]> Raphaël G. Git Repositories - youtubedl/blobdiff - youtube_dl/extractor/dailymotion.py
Merge tag 'upstream/2015.11.10'
[youtubedl] / youtube_dl / extractor / dailymotion.py
index 6685c94a3d6b283e0b7f2240ebfcf35ce462edc2..bc782393173f315b096a640bf20bfa45cb89eb8d 100644 (file)
+# coding: utf-8
+from __future__ import unicode_literals
+
 import re
 import json
 import itertools
 
 from .common import InfoExtractor
-from .subtitles import SubtitlesInfoExtractor
 
-from ..utils import (
-    compat_urllib_request,
+from ..compat import (
     compat_str,
-    get_element_by_attribute,
-    get_element_by_id,
-    orderedSet,
-    str_to_int,
-
+    compat_urllib_request,
+)
+from ..utils import (
     ExtractorError,
+    determine_ext,
+    int_or_none,
+    parse_iso8601,
+    str_to_int,
+    unescapeHTML,
 )
 
+
 class DailymotionBaseInfoExtractor(InfoExtractor):
     @staticmethod
     def _build_request(url):
         """Build a request with the family filter disabled"""
         request = compat_urllib_request.Request(url)
-        request.add_header('Cookie', 'family_filter=off')
-        request.add_header('Cookie', 'ff=off')
+        request.add_header('Cookie', 'family_filter=off; ff=off')
         return request
 
-class DailymotionIE(DailymotionBaseInfoExtractor, SubtitlesInfoExtractor):
-    """Information Extractor for Dailymotion"""
+    def _download_webpage_handle_no_ff(self, url, *args, **kwargs):
+        request = self._build_request(url)
+        return self._download_webpage_handle(request, *args, **kwargs)
+
+    def _download_webpage_no_ff(self, url, *args, **kwargs):
+        request = self._build_request(url)
+        return self._download_webpage(request, *args, **kwargs)
+
 
+class DailymotionIE(DailymotionBaseInfoExtractor):
     _VALID_URL = r'(?i)(?:https?://)?(?:(www|touch)\.)?dailymotion\.[a-z]{2,3}/(?:(embed|#)/)?video/(?P<id>[^/?_]+)'
-    IE_NAME = u'dailymotion'
+    IE_NAME = 'dailymotion'
 
     _FORMATS = [
-        (u'stream_h264_ld_url', u'ld'),
-        (u'stream_h264_url', u'standard'),
-        (u'stream_h264_hq_url', u'hq'),
-        (u'stream_h264_hd_url', u'hd'),
-        (u'stream_h264_hd1080_url', u'hd180'),
+        ('stream_h264_ld_url', 'ld'),
+        ('stream_h264_url', 'standard'),
+        ('stream_h264_hq_url', 'hq'),
+        ('stream_h264_hd_url', 'hd'),
+        ('stream_h264_hd1080_url', 'hd180'),
     ]
 
