Parse Rust source from Python

parse_rust_files.py

#!/usr/bin/env python3
# Inspired by https://stackoverflow.com/a
# Posted by s-m-e
# Retrieved 2025-12-04, License - CC BY-SA 4.0
#
# Extended by L. Woodtli

import os
import sys
from pathlib import Path

import tree_sitter_rust as tsrust
from tree_sitter import Language, Parser

def get_struct_name(node):
    struct_name = [n.text for n in node.children if n.type == 'type_identifier']
    assert len(struct_name) == 1
    return f"Struct name: {struct_name[0]}"

def get_function_name(node):
    function_name = [n.text for n in node.children if n.type == 'identifier']
    assert len(function_name) == 1
    return f"Function: {function_name[0]}"

def get_struct_and_function_names(node):
    items = []
    for node in node.root_node.children:
        if node.type == 'struct_item':
            n = get_struct_name(node)
            items.append(n)
        elif node.type == 'function_item':
            n = get_function_name(node)
            items.append(n)
    return items

def parse_rust_files(dir):
    output = ""
    for dir_path, _dir_names, file_names in os.walk(dir):
        if 'target' in dir_path:
            continue
        for file_name in file_names:
            if file_name.endswith(".rs"):
                file_path = os.path.join(dir_path, file_name)
                with open(file_path, mode = "rb") as f:
                    raw = f.read()
                    parser = Parser(Language(tsrust.language()))
                    tree = parser.parse(raw)
                    output += f"File: {file_path}\n"
                    names = get_struct_and_function_names(tree)
                    output += "\n".join(names)
                    output += "\n\n"
    return output


if __name__ == "__main__":
    dir = sys.argv[1]
    print(parse_rust_files(dir))