A quick script to run through files in a directory changing leading tabs to spaces, trimming trailing whitespace, and converting line endings to UNIX (LF).

tabs2spaces.py

#!/usr/bin/env python

import argparse
import fnmatch
import os
import sys

DEFAULT_EXTENSIONS = ['py', 'css', 'html', 'js', 'txt', 'json', 'sh', 'sql', 'yml', 'yaml', 'xml']


def fix_line(line, opts):
    line = line.rstrip()
    if not line:
        return line
    tabs = 0
    spaces = 0
    for ch in line:
        if ch == '\t':
            tabs += 1
            spaces = 0
        elif ch == ' ':
            spaces += 1
            if spaces == opts.spaces:
                tabs += 1
                spaces = 0
        else:
            break
    indent = (tabs * opts.spaces) + spaces
    return ' ' * indent + line.lstrip()


def fix_file(path, opts):
    for exclude in opts.exclude:
        if fnmatch.fnmatch(path, exclude):
            return
    if os.path.splitext(path)[1][1:] not in opts.ext:
        return
    print(path)
    if opts.dry_run:
        return
    lines = [fix_line(line, opts) for line in open(path, 'r')]
    data = '\n'.join(lines).strip()
    with open(path, 'w') as f:
        f.write(data)
        if data:
            f.write('\n')


def fix_directories(opts):
    for path in opts.dirs:
        if os.path.isfile(path):
            fix_file(os.path.abspath(path), opts)
        else:
            for root, dirs, files in os.walk(path):
                for name in files:
                    if name.startswith('.'):
                        continue
                    fix_file(os.path.join(root, name), opts)


if __name__ == '__main__':
    parser = argparse.ArgumentParser(
        description='Replaces leading tabs with spaces, and fixes line endings.')
    parser.add_argument('-s', '--spaces', type=int, default=4)
    parser.add_argument('-e', '--ext', action='append')
    parser.add_argument('-x', '--exclude', action='append', default=[])
    parser.add_argument('-n', '--dry-run', action='store_true', default=False)
    parser.add_argument('dirs', nargs='+')
    opts = parser.parse_args()
    if not opts.ext:
        opts.ext = DEFAULT_EXTENSIONS
    fix_directories(opts)