     _TESTS = [
         {
-            u'url': u'http://www.dailymotion.com/video/x33vw9_tutoriel-de-youtubeur-dl-des-video_tech',
-            u'file': u'x33vw9.mp4',
-            u'md5': u'392c4b85a60a90dc4792da41ce3144eb',
-            u'info_dict': {
-                u"uploader": u"Amphora Alex and Van .", 
-                u"title": u"Tutoriel de Youtubeur\"DL DES VIDEO DE YOUTUBE\""
+            'url': 'https://www.dailymotion.com/video/x2iuewm_steam-machine-models-pricing-listed-on-steam-store-ign-news_videogames',
+            'md5': '2137c41a8e78554bb09225b8eb322406',
+            'info_dict': {
+                'id': 'x2iuewm',
+                'ext': 'mp4',
+                'title': 'Steam Machine Models, Pricing Listed on Steam Store - IGN News',
+                'description': 'Several come bundled with the Steam Controller.',
+                'thumbnail': 're:^https?:.*\.(?:jpg|png)$',
+                'duration': 74,
+                'timestamp': 1425657362,
+                'upload_date': '20150306',
+                'uploader': 'IGN',
+                'uploader_id': 'xijv66',
+                'age_limit': 0,
+                'view_count': int,
+                'comment_count': int,
             }
         },
         # Vevo video
         {
-            u'url': u'http://www.dailymotion.com/video/x149uew_katy-perry-roar-official_musi',
-            u'file': u'USUV71301934.mp4',
-            u'info_dict': {
-                u'title': u'Roar (Official)',
-                u'uploader': u'Katy Perry',
-                u'upload_date': u'20130905',
+            'url': 'http://www.dailymotion.com/video/x149uew_katy-perry-roar-official_musi',
+            'info_dict': {
+                'title': 'Roar (Official)',
+                'id': 'USUV71301934',
+                'ext': 'mp4',
+                'uploader': 'Katy Perry',
+                'upload_date': '20130905',
             },
-            u'params': {
-                u'skip_download': True,
+            'params': {
+                'skip_download': True,
             },
-            u'skip': u'VEVO is only available in some countries',
+            'skip': 'VEVO is only available in some countries',
         },
         # age-restricted video
         {
-            u'url': u'http://www.dailymotion.com/video/xyh2zz_leanna-decker-cyber-girl-of-the-year-desires-nude-playboy-plus_redband',
-            u'file': u'xyh2zz.mp4',
-            u'md5': u'0d667a7b9cebecc3c89ee93099c4159d',
-            u'info_dict': {
-                u'title': 'Leanna Decker - Cyber Girl Of The Year Desires Nude [Playboy Plus]',
-                u'uploader': 'HotWaves1012',
-                u'age_limit': 18,
+            'url': 'http://www.dailymotion.com/video/xyh2zz_leanna-decker-cyber-girl-of-the-year-desires-nude-playboy-plus_redband',
+            'md5': '0d667a7b9cebecc3c89ee93099c4159d',
+            'info_dict': {
+                'id': 'xyh2zz',
+                'ext': 'mp4',
+                'title': 'Leanna Decker - Cyber Girl Of The Year Desires Nude [Playboy Plus]',
+                'uploader': 'HotWaves1012',
+                'age_limit': 18,
             }
-
+        },
+        # geo-restricted, player v5
+        {
+            'url': 'http://www.dailymotion.com/video/xhza0o',
+            'only_matching': True,
         }
     ]
 
     def _real_extract(self, url):
-        # Extract id and simplified title from URL
-        mobj = re.match(self._VALID_URL, url)
+        video_id = self._match_id(url)
 
-        video_id = mobj.group('id')
+        webpage = self._download_webpage_no_ff(
+            'https://www.dailymotion.com/video/%s' % video_id, video_id)
 
-        url = 'http://www.dailymotion.com/video/%s' % video_id
+        age_limit = self._rta_search(webpage)
 
