Converter Bot · md2gost

A Markdown to DOCX converter that follows GOST 7.32-2017 (the Russian standard for research and student reports), plus a Telegram bot for study routine: send a .md and get a report ready to submit, send a project .zip and get a .txt with the folder tree and file contents. The converter is installable on its own as the md2gost package on PyPI. The bot speaks Russian.

Status: personal project, maintained · bot @my_convbot · package md2gost

Stack: Python 3.9+ · python-docx · aiogram 3 · Docker

md2gost — Markdown → DOCX per GOST

pip install md2gost
md2gost report.md                 # writes report.docx next to it
md2gost report.md -o out.docx --no-heading-numbers

What it does to the document:

margins: left 30 mm, right 15, top and bottom 20; Times New Roman 14 pt, 1.5 line spacing, 1.25 cm first-line indent, justified text
page numbers at the bottom center, none on the title page
section numbering 1., 1.1., 1.1.1.; structural elements ("Введение", "Заключение", "Список литературы" and others) unnumbered, uppercase, centered
a page break before every second-level section
bulleted lists with dashes and nesting, numbered lists as 1) with numbering restarted per list
tables with a "Таблица N" caption on the top left, code blocks and inline code in a monospace font, quotes, footnotes [^1], bold and italic

Command-line options: --font, --size, --spacing, --no-heading-numbers, --no-page-numbers, --number-title-page. From Python:

from md2gost import DocumentSettings, MarkdownToDocxConverter

settings = DocumentSettings()
settings.auto_numbering_headings = True
MarkdownToDocxConverter(settings).convert("report.md", "report.docx")

The Markdown parser is custom and line-based: nested tables and lists inside tables are not supported.

The bot

Send	Get
`.md`	a GOST-formatted `.docx` (the same md2gost, with heading numbers)
project `.zip`	a `.txt`: folder tree and contents of text files with line numbers, handy for an LLM or a report appendix

Files up to 20 MB. Archives are checked before extraction: at most 10,000,000 files and 15 GB unpacked. Service folders (.git, node_modules, __pycache__, build…) and binary files are skipped. Conversion runs in a separate thread, so the bot never freezes on big files.

git clone https://github.com/EDeev/converterbot.git && cd converterbot
cp .env.example .env      # BOT_TOKEN from @BotFather
docker compose up -d

Prebuilt image: docker pull ghcr.io/edeev/converterbot or docker pull dcr.deev.su/edeev/converterbot. Without Docker: pip install -r requirements.txt, then BOT_TOKEN=… python bot.py.

rep_to_txt.py also works on its own: python rep_to_txt.py path/to/project.

Development

pip install -r requirements-dev.txt
ruff check --select E9,F,B . && pytest

CI tests the package on Python 3.9, 3.12 and 3.13 and builds it. On v* tags the package is published to PyPI and the bot's Docker image to GitHub Packages and dcr.deev.su.

License

MIT — see LICENSE.

Author

Egor Deev — GitHub · Telegram · egor@deev.space

4 KiB Raw Blame History