#?-*-?coding:utf-8?-*-
‘‘‘
@author:Zhong?Peng
@createDate:2015-11-05
@version:1.0.0
‘‘‘
import?sys
from?xml.sax?import?handler?make_parser
import?pickle
import?os

DBLP_xml_PATH?=?r‘C:\Users\zhongpeng\Desktop\dblp.xml‘

#此處需要完整添加所有“塊”結(jié)構(gòu)的標(biāo)簽，或者需要處理的類型的標(biāo)簽
paperTag?=?（‘inproceedings‘‘proceedings‘
????????‘incollection‘‘phdthesis‘‘mastersthesis‘‘www‘）



class?CoauthorHandler（handler.ContentHandler）:

????def?__init__（self）:
????????self.title?=?‘‘
????????self.year?=?‘‘
????????self.author?=?‘‘
????????self.count?=?0
????????self.isPaperTag?=?0
????????self.istitleTag?=?0
????????self.isYearTag?=?0
????????self.isAuthorTag?=?0
????????self.authors?=?[]#存儲(chǔ)每個(gè)“塊”中的所有author
????????self.storage?=?{}#用來(lái)存儲(chǔ)生成的數(shù)據(jù)，結(jié)構(gòu)為{‘title‘:[year?[author1?author2?...]]}
??????