-        # Retrieve video webpage to extract further information
-        request = self._build_request(url)
-        webpage = self._download_webpage(request, video_id)
+        description = self._og_search_description(webpage) or self._html_search_meta(
+            'description', webpage, 'description')
+
+        view_count = str_to_int(self._search_regex(
+            [r'<meta[^>]+itemprop="interactionCount"[^>]+content="UserPlays:(\d+)"',
+             r'video_views_count[^>]+>\s+([\d\.,]+)'],
+            webpage, 'view count', fatal=False))
+        comment_count = int_or_none(self._search_regex(
+            r'<meta[^>]+itemprop="interactionCount"[^>]+content="UserComments:(\d+)"',
+            webpage, 'comment count', fatal=False))
+
+        player_v5 = self._search_regex(
+            [r'buildPlayer\(({.+?})\);', r'playerV5\s*=\s*dmp\.create\([^,]+?,\s*({.+?})\);'],
+            webpage, 'player v5', default=None)
+        if player_v5:
+            player = self._parse_json(player_v5, video_id)
+            metadata = player['metadata']
+
+            self._check_error(metadata)
+
+            formats = []
+            for quality, media_list in metadata['qualities'].items():
+                for media in media_list:
+                    media_url = media.get('url')
+                    if not media_url:
+                        continue
+                    type_ = media.get('type')
+                    if type_ == 'application/vnd.lumberjack.manifest':
+                        continue
+                    ext = determine_ext(media_url)
+                    if type_ == 'application/x-mpegURL' or ext == 'm3u8':
+                        m3u8_formats = self._extract_m3u8_formats(
+                            media_url, video_id, 'mp4', m3u8_id='hls', fatal=False)
+                        if m3u8_formats:
+                            formats.extend(m3u8_formats)
+                    elif type_ == 'application/f4m' or ext == 'f4m':
+                        f4m_formats = self._extract_f4m_formats(
+                            media_url, video_id, preference=-1, f4m_id='hds', fatal=False)
+                        if f4m_formats:
+                            formats.extend(f4m_formats)
+                    else:
+                        f = {
+                            'url': media_url,
+                            'format_id': quality,
+                        }
+                        m = re.search(r'H264-(?P<width>\d+)x(?P<height>\d+)', media_url)
+                        if m:
+                            f.update({
+                                'width': int(m.group('width')),
+                                'height': int(m.group('height')),
+                            })
+                        formats.append(f)
+            self._sort_formats(formats)
+
+            title = metadata['title']
+            duration = int_or_none(metadata.get('duration'))
+            timestamp = int_or_none(metadata.get('created_time'))
+            thumbnail = metadata.get('poster_url')
+            uploader = metadata.get('owner', {}).get('screenname')
+            uploader_id = metadata.get('owner', {}).get('id')
+
+            subtitles = {}
+            for subtitle_lang, subtitle in metadata.get('subtitles', {}).get('data', {}).items():
+                subtitles[subtitle_lang] = [{
+                    'ext': determine_ext(subtitle_url),
+                    'url': subtitle_url,
+                } for subtitle_url in subtitle.get('urls', [])]
+
+            return {
+                'id': video_id,
+                'title': title,
+                'description': description,
+                'thumbnail': thumbnail,
+                'duration': duration,
+                'timestamp': timestamp,
+                'uploader': uploader,
+                'uploader_id': uploader_id,
+                'age_limit': age_limit,
+                'view_count': view_count,
+                'comment_count': comment_count,
+                'formats': formats,
+                'subtitles': subtitles,
+            }
 
-        # Extract URL, uploader and title from webpage
-        self.report_extraction(video_id)
+        # vevo embed
+        vevo_id = self._search_regex(
+            r'<link rel="video_src" href="[^"]*?vevo.com[^"]*?video=(?P<id>[\w]*)',
+            webpage, 'vevo embed', default=None)
+        if vevo_id:
+            return self.url_result('vevo:%s' % vevo_id, 'Vevo')
 
-        # It may just embed a vevo video:
-        m_vevo = re.search(
-            r'<link rel="video_src" href="[^"]*?vevo.com[^"]*?videoId=(?P<id>[\w]*)',
-            webpage)
-        if m_vevo is not None:
-            vevo_id = m_vevo.group('id')
-            self.to_screen(u'Vevo video detected: %s' % vevo_id)
-            return self.url_result(u'vevo:%s' % vevo_id, ie='Vevo')
+        # fallback old player
+        embed_page = self._download_webpage_no_ff(
+            'https://www.dailymotion.com/embed/video/%s' % video_id,
+            video_id, 'Downloading embed page')
 
-        age_limit = self._rta_search(webpage)
+        timestamp = parse_iso8601(self._html_search_meta(
+            'video:release_date', webpage, 'upload date'))
 
