Py2C — Python to C Transpiler (MVP)

Py2C is a lightweight MVP transpiler that converts Python 3.10 source code into valid C code using AST transformation. It is designed to demonstrate the feasibility of converting high-level language constructs to C while maintaining correctness and basic structure.

Note: Source code comments are primarily in Chinese, as this is an MVP stage prototype.

*Note: This project was built by a first-year undergraduate student as an initial attempt to explore compiler design.

Features

Supported

• Variables and expressions

  • Arithmetic: +, -, *, /, %, **
  • Comparison and logical operators
  • Automatic type inference: int, double, char

• Control flow

  • if / elif / else
  • while, for x in range(n)
  • break, continue, pass

• Functions

  • Definition and invocation
  • Return values
  • Type inference for parameters and return types

• print()

  • Supports multi-argument
  • Automatically chooses format specifier (%d, %f, %s)

• Classes and objects

  • class converted to struct
  • init, methods, attribute access
  • self mapped to struct pointer

• Lists

  • Simple list converted to C array (no slicing or append)

• Global code

  • All top-level code placed inside main()

Not supported (output as comments in generated C code)

• try/except/finally
• with
• import, from ... import
• dict, set, tuple
• lambda, decorators, yield, async/await

Usage

python3 py2ast.py example.py > example_ast.json go run ast2c.go example_ast.json > example.c gcc -o example example.c ./example

Example

The included example.py demonstrates support for:

• Functions
• Class definition and usage
• Control structures
• Print output

The corresponding C output compiles via GCC and executes partially with correct output.

Notes

• Some generated C files may require manually adding missing #include lines.
• Output structure is formatted with readable indentation and fallback comments.
• Only one test file (example.py) has been verified to compile and partially execute correctly.

Contact

I'd be glad to receive feedback if you'd like.

Email:[email protected]

License

MIT License. See LICENSE for details.