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

fc2.py: update

Browse Source
This commit is contained in:
lededev
2021-09-26 11:38:48 +08:00
parent 8cfefc60ef
commit 161f4063b9
1 changed files with 11 additions and 10 deletions
Download Patch File Download Diff File Expand all files Collapse all files

21
WebCrawler/fc2.py
View File

@@ -46,9 +46,8 @@ def getCover_fc2com(htmlcode2): #获取厂商 #
# print(ADC_function.get_html('https://adult.contents.fc2.com'+path,cookies={'wei6H':'1'})) # print(ADC_function.get_html('https://adult.contents.fc2.com'+path,cookies={'wei6H':'1'}))
# result = str(html.xpath('/html/body/div/text()')).strip(" ['']").replace("\\n",'',10000).replace("'",'',10000).replace(', ,','').strip(' ').replace('。,',',') # result = str(html.xpath('/html/body/div/text()')).strip(" ['']").replace("\\n",'',10000).replace("'",'',10000).replace(', ,','').strip(' ').replace('。,',',')
# return result # return result
def getTag_fc2com(number): #获取番号 def getTag_fc2com(lx):
htmlcode = str(bytes(ADC_function.get_html('http://adult.contents.fc2.com/api/v4/article/'+number+'/tag?'),'utf-8').decode('unicode-escape')) result = lx.xpath("//a[@class='tag tagTag']/text()")
result = re.findall('"tag":"(.*?)"', htmlcode)
tag = [] tag = []
for i in result: for i in result:
tag.append(ADC_function.translateTag_to_sc(i)) tag.append(ADC_function.translateTag_to_sc(i))
@@ -89,26 +88,28 @@ def main(number):
try: try:
number = number.replace('FC2-', '').replace('fc2-', '') number = number.replace('FC2-', '').replace('fc2-', '')
htmlcode2 = ADC_function.get_html('https://adult.contents.fc2.com/article/' + number + '/') htmlcode2 = ADC_function.get_html('https://adult.contents.fc2.com/article/' + number + '/')
#print(htmlcode2)
actor = getActor_fc2com(htmlcode2) actor = getActor_fc2com(htmlcode2)
if getActor_fc2com(htmlcode2) == '': if not actor:
actor = 'FC2系列' actor = '素人'
lx = etree.fromstring(htmlcode2, etree.HTMLParser())
cover = str(lx.xpath("//div[@class='items_article_MainitemThumb']/span/img/@src")).strip(" ['']")
dic = { dic = {
'title': getTitle_fc2com(htmlcode2), 'title': lx.xpath('/html/head/title/text()')[0],
'studio': getStudio_fc2com(htmlcode2), 'studio': getStudio_fc2com(htmlcode2),
'year': getYear_fc2com(getRelease_fc2com(htmlcode2)), 'year': getYear_fc2com(getRelease_fc2com(htmlcode2)),
'outline': '', # getOutline_fc2com(htmlcode2), 'outline': '', # getOutline_fc2com(htmlcode2),
'runtime': '', 'runtime': str(lx.xpath("//p[@class='items_article_info']/text()")[0]),
'director': getStudio_fc2com(htmlcode2), 'director': getStudio_fc2com(htmlcode2),
'actor': actor, 'actor': actor,
'release': getRelease_fc2com(htmlcode2), 'release': getRelease_fc2com(htmlcode2),
'number': 'FC2-' + number, 'number': 'FC2-' + number,
'label': '', 'label': '',
'cover': getCover_fc2com(htmlcode2), 'cover': cover,
'thumb': cover,
'extrafanart': getExtrafanart(htmlcode2), 'extrafanart': getExtrafanart(htmlcode2),
"trailer": getTrailer(htmlcode2, number), "trailer": getTrailer(htmlcode2, number),
'imagecut': 0, 'imagecut': 0,
'tag': getTag_fc2com(number), 'tag': getTag_fc2com(lx),
'actor_photo': '', 'actor_photo': '',
'website': 'https://adult.contents.fc2.com/article/' + number + '/', 'website': 'https://adult.contents.fc2.com/article/' + number + '/',
'source': 'https://adult.contents.fc2.com/article/' + number + '/', 'source': 'https://adult.contents.fc2.com/article/' + number + '/',