python查找字符串

#!/usr/bin/env python2.7 import os import sys import string def get_hash(basedir): hash = {} outchain = open("./chain",'w') for path in os.walk(basedir): if len(path[2]) == 0 : continue else: for file in path[2]: fullname = file.split('.') if fullname[1] == "chain": file_name = path[0] + "/" + file doc_id =path[0][path[0].find("annotations")+12:] + "/" + fullname[0] infile = open(file_name) part_num = "000" while True: line = infile.readline() if len(line) == 0 : break if line.startswith("Chain"): part_num = line.split()[1].split('-')[0] elif not line.startswith("\n"): line_list = line.split() position = line_list[0] token_list = line_list[1:] tokens = " ".join(token_list) sentence_id = position.split('.')[0] token_id = position.split('.')[1] key = doc_id + ".part" + part_num + "." + str(string.atoi(sentence_id)) + "." + str(string.atoi(token_id.split('-')[0])) + "-" + str(string.atoi(token_id.split('-')[1])) hash[key] = tokens outchain.write(key+"\t"+tokens+"\n") infile.close() outchain.close() return hash lang = "english" #can change to english or chinese base_dir = "/home/lzj/CONLLdata/" + lang infile_name = "./mention" if lang == "english": infile_name = infile_name + "_en" elif lang == "chinese": infile_name = infile_name + "_ch" else: infile_name = infile_name + "_ar" in_mention_file = "./in_mention" in_chain_file = "./in_chain" in_both_file = "./in_both" chain_hash = get_hash(base_dir) print "get hash finished." infile = open(infile_name) in_mention = open(in_mention_file,'w') in_chain = open(in_chain_file,'w') in_both = open(in_both_file,'w') while True: line = infile.readline() if len(line) == 0 : break key = line.split("\t")[0] if chain_hash.has_key(key): in_both.write(line) del chain_hash[key] else: in_mention.write(line) left_chain_items = chain_hash.items() for item in left_chain_items: in_chain.write(item[0]+"\t"+item[1]+"\n") print "search finished."