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

修正 AVSOX 重定向到发布页的问题

Browse Source 
AVSOX 硬编码的地址有时候会 302 跳转到发布页，不如直接从发布页获取最新地址。
This commit is contained in:
Wait
2020-08-19 03:07:56 +08:00
committed by GitHub
parent b6dd790d19
commit e7ad2b085a
1 changed files with 6 additions and 4 deletions
Download Patch File Download Diff File Expand all files Collapse all files

10
WebCrawler/avsox.py
View File

@@ -83,15 +83,17 @@ def getSeries(htmlcode):
return ''
def main(number):
a = get_html('https://avsox.host/cn/search/' + number)
html = get_html('https://tellme.pw/avsox')
site = etree.HTML(html).xpath('//div[@class="container"]/div/a/@href')[0]
a = get_html(site + '/cn/search/' + number)
html = etree.fromstring(a, etree.HTMLParser()) # //table/tr[1]/td[1]/text()
result1 = str(html.xpath('//*[@id="waterfall"]/div/a/@href')).strip(" ['']")
if result1 == '' or result1 == 'null' or result1 == 'None':
a = get_html('https://avsox.host/cn/search/' + number.replace('-', '_'))
a = get_html(site + '/cn/search/' + number.replace('-', '_'))
html = etree.fromstring(a, etree.HTMLParser()) # //table/tr[1]/td[1]/text()
result1 = str(html.xpath('//*[@id="waterfall"]/div/a/@href')).strip(" ['']")
if result1 == '' or result1 == 'null' or result1 == 'None':
a = get_html('https://avsox.host/cn/search/' + number.replace('_', ''))
a = get_html(site + '/cn/search/' + number.replace('_', ''))
html = etree.fromstring(a, etree.HTMLParser()) # //table/tr[1]/td[1]/text()
result1 = str(html.xpath('//*[@id="waterfall"]/div/a/@href')).strip(" ['']")
web = get_html(result1)
@@ -121,4 +123,4 @@ def main(number):
return js
if __name__ == "__main__":
print(main('012717_472'))
print(main('012717_472'))