Naive brainfuck-to-C transpiler

bf2c.py

#!/usr/bin/env python3
import argparse
import sys

def tokenize(source):
    return list(source)

def parse(tokens):
    ast = []
    stack = [ast]
    node = ast
    for token in tokens:
        if token == '[':
            node.append([])
            stack.append(node)
            node = node[-1]
        elif token == ']':
            node = stack.pop()
        elif token in "+-<>,.":
            node.append(token)
    return ast

def compile_to_c(tree, depth=1):
    buf = ""
    def emit(code):
        nonlocal buf
        buf += ("  " * depth) + code + "\n"
    for node in tree:
        if isinstance(node, list):
            emit("while (*p) {")
            emit(compile_to_c(node, depth=depth+1).rstrip()[2:])
            emit("}")
        elif node == '+': emit("++*p;")
        elif node == '-': emit("--*p;")
        elif node == '>': emit("++p;")
        elif node == '<': emit("--p;")
        elif node == ",": emit("*p = getchar();")
        elif node == ".": emit("putchar(*p);")
    return buf

if __name__ == '__main__':
    argparser = argparse.ArgumentParser(description="Transpile a Brainfuck program to a different host language")
    argparser.add_argument('infile', nargs='?', type=argparse.FileType('r'), \
        default='-', help="File to read source from, defaults to stdin")
    argparser.add_argument('-t', '--target', action='store', type=str, \
        choices=set(['c']), default='c', help="Output language to target")
    args = argparser.parse_args()

    code = args.infile.read()
    tokens = tokenize(code)
    ast = parse(tokens)

    if args.target == 'c':
        print("#include <stdio.h>")
        print("#include <stdlib.h>")
        print("int main(int argc, char **argv) {")
        print("  int *p = (int *)malloc(30000 * sizeof(int));")
        print(compile_to_c(ast))
        print("  return 0;")
        print("}")

helloworld.bf

>+++++++++
[<++++++++>-]
<.>+++++++
[<++++>-]
<+.+++++++..+++.>>>++++++++
[<++++>-]
<.>>>++++++++++
[<+++++++++>-]
<---.<<<<.+++.------.--------.>>+.

helloworld.c

#include <stdio.h>
#include <stdlib.h>
int main(int argc, char **argv) {
  int *p = (int *)malloc(30000 * sizeof(int));
  ++p;
  ++*p;
  ++*p;
  ++*p;
  ++*p;
  ++*p;
  ++*p;
  ++*p;
  ++*p;
  ++*p;
  while (*p) {
    --p;
    ++*p;
    ++*p;
    ++*p;
    ++*p;
    ++*p;
    ++*p;
    ++*p;
    ++*p;
    ++p;
    --*p;
  }
  --p;
  putchar(*p);
  ++p;
  ++*p;
  ++*p;
  ++*p;
  ++*p;
  ++*p;
  ++*p;
  ++*p;
  while (*p) {
    --p;
    ++*p;
    ++*p;
    ++*p;
    ++*p;
    ++p;
    --*p;
  }
  --p;
  ++*p;
  putchar(*p);
  ++*p;
  ++*p;
  ++*p;
  ++*p;
  ++*p;
  ++*p;
  ++*p;
  putchar(*p);
  putchar(*p);
  ++*p;
  ++*p;
  ++*p;
  putchar(*p);
  ++p;
  ++p;
  ++p;
  ++*p;
  ++*p;
  ++*p;
  ++*p;
  ++*p;
  ++*p;
  ++*p;
  ++*p;
  while (*p) {
    --p;
    ++*p;
    ++*p;
    ++*p;
    ++*p;
    ++p;
    --*p;
  }
  --p;
  putchar(*p);
  ++p;
  ++p;
  ++p;
  ++*p;
  ++*p;
  ++*p;
  ++*p;
  ++*p;
  ++*p;
  ++*p;
  ++*p;
  ++*p;
  ++*p;
  while (*p) {
    --p;
    ++*p;
    ++*p;
    ++*p;
    ++*p;
    ++*p;
    ++*p;
    ++*p;
    ++*p;
    ++*p;
    ++p;
    --*p;
  }
  --p;
  --*p;
  --*p;
  --*p;
  putchar(*p);
  --p;
  --p;
  --p;
  --p;
  putchar(*p);
  ++*p;
  ++*p;
  ++*p;
  putchar(*p);
  --*p;
  --*p;
  --*p;
  --*p;
  --*p;
  --*p;
  putchar(*p);
  --*p;
  --*p;
  --*p;
  --*p;
  --*p;
  --*p;
  --*p;
  --*p;
  putchar(*p);
  ++p;
  ++p;
  ++*p;
  putchar(*p);

  return 0;
}