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

Merge pull request #434 from mcdull/master

Browse Source 
把 AIRAV 的大姐姐名稱改由 javbus 取得
This commit is contained in:
Yoshiko2
2021-03-02 10:40:25 +08:00
committed by GitHub
parent fa1d99be4c 3d45e27f21
commit a7e2eb210d
1 changed files with 2 additions and 4 deletions
Download Patch File Download Diff File Expand all files Collapse all files

6
WebCrawler/airav.py
View File

@@ -68,15 +68,13 @@ def getRuntime(htmlcode): #获取分钟 已修改
html = etree.fromstring(htmlcode, etree.HTMLParser()) html = etree.fromstring(htmlcode, etree.HTMLParser())
result = str(html.xpath('/html/body/div[5]/div[1]/div[2]/p[3]/text()')).strip(" ['']分鐘") result = str(html.xpath('/html/body/div[5]/div[1]/div[2]/p[3]/text()')).strip(" ['']分鐘")
return result return result
def getActor(htmlcode): #获取女优 def getActor(htmlcode): #获取女优
b=[] b=[]
soup=BeautifulSoup(htmlcode,'lxml') soup=BeautifulSoup(htmlcode,'lxml')
a=soup.find_all(attrs={'class':'videoAvstarListItem'}) a=soup.find_all(attrs={'class':'star-name'})
for i in a: for i in a:
b.append(i.get_text()) b.append(i.get_text())
return b return b
def getNum(htmlcode): #获取番号 def getNum(htmlcode): #获取番号
html = etree.fromstring(htmlcode, etree.HTMLParser()) html = etree.fromstring(htmlcode, etree.HTMLParser())
result = str(html.xpath('/html/body/div[5]/div[1]/div[2]/p[1]/span[2]/text()')).strip(" ['']") result = str(html.xpath('/html/body/div[5]/div[1]/div[2]/p[1]/span[2]/text()')).strip(" ['']")
@@ -190,7 +188,7 @@ def main(number):
# 导演 使用javbus # 导演 使用javbus
'director': getDirector(javbus_htmlcode), 'director': getDirector(javbus_htmlcode),
# 作者 使用airav # 作者 使用airav
'actor': getActor(htmlcode), 'actor': getActor(javbus_htmlcode),
# 发售日使用javbus # 发售日使用javbus
'release': getRelease(javbus_htmlcode), 'release': getRelease(javbus_htmlcode),
# 番号使用javbus # 番号使用javbus