Dekker1/Fourmi
Archived
1
0
Fork 0
Code

chemspider new_synonym is now PEP-8 compliant

Browse Source
This commit is contained in:
RTB 2014-04-18 16:07:06 +02:00
parent 319e028717
commit 479182d77e
1 changed files with 12 additions and 9 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

21
FourmiCrawler/parsers/ChemSpider.py
View File

@ -119,9 +119,9 @@ class ChemSpider(Parser):
def new_synonym(self, sel, name, category): def new_synonym(self, sel, name, category):
self.ignore_list.append(name) self.ignore_list.append(name)
language = sel.xpath('span[@class="synonym_language"]/text()').extract() language = sel.xpath('span[@class="synonym_language"]/text()')
if language: if language:
language = language[0][1:-1] language = language.extract()[0][1:-1]
else: else:
language = 'English' language = 'English'
log.msg('CS synonym: %s (%s) (%s)' % (name, category, language), log.msg('CS synonym: %s (%s) (%s)' % (name, category, language),
@ -129,7 +129,10 @@ class ChemSpider(Parser):
references = [] references = []
for ref in sel.xpath('span[@class="synonym_ref"]'): for ref in sel.xpath('span[@class="synonym_ref"]'):
refname = ref.xpath('normalize-space(string())') refname = ref.xpath('normalize-space(string())')
references.append({'name': refname.extract()[0][1:-1], 'URI': ''}) references.append({
'name': refname.extract()[0][1:-1],
'URI': ''
})
for ref in sel.xpath('a[@class="synonym_ref"]'): for ref in sel.xpath('a[@class="synonym_ref"]'):
references.append({ references.append({
'name': ref.xpath('@title').extract()[0], 'name': ref.xpath('@title').extract()[0],
@ -137,13 +140,13 @@ class ChemSpider(Parser):
}) })
for ref in references: for ref in references:
log.msg('CS synonym ref: %s %s' % (ref['name'], ref['URI']), log.msg('CS synonym ref: %s %s' % (ref['name'], ref['URI']),
level=log.DEBUG) level=log.DEBUG)
synonym = { synonym = {
'name': name, 'name': name,
'category': category, 'category': category,
'language': language, 'language': language,
'references': references 'references': references
} }
return synonym return synonym
def parse_extendedinfo(self, response): def parse_extendedinfo(self, response):