If you've ever moved complex math from Python code into documentation, Javadoc, or a scientific paper, you know the pain. Manually typesetting multi-story formulas in LaTeX is a special kind of masochism, where one forgotten } costs twenty minutes of debugging.

The latexify package does exactly one thing: translates Python functions into LaTeX mathematical formulas.

This is not another LLM wrapper. The library works in a maximally deterministic way: it uses the built-in ast module to parse the abstract syntax tree of your Python function, and translates AST nodes into LaTeX markup according to strict rules.

What it can do at the AST level:
🟢 Recognizes and converts math and numpy modules (math.sqrt becomes \sqrt, math.gamma becomes \Gamma).
🟢 Expands if/elif/else into beautiful piecewise functions.
🟢 Renders NumPy matrices.
🟢 Understands set operations (&, | turn into \cap, \cup if you set use_set_symbols=True).
🟢 Can generate not only formulas but also pseudocode (decorator @latexify.algorithmic).

In Jupyter Notebook or Colab, the @latexify.function decorator automatically makes the cell render a beautiful formula. You don't even need to call print().

🔗 Source code is on GitHub, and you can try it live in Google Colab.

#годный_опенсорс