parse dblp

parse_dblp.py

#!/usr/bin/python
# filename: parse_dblp.py
# author: ivanchou
 
import codecs, os
import xml.etree.ElementTree as ET
paper_tag = ('article','inproceedings','proceedings','book',
             'incollection','phdthesis','mastersthesis','www')
 
class AllEntities:
    def __getitem__(self, key):
        return key
 
 
print ('----------parse begin----------')
result = codecs.open('authors','w','utf-8')
 
parser = ET.XMLParser()
parser.parser.UseForeignDTD(True)
parser.entity = AllEntities()
 
for event, article in ET.iterparse('dblp_part.xml', events=("start", "end"), parser=parser):
	for author in article.findall('author'):
		result.write(author.text + u'|')
	if event == 'end' and article.tag in paper_tag:
		result.write(os.linesep)
	article.clear()
print ('----------parse end----------')