Rough-and-ready approximation to Arduino's multi-file sketch assembly process

arduino_assemble.py

#!/usr/bin/python

import glob
import re
import os
import sys


def assemble_paths(paths):
    """Assemble multiple files together into a main sketch.  paths is a list
       of paths to individual files, the first element is assumed to be the
       main file.  The assembled file is written to sys.stdout."""
    # find all prototypes in all input files
    prototypes = []
    for path in paths:
        with open(path) as f:
            m = re.findall(
                r"""^(
                (?:\s*[a-z_\d]+){1,2}       # return type
                \s+[a-z_\d]+\s*             # name of prototype
                \([a-z_,\.\*\&\[\]\s\d]*\)  # args
                )\s*\{                      # must end with {
                """,
                f.read(),
                re.X | re.M | re.I
            )
            prototypes.extend(p.strip() for p in m)
    # write output file header
    sys.stdout.write('#include "Arduino.h"\n')
    # copy the main file, up to the first non comment/preprocessor/blank line
    f = open(paths[0])
    in_comment = False
    for line in f:
        if in_comment:
            if line.find('*/') != -1:
                in_comment = False
        else:
            if line.find('/*') != -1 and line.find('*/') == -1:  # start of multi-line comment
                in_comment = True
            elif not re.match(r'\s*(#.*|//.*)?$', line, re.M):
                break
        sys.stdout.write(line)
    # write out the prototypes
    for p in prototypes:
        sys.stdout.write(p + ';\n')
    # continue writing out the rest of the main file
    sys.stdout.write(line)
    for line in f:
        sys.stdout.write(line)
    f.close()
    # write out the contents of each of the non-main files
    for path in paths[1:]:
        with open(path) as g:
            sys.stdout.write(g.read())


def assemble_directory(directory):
    """Assemble multiple files together in a single directory.  The assembled
       file is written to sys.stdout."""
    # find all *.ino and *.pde files
    paths = glob.glob(directory + '/*.ino')
    paths += glob.glob(directory + '/*.pde')
    # if there's a file whose name matches the name of the directory then
    # assume that this the main file, otherwise assume main.cpp
    main = directory + '/main.cpp'
    for path in paths:
        if os.path.splitext(os.path.basename(path))[0] == os.path.basename(directory):
            main = path
            break
    if main in paths:
        paths.remove(main)
    return assemble_paths([main] + paths)


if __name__ == '__main__':
    assemble_directory(sys.argv[1])