feat: add RTF (Rich Text Format) converter

[hoangsonww]

Summary

Adds a converter for RTF (Rich Text Format) files, which MarkItDown did not previously support. RTF is a common interchange format (WordPad, TextEdit, legacy office exports), so this fills a real gap in the supported file types.

RtfConverter accepts .rtf files (the file extension and the application/rtf, application/x-rtf, text/rtf, text/richtext mimetypes), strips RTF control words, and returns the underlying text content as Markdown.

Implementation

• New converters/_rtf_converter.py: RtfConverter(DocumentConverter) following the same accepts() / convert() shape as the existing converters. It uses the lightweight, pure-Python striprtf library, imported lazily and gated behind the standard MissingDependencyException (matching the docx/pptx/pdf pattern), so the base install is unaffected.
• Registered in _markitdown.py and exported from converters/__init__.py.
• Added an rtf = ["striprtf"] optional-dependency extra in pyproject.toml and folded striprtf into [all].
• Charset handling mirrors the CSV converter (stream_info.charset with a charset-normalizer fallback).

Scope

This is a v1 text-extraction converter: it preserves the document's textual content with control words removed (the same altitude as the plain-text and CSV converters). It does not attempt to map RTF bold/heading runs onto Markdown syntax, which would require a substantially heavier dependency.

Testing

• Added tests/test_files/test.rtf (a small fixture with heading, bold, italic, and paragraph content) and tests/test_rtf_converter.py (local-path and binary-stream conversion).
• New RTF tests pass, and the full existing offline suite passes with [all] installed (no regressions).
• python -m markitdown test.rtf produces the expected Markdown.
• ruff check and ruff format are clean.

[hoangsonww]

@microsoft-github-policy-service agree