Use the re.UNICODE
flag:
re.UNICODE
Make \w, \W, \b, \B, \d, \D, \s and \S dependent on the Unicode character
properties database.
tweet = u"//@lilei: dd //@Bob: cc//@Girl: dd//@魏武: 利益所致 自然念念不忘// @诺什: 吸引优质 客户,摆脱屌丝男!!!//@MarkGreene: 转发微博"
RTpattern = r'''//?@(\w+)'''
for word in re.findall(RTpattern, tweet, re.UNICODE):
print word
# lilei
# Bob
# Girl
# 魏武
# MarkGreene