-        video_upload_date = None
-        mobj = re.search(r'<div class="[^"]*uploaded_cont[^"]*" title="[^"]*">([0-9]{2})-([0-9]{2})-([0-9]{4})</div>', webpage)
-        if mobj is not None:
-            video_upload_date = mobj.group(3) + mobj.group(2) + mobj.group(1)
-
-        embed_url = 'http://www.dailymotion.com/embed/video/%s' % video_id
-        embed_page = self._download_webpage(embed_url, video_id,
-                                            u'Downloading embed page')
-        info = self._search_regex(r'var info = ({.*?}),$', embed_page,
-            'video info', flags=re.MULTILINE)
-        info = json.loads(info)
-        if info.get('error') is not None:
-            msg = 'Couldn\'t get video, Dailymotion says: %s' % info['error']['title']
-            raise ExtractorError(msg, expected=True)
+        info = self._parse_json(
+            self._search_regex(
+                r'var info = ({.*?}),$', embed_page,
+                'video info', flags=re.MULTILINE),
+            video_id)
+
+        self._check_error(info)
 
         formats = []
         for (key, format_id) in self._FORMATS:
@@ -124,7 +225,7 @@ class DailymotionIE(DailymotionBaseInfoExtractor, SubtitlesInfoExtractor):
             if video_url is not None:
                 m_size = re.search(r'H264-(\d+)x(\d+)', video_url)
                 if m_size is not None:
-                    width, height = m_size.group(1), m_size.group(2)
+                    width, height = map(int_or_none, (m_size.group(1), m_size.group(2)))
                 else:
                     width, height = None, None
                 formats.append({
@@ -134,94 +235,133 @@ class DailymotionIE(DailymotionBaseInfoExtractor, SubtitlesInfoExtractor):
                     'width': width,
                     'height': height,
                 })
-        if not formats:
-            raise ExtractorError(u'Unable to extract video URL')
+        self._sort_formats(formats)
 
         # subtitles
         video_subtitles = self.extract_subtitles(video_id, webpage)
-        if self._downloader.params.get('listsubtitles', False):
-            self._list_available_subtitles(video_id, webpage)
-            return
 
-        view_count = self._search_regex(
-            r'video_views_count[^>]+>\s+([\d\.,]+)', webpage, u'view count', fatal=False)
-        if view_count is not None:
-            view_count = str_to_int(view_count)
+        title = self._og_search_title(webpage, default=None)
+        if title is None:
+            title = self._html_search_regex(
+                r'(?s)<span\s+id="video_title"[^>]*>(.*?)</span>', webpage,
+                'title')
 
         return {
-            'id':       video_id,
+            'id': video_id,
             'formats': formats,
-            'uploader': info['owner_screenname'],
-            'upload_date':  video_upload_date,
-            'title':    self._og_search_title(webpage),
-            'subtitles':    video_subtitles,
+            'uploader': info['owner.screenname'],
+            'timestamp': timestamp,
+            'title': title,
+            'description': description,
+            'subtitles': video_subtitles,
             'thumbnail': info['thumbnail_url'],
             'age_limit': age_limit,
             'view_count': view_count,
+            'duration': info['duration']
         }
 
-    def _get_available_subtitles(self, video_id, webpage):
+    def _check_error(self, info):
+        if info.get('error') is not None:
+            raise ExtractorError(
+                '%s said: %s' % (self.IE_NAME, info['error']['title']), expected=True)
+
+    def _get_subtitles(self, video_id, webpage):
         try:
             sub_list = self._download_webpage(
                 'https://api.dailymotion.com/video/%s/subtitles?fields=id,language,url' % video_id,
                 video_id, note=False)
         except ExtractorError as err:
-            self._downloader.report_warning(u'unable to download video subtitles: %s' % compat_str(err))
+            self._downloader.report_warning('unable to download video subtitles: %s' % compat_str(err))
             return {}
         info = json.loads(sub_list)
         if (info['total'] > 0):
-            sub_lang_list = dict((l['language'], l['url']) for l in info['list'])
+            sub_lang_list = dict((l['language'], [{'url': l['url'], 'ext': 'srt'}]) for l in info['list'])
             return sub_lang_list
-        self._downloader.report_warning(u'video doesn\'t have subtitles')
+        self._downloader.report_warning('video doesn\'t have subtitles')
         return {}
 
