bacow/AV_Data_Capture
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

Fix search func in source dlsite

Browse Source
This commit is contained in:
yoshiko2
2022-05-20 20:07:10 +08:00
parent da76d33b8b
commit f82d6a5317
1 changed files with 9 additions and 1 deletions
Download Patch File Download Diff File Expand all files Collapse all files

10
WebCrawler/dlsite.py
View File

@@ -120,7 +120,15 @@ def main(number):
else:
htmlcode = get_html(f'https://www.dlsite.com/maniax/fsr/=/language/jp/sex_category/male/keyword/{number}/order/trend/work_type_category/movie',cookies={'locale': 'zh-cn'})
html = etree.HTML(htmlcode)
a = html.xpath('//*[@id="search_result_img_box"]/li[1]/dl/dd[2]/div[2]/a/@href')[0]
search_result = html.xpath('//*[@id="search_result_img_box"]/li[1]/dl/dd[2]/div[2]/a/@href')
if len(search_result) == 0:
number = number.replace("THE ANIMATION", "").replace("THE ANIMATION", "").replace('上巻','').replace('下巻','').replace('前編','').replace('後編','')
htmlcode = get_html(
f'https://www.dlsite.com/maniax/fsr/=/language/jp/sex_category/male/keyword/{number}/order/trend/work_type_category/movie',
cookies={'locale': 'zh-cn'})
html = etree.HTML(htmlcode)
search_result = html.xpath('//*[@id="search_result_img_box"]/li[1]/dl/dd[2]/div[2]/a/@href')
a = search_result[0]
html = etree.HTML(get_html(a,cookies={'locale': 'zh-cn'}))
number = str(re.findall("\wJ\w+",a)).strip(" [']")
dic = {