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

Update 1.3

Browse Source
This commit is contained in:
Yoshiko
2019-10-06 01:14:12 +08:00
committed by GitHub
parent ae99c652f5
commit 73c0126fb8
3 changed files with 56 additions and 75 deletions
Download Patch File Download Diff File Expand all files Collapse all files

11
fc2fans_club.py
View File

@@ -4,10 +4,8 @@ import json
import ADC_function
def getTitle(htmlcode): #获取厂商
#print(htmlcode)
html = etree.fromstring(htmlcode,etree.HTMLParser())
result = str(html.xpath('//*[@id="container"]/div[1]/div/article/section[1]/h2/text()')).strip(" ['']")
#print(result2)
return result
def getActor(htmlcode):
try:
@@ -26,21 +24,14 @@ def getStudio(htmlcode): #获取厂商
def getNum(htmlcode): #获取番号
html = etree.fromstring(htmlcode, etree.HTMLParser())
result = str(html.xpath('/html/body/div[5]/div[1]/div[2]/p[1]/span[2]/text()')).strip(" ['']")
#print(result)
return result
def getRelease(htmlcode2): #
#a=ADC_function.get_html('http://adult.contents.fc2.com/article_search.php?id='+str(number).lstrip("FC2-").lstrip("fc2-").lstrip("fc2_").lstrip("fc2-")+'&utm_source=aff_php&utm_medium=source_code&utm_campaign=from_aff_php')
html=etree.fromstring(htmlcode2,etree.HTMLParser())
result = str(html.xpath('//*[@id="container"]/div[1]/div/article/section[1]/div/div[2]/dl/dd[4]/text()')).strip(" ['']")
return result
def getCover(htmlcode2): #获取厂商 #
#a = ADC_function.get_html('http://adult.contents.fc2.com/article_search.php?id=' + str(number).lstrip("FC2-").lstrip("fc2-").lstrip("fc2_").lstrip("fc2-") + '&utm_source=aff_php&utm_medium=source_code&utm_campaign=from_aff_php')
html = etree.fromstring(htmlcode2, etree.HTMLParser())
result = str(html.xpath('//*[@id="container"]/div[1]/div/article/section[1]/div/div[1]/a/img/@src')).strip(" ['']")
# if result == '':
# html = etree.fromstring(htmlcode, etree.HTMLParser())
# result2 = str(html.xpath('//*[@id="slider"]/ul[1]/li[1]/img/@src')).strip(" ['']")
# return result2
return 'http:' + result
def getOutline(htmlcode2): #获取番号 #
html = etree.fromstring(htmlcode2, etree.HTMLParser())
@@ -58,7 +49,7 @@ def getYear(release):
return ''
def main(number):
number=number.replace('PPV','').replace('ppv','').strip('fc2_').strip('fc2-').strip('ppv-').strip('PPV-').strip('FC2_').strip('FC2-').strip('ppv-').strip('PPV-')
number=number.replace('PPV','').replace('ppv','').strip('fc2_').strip('fc2-').strip('ppv-').strip('PPV-').strip('FC2_').strip('FC2-').strip('ppv-').strip('PPV-').replace('fc2ppv-','').replace('FC2PPV-','')
htmlcode2 = ADC_function.get_html('http://adult.contents.fc2.com/article_search.php?id='+str(number).lstrip("FC2-").lstrip("fc2-").lstrip("fc2_").lstrip("fc2-")+'')
#htmlcode = ADC_function.get_html('http://fc2fans.club/html/FC2-' + number + '.html')
dic = {