[pornhub] Fix uploader extraction and extract counts

This commit is contained in:
Sergey M․ 2014-03-22 21:29:01 +07:00
parent 56dd55721c
commit 0320ddc192

View File

@ -8,6 +8,7 @@
compat_urllib_parse_urlparse, compat_urllib_parse_urlparse,
compat_urllib_request, compat_urllib_request,
compat_urllib_parse, compat_urllib_parse,
str_to_int,
) )
from ..aes import ( from ..aes import (
aes_decrypt_text aes_decrypt_text
@ -27,6 +28,12 @@ class PornHubIE(InfoExtractor):
} }
} }
def _extract_count(self, pattern, webpage, name):
count = self._html_search_regex(pattern, webpage, '%s count' % name, fatal=False)
if count:
count = str_to_int(count)
return count
def _real_extract(self, url): def _real_extract(self, url):
mobj = re.match(self._VALID_URL, url) mobj = re.match(self._VALID_URL, url)
video_id = mobj.group('videoid') video_id = mobj.group('videoid')
@ -37,11 +44,19 @@ def _real_extract(self, url):
webpage = self._download_webpage(req, video_id) webpage = self._download_webpage(req, video_id)
video_title = self._html_search_regex(r'<h1 [^>]+>([^<]+)', webpage, 'title') video_title = self._html_search_regex(r'<h1 [^>]+>([^<]+)', webpage, 'title')
video_uploader = self._html_search_regex(r'<b>From: </b>(?:\s|<[^>]*>)*(.+?)<', webpage, 'uploader', fatal=False) video_uploader = self._html_search_regex(
r'(?s)<div class="video-info-row">\s*From:&nbsp;.+?<(?:a href="/users/|<span class="username)[^>]+>(.+?)<',
webpage, 'uploader', fatal=False)
thumbnail = self._html_search_regex(r'"image_url":"([^"]+)', webpage, 'thumbnail', fatal=False) thumbnail = self._html_search_regex(r'"image_url":"([^"]+)', webpage, 'thumbnail', fatal=False)
if thumbnail: if thumbnail:
thumbnail = compat_urllib_parse.unquote(thumbnail) thumbnail = compat_urllib_parse.unquote(thumbnail)
view_count = self._extract_count(r'<span class="count">([\d,\.]+)</span> views', webpage, 'view')
like_count = self._extract_count(r'<span class="votesUp">([\d,\.]+)</span>', webpage, 'like')
dislike_count = self._extract_count(r'<span class="votesDown">([\d,\.]+)</span>', webpage, 'dislike')
comment_count = self._extract_count(
r'All comments \(<var class="videoCommentCount">([\d,\.]+)</var>', webpage, 'comment')
video_urls = list(map(compat_urllib_parse.unquote , re.findall(r'"quality_[0-9]{3}p":"([^"]+)', webpage))) video_urls = list(map(compat_urllib_parse.unquote , re.findall(r'"quality_[0-9]{3}p":"([^"]+)', webpage)))
if webpage.find('"encrypted":true') != -1: if webpage.find('"encrypted":true') != -1:
password = compat_urllib_parse.unquote_plus(self._html_search_regex(r'"video_title":"([^"]+)', webpage, 'password')) password = compat_urllib_parse.unquote_plus(self._html_search_regex(r'"video_title":"([^"]+)', webpage, 'password'))
@ -77,6 +92,10 @@ def _real_extract(self, url):
'uploader': video_uploader, 'uploader': video_uploader,
'title': video_title, 'title': video_title,
'thumbnail': thumbnail, 'thumbnail': thumbnail,
'view_count': view_count,
'like_count': like_count,
'dislike_count': dislike_count,
'comment_count': comment_count,
'formats': formats, 'formats': formats,
'age_limit': 18, 'age_limit': 18,
} }