python 读取tmx文件

转载自http://www.pianshen.com/article/4142315326/

# -*- coding:utf-8 -*-
import codecs
import xml.etree.ElementTree as ET
import sys

defaultencoding = 'utf-8'

if sys.getdefaultencoding() != defaultencoding:
   reload(sys)
   sys.setdefaultencoding(defaultencoding)
def tmx_process(file, target_file):

   tree = ET.parse(file)
   root = tree.getroot()
   body = root[1]

   txtfile = codecs.open(target_file, 'w+', 'utf-8')
   print(len(body))
   for item in body:
      try:
         bs = item[0][0].text
         zh = item[1][0].text
         bs = bs.decode('utf-8')
         zh = zh.decode('utf-8')
         txtfile.write(zh + r' ||| ' + bs + '\n')
      except Exception as ex:
         print ex.message
         countinue

   txtfile.close()
   return 'Done'