Dekker1/Fourmi
Archived
1
0
Fork 0
Code

added newrresult function to WikipediaParser.py

Browse Source
This commit is contained in:
RTB 2014-06-06 19:30:31 +02:00
parent 68139b4839
commit 69664d3ac0
1 changed files with 9 additions and 0 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

9
FourmiCrawler/sources/WikipediaParser.py
View File

@ -124,3 +124,12 @@ class WikipediaParser(Source):
links = sel.xpath('//span[contains(concat(" ",normalize-space(@class)," "),"reflink")]/a' links = sel.xpath('//span[contains(concat(" ",normalize-space(@class)," "),"reflink")]/a'
'[contains(concat(" ",normalize-space(@class)," "),"external")]/@href').extract() '[contains(concat(" ",normalize-space(@class)," "),"external")]/@href').extract()
return links return links
def newresult(self, attribute, value):
return Result({
'attribute': attribute,
'value': value,
'source': 'Wikipedia',
'reliability': self.cfg['reliability'],
'conditions': ''
})