magic_codec is a small utility to make writng preprocessors for Python easier. This uses a custom codec to kick off preprocessing before passing the result to the Python interpreter. You can find a more in-depth explanation of how this works over at pydong.org

Currently the following preprocessors are available:

• braces Python with braces - inspired by Bython
  • magic line: # coding: magic.braces
  • example: python tests/braces/test.by
• incdec Extends python with unary prefix ++i and postfix i++ increment/decrement expressions
  • magic line: # coding: magic.incdec
  • example: python tests/incdec/incdec.py
• cpp lets the Python interpreter interpret C++ via cppyy
  • magic line: #define CODEC "coding:magic.cpp"
  • example: python tests/cpp/test.cpp
• toml validate toml files using json schemas
  • magic line: # coding: magic.toml
  • example: python tests/toml/data_valid.toml -s tests/toml/schema.json

Builtins can be loaded by setting the codec to magic.builtin_name where builtin_name is the name of the builtin.

To extend magic_codec with your own preprocessors, you can create another Python package whose name is prefixed with magic_. Setting the codec to magic_foo would load the magic_foo package and check if it has a function preprocess.

The expected signature of preprocess is as follows:

def preprocess(data: str) -> str:
    raise NotImplementedError

You can find an example extension in example.