 
 class DailymotionPlaylistIE(DailymotionBaseInfoExtractor):
-    IE_NAME = u'dailymotion:playlist'
+    IE_NAME = 'dailymotion:playlist'
     _VALID_URL = r'(?:https?://)?(?:www\.)?dailymotion\.[a-z]{2,3}/playlist/(?P<id>.+?)/'
-    _MORE_PAGES_INDICATOR = r'<div class="next">.*?<a.*?href="/playlist/.+?".*?>.*?</a>.*?</div>'
+    _MORE_PAGES_INDICATOR = r'(?s)<div class="pages[^"]*">.*?<a\s+class="[^"]*?icon-arrow_right[^"]*?"'
     _PAGE_TEMPLATE = 'https://www.dailymotion.com/playlist/%s/%s'
+    _TESTS = [{
+        'url': 'http://www.dailymotion.com/playlist/xv4bw_nqtv_sport/1#video=xl8v3q',
+        'info_dict': {
+            'title': 'SPORT',
+            'id': 'xv4bw_nqtv_sport',
+        },
+        'playlist_mincount': 20,
+    }]
 
     def _extract_entries(self, id):
-        video_ids = []
+        video_ids = set()
+        processed_urls = set()
         for pagenum in itertools.count(1):
-            request = self._build_request(self._PAGE_TEMPLATE % (id, pagenum))
-            webpage = self._download_webpage(request,
-                                             id, u'Downloading page %s' % pagenum)
+            page_url = self._PAGE_TEMPLATE % (id, pagenum)
+            webpage, urlh = self._download_webpage_handle_no_ff(
+                page_url, id, 'Downloading page %s' % pagenum)
+            if urlh.geturl() in processed_urls:
+                self.report_warning('Stopped at duplicated page %s, which is the same as %s' % (
+                    page_url, urlh.geturl()), id)
+                break
+
+            processed_urls.add(urlh.geturl())
 
-            playlist_el = get_element_by_attribute(u'class', u'row video_list', webpage)
-            video_ids.extend(re.findall(r'data-id="(.+?)"', playlist_el))
+            for video_id in re.findall(r'data-xid="(.+?)"', webpage):
+                if video_id not in video_ids:
+                    yield self.url_result('http://www.dailymotion.com/video/%s' % video_id, 'Dailymotion')
+                    video_ids.add(video_id)
 
-            if re.search(self._MORE_PAGES_INDICATOR, webpage, re.DOTALL) is None:
+            if re.search(self._MORE_PAGES_INDICATOR, webpage) is None:
                 break
-        return [self.url_result('http://www.dailymotion.com/video/%s' % video_id, 'Dailymotion')
-                   for video_id in orderedSet(video_ids)]
 
     def _real_extract(self, url):
         mobj = re.match(self._VALID_URL, url)
         playlist_id = mobj.group('id')
         webpage = self._download_webpage(url, playlist_id)
 
-        return {'_type': 'playlist',
-                'id': playlist_id,
-                'title': get_element_by_id(u'playlist_name', webpage),
-                'entries': self._extract_entries(playlist_id),
-                }
+        return {
+            '_type': 'playlist',
+            'id': playlist_id,
+            'title': self._og_search_title(webpage),
+            'entries': self._extract_entries(playlist_id),
+        }
 
 
 class DailymotionUserIE(DailymotionPlaylistIE):
-    IE_NAME = u'dailymotion:user'
-    _VALID_URL = r'(?:https?://)?(?:www\.)?dailymotion\.[a-z]{2,3}/user/(?P<user>[^/]+)'
-    _MORE_PAGES_INDICATOR = r'<div class="next">.*?<a.*?href="/user/.+?".*?>.*?</a>.*?</div>'
+    IE_NAME = 'dailymotion:user'
+    _VALID_URL = r'https?://(?:www\.)?dailymotion\.[a-z]{2,3}/(?!(?:embed|#|video|playlist)/)(?:(?:old/)?user/)?(?P<user>[^/]+)'
     _PAGE_TEMPLATE = 'http://www.dailymotion.com/user/%s/%s'
+    _TESTS = [{
+        'url': 'https://www.dailymotion.com/user/nqtv',
+        'info_dict': {
+            'id': 'nqtv',
+            'title': 'Rémi Gaillard',
+        },
+        'playlist_mincount': 100,
+    }, {
+        'url': 'http://www.dailymotion.com/user/UnderProject',
+        'info_dict': {
+            'id': 'UnderProject',
+            'title': 'UnderProject',
+        },
+        'playlist_mincount': 1800,
+        'expected_warnings': [
+            'Stopped at duplicated page',
+        ],
+        'skip': 'Takes too long time',
+    }]
 
