter block.

---

🛠️ New Utility: get_document_stats()

Added get_document_stats(markdown_text: str) -> dict to converters module:

from markitdown.converters import get_document_stats

stats = get_document_stats(result.markdown)
# {'word_count': 342, 'char_count': 2187, 'line_count': 45,
#  'heading_count': 5, 'code_block_count': 3, 'link_count': 8, 'image_count': 2}

Also available as DocumentConverterResult.stats() method directly on any conversion result.

---

🧪 Tests

• 75 tests added across 2 test files
• All passing: 75 passed in 1.86s
• Full coverage of every new converter (accepts/rejects + output correctness) and every bug fix

[AKIB473]

@microsoft-github-policy-service agree

[AKIB473]

Closing this PR to resubmit as smaller, focused pull requests — one per converter/fix — to make review easier. Will follow up with separate PRs shortly.