Celeste langfile parser (to translator-friendly CSV format)

celeste-parse.py

#!/usr/bin/env python2

import sys
import csv
import re

_in = sys.argv[1] # input .txt
try:
    _out = sys.argv[2] # output .csv
except IndexError:
    _out = 'output.csv'

if not _in:
    raise IOError, 'No input file specified'

keys = []
result = {}
mode = None
lastkey = None

regex_comment = re.compile('^\t*(#.*)')
regex_key = re.compile('^\t(?P<key>[A-Za-z0-9_]+)=[\t ]*(?P<value>.*)$')
regex_indented = re.compile('^\t\t(.+)')

with open(_in, 'r') as f:
    content = f.readlines()

for line in content:
    if lastkey:
        if line.isspace() or regex_indented.match(line):
            result[lastkey] += line.strip() + '\n'
            continue
        else:
            result[lastkey] = result[lastkey].rstrip()
            lastkey = None

    comment_match = regex_comment.match(line)
    key_match = regex_key.match(line)

    if comment_match:
        keys.append(line.strip())
    elif key_match:
        groups = key_match.groupdict()
        key = groups['key']
        value = groups['value']
        keys.append(key)
        if value == '':
            lastkey = key
            result[key] = ''
        else:
            result[key] = value

with open(_out, 'w+') as f:
    writer = csv.writer(f)
    for i in keys:
        if i[0] == '#':
            writer.writerow([i])
        else:
            writer.writerow([i, result.get(i)])