[pornhub] Fix extraction (closes #12007)

[youtube-dl] / youtube_dl / extractor / pornhub.py

diff --git a/youtube_dl/extractor/pornhub.py b/youtube_dl/extractor/pornhub.py
index 0724efc09963d1f84fc7b24ec99d89b58e1bdfb6..5e930f45e0ac271a3c37e5bffa4a23f791a2ebf0 100644 (file)
--- a/youtube_dl/extractor/pornhub.py
+++ b/youtube_dl/extractor/pornhub.py
@@ -33,7 +33,7 @@ class PornHubIE(InfoExtractor):
                             (?:[a-z]+\.)?pornhub\.com/(?:view_video\.php\?viewkey=|embed/)|
                             (?:www\.)?thumbzilla\.com/video/
                         )
-                        (?P<id>[0-9a-z]+)
+                        (?P<id>[\da-z]+)
                     '''
     _TESTS = [{
         'url': 'http://www.pornhub.com/view_video.php?viewkey=648719015',
@@ -96,12 +96,11 @@ class PornHubIE(InfoExtractor):
         'only_matching': True,
     }]
 
-    @classmethod
-    def _extract_url(cls, webpage):
-        mobj = re.search(
-            r'<iframe[^>]+?src=(["\'])(?P<url>(?:https?:)?//(?:www\.)?pornhub\.com/embed/\d+)\1', webpage)
-        if mobj:
-            return mobj.group('url')
+    @staticmethod
+    def _extract_urls(webpage):
+        return re.findall(
+            r'<iframe[^>]+?src=["\'](?P<url>(?:https?:)?//(?:www\.)?pornhub\.com/embed/[\da-z]+)',
+            webpage)
 
     def _extract_count(self, pattern, webpage, name):
         return str_to_int(self._search_regex(
@@ -157,7 +156,25 @@ class PornHubIE(InfoExtractor):
         comment_count = self._extract_count(
             r'All Comments\s*<span>\(([\d,.]+)\)', webpage, 'comment')
 
-        video_urls = list(map(compat_urllib_parse_unquote, re.findall(r"player_quality_[0-9]{3}p\s*=\s*'([^']+)'", webpage)))
+        video_variables = {}
+        for video_variablename, quote, video_variable in re.findall(
+                r'(player_quality_[0-9]{3,4}p[0-9a-z]+?)=\s*(["\'])(.*?)\2;', webpage):
+            video_variables[video_variablename] = video_variable
+
+        encoded_video_urls = []
+        for encoded_video_url in re.findall(
+                r'player_quality_[0-9]{3,4}p\s*=(.*?);', webpage):
+            encoded_video_urls.append(encoded_video_url)
+
+        # Decode the URLs 
+        video_urls = []
+        for url in encoded_video_urls:
+            for varname, varval in video_variables.items():
+                url = url.replace(varname, varval)
+            url = url.replace('+', '')
+            url = url.replace(' ', '')
+            video_urls.append(url)
+
         if webpage.find('"encrypted":true') != -1:
             password = compat_urllib_parse_unquote_plus(
                 self._search_regex(r'"video_title":"([^"]+)', webpage, 'password'))
@@ -230,7 +247,14 @@ class PornHubPlaylistBaseIE(InfoExtractor):
 
         webpage = self._download_webpage(url, playlist_id)
 
-        entries = self._extract_entries(webpage)
+        # Only process container div with main playlist content skipping
+        # drop-down menu that uses similar pattern for videos (see
+        # https://github.com/rg3/youtube-dl/issues/11594).
+        container = self._search_regex(
+            r'(?s)(<div[^>]+class=["\']container.+)', webpage,
+            'container', default=webpage)
+
+        entries = self._extract_entries(container)
 
         playlist = self._parse_json(
             self._search_regex(
@@ -244,12 +268,12 @@ class PornHubPlaylistBaseIE(InfoExtractor):
 class PornHubPlaylistIE(PornHubPlaylistBaseIE):
     _VALID_URL = r'https?://(?:www\.)?pornhub\.com/playlist/(?P<id>\d+)'
     _TESTS = [{
-        'url': 'http://www.pornhub.com/playlist/6201671',
+        'url': 'http://www.pornhub.com/playlist/4667351',
         'info_dict': {
-            'id': '6201671',
-            'title': 'P0p4',
+            'id': '4667351',
+            'title': 'Nataly Hot',
         },
-        'playlist_mincount': 35,
+        'playlist_mincount': 2,
     }]