def get_image_urls(response):
#sgml seems not work.htmlparser slow than lxml.user lxml
# img_ex = SgmlLinkExtractor(allow=(), tags=('img',), attrs=('src',))
# img_ex = HtmlParserLinkExtractor(tag='img', attr='src', unique=True)
img_ex = LxmlParserLinkExtractor(tag='img', attr='src', unique=True)
links = img_ex.extract_links(response)
urls = [link.url for link in links]
return urls
def parse_detail_page(self, response):
item = response.meta['item']
html5_response = response_html5parse(response)
hxs = HtmlXPathSelector(html5_response)
texts_hxs = hxs.select('//table[contains(@class,"patent-bibdata")]//tr')
texts = [''.join(hxs.select('.//text()').extract()) for hxs in texts_hxs]
# texts = clean_google_texts(texts)
result_doc = blur_ana_patent(texts)
classification_hxs = hxs.select('//td[text()="國際專利分類號"]/parent::* | '
'//td[text()="国际分类号"]/parent::* |'
'//td[text()="International Classification"]/parent::*')
patent_state = ''.join(hxs.select('//td[text()="出版類型"]/../td[2]//text()').extract())
texts1 = [''.join(classification_hxs.select('.//text()').extract())]
result_doc1 = blur_ana_patent(texts1)
doc = item['doc']
doc.update(result_doc)
doc.update(result_doc1)
patent_name = ''.join(hxs.select('//span[@class="patent-title"]//text()').extract())
abstract = ''.join(
hxs.select('//div[@class="patent-section patent-abstract-section"]//div[@class="patent-text"]//text()')
.extract())
description = ''.join(
hxs.select('//div[@class="patent-section patent-description-section"]//div[@class="patent-text"]//text()')
.extract())
claims = ''.join(
hxs.select('//div[@class="patent-section patent-claims-section"]//div[@class="patent-text"]//text()')
.extract())
doc['patent_name'] = patent_name
doc['abstract'] = abstract
doc['description'] = description
doc['claims'] = claims
doc['patent_state'] = patent_state
doc['patent_type'] = ''
attachments = item['attachments']
attach1 = {
'url': response.url,
'data': response.body_as_unicode(),
'mime_type': get_mime_type_in_response(response)
}
attachments.append(attach1)
image_urls = get_image_urls(response)
item['attachment_urls'] += image_urls
yield self.item_or_request(item)
#如果有中文版本，对中文版本进行抓取
link_ex = LxmlParserLinkExtractor(unique=False)
links = link_ex.extract_links(response)
for link in links:
if link.text in ['Chinese', 'chinese', '中文']:
request = Request(link.url, callback=self.parse_detail_page)
doc = {
'data_source': 'google专利搜索',
'url': link.url,
}
cn_item = PatentItem(doc=doc,
next_request=request, list_url=item['list_url'], query=item['query'],
attachments=[], attachment_urls=[])
yield self.item_or_request(cn_item)
break
