Files for "How to Translate an RTF File: What Survives, What Breaks, and What to Check"
https://www.booktranslator.app/blog/translate-rtf-file

check_rtf.py
  Python 3 standard library only.
  python3 check_rtf.py sample-word-style.rtf                                   inspect a source file
  python3 check_rtf.py sample-word-style.rtf sample-word-style-translated.rtf  compare a translation
  It reads the RTF markup directly. It is a pre-flight check, not a full RTF parser.

sample-textedit.rtf
  Made on macOS with: textutil -convert rtf (Apple's text system, the same one TextEdit uses).
  Lines that ended in a space were joined to the next line so the file has no trailing
  whitespace; RTF ignores raw line breaks, so the content is unchanged.
  A heading, bold and italic words, a link, a bulleted list, a 2x2 table, and a line with
  braces and backslashes. Its paragraphs end with a backslash + line break, not \par.

sample-textedit-source.html
  The HTML that sample-textedit.rtf was converted from.

sample-word-style.rtf
  Hand-written in the style Microsoft Word uses: \par paragraph breaks, a page header,
  a footnote, a HYPERLINK field, a 2x2 table, a 1x1 PNG image, a page break,
  a cp1252 accent (\'e9) and two Unicode characters (\u26085? \u26412?).

sample-long-numbers.rtf
  Three short paragraphs: a 20-digit order reference, a 40-character commit hash,
  and a phone number.

*-translated.rtf, *-bilingual.rtf
  What came back when each sample went through a local copy of BookTranslator's RTF
  parsing and writing code (backend main 89dcc394, 2026-11-03). No model was used:
  every paragraph that would be sent for translation was replaced with "[ZH] " + the
  original text, so the files show exactly what is sent and how the file is rebuilt.