     def _real_extract(self, url):
         mobj = re.match(self._VALID_URL, url)
         user = mobj.group('user')
-        webpage = self._download_webpage(url, user)
-        full_user = self._html_search_regex(
-            r'<a class="label" href="/%s".*?>(.*?)</' % re.escape(user),
-            webpage, u'user', flags=re.DOTALL)
+        webpage = self._download_webpage(
+            'https://www.dailymotion.com/user/%s' % user, user)
+        full_user = unescapeHTML(self._html_search_regex(
+            r'<a class="nav-image" title="([^"]+)" href="/%s">' % re.escape(user),
+            webpage, 'user'))
 
         return {
             '_type': 'playlist',
@@ -229,3 +369,52 @@ class DailymotionUserIE(DailymotionPlaylistIE):
             'title': full_user,
             'entries': self._extract_entries(user),
         }
+
+
+class DailymotionCloudIE(DailymotionBaseInfoExtractor):
+    _VALID_URL_PREFIX = r'http://api\.dmcloud\.net/(?:player/)?embed/'
+    _VALID_URL = r'%s[^/]+/(?P<id>[^/?]+)' % _VALID_URL_PREFIX
+    _VALID_EMBED_URL = r'%s[^/]+/[^\'"]+' % _VALID_URL_PREFIX
+
+    _TESTS = [{
+        # From http://www.francetvinfo.fr/economie/entreprises/les-entreprises-familiales-le-secret-de-la-reussite_933271.html
+        # Tested at FranceTvInfo_2
+        'url': 'http://api.dmcloud.net/embed/4e7343f894a6f677b10006b4/556e03339473995ee145930c?auth=1464865870-0-jyhsm84b-ead4c701fb750cf9367bf4447167a3db&autoplay=1',
+        'only_matching': True,
+    }, {
+        # http://www.francetvinfo.fr/societe/larguez-les-amarres-le-cobaturage-se-developpe_980101.html
+        'url': 'http://api.dmcloud.net/player/embed/4e7343f894a6f677b10006b4/559545469473996d31429f06?auth=1467430263-0-90tglw2l-a3a4b64ed41efe48d7fccad85b8b8fda&autoplay=1',
+        'only_matching': True,
+    }]
+
+    @classmethod
+    def _extract_dmcloud_url(self, webpage):
+        mobj = re.search(r'<iframe[^>]+src=[\'"](%s)[\'"]' % self._VALID_EMBED_URL, webpage)
+        if mobj:
+            return mobj.group(1)
+
+        mobj = re.search(
+            r'<input[^>]+id=[\'"]dmcloudUrlEmissionSelect[\'"][^>]+value=[\'"](%s)[\'"]' % self._VALID_EMBED_URL,
+            webpage)
+        if mobj:
+            return mobj.group(1)
+
+    def _real_extract(self, url):
+        video_id = self._match_id(url)
+
+        webpage = self._download_webpage_no_ff(url, video_id)
+
+        title = self._html_search_regex(r'<title>([^>]+)</title>', webpage, 'title')
+
+        video_info = self._parse_json(self._search_regex(
+            r'var\s+info\s*=\s*([^;]+);', webpage, 'video info'), video_id)
+
+        # TODO: parse ios_url, which is in fact a manifest
+        video_url = video_info['mp4_url']
+
+        return {
+            'id': video_id,
+            'url': video_url,
+            'title': title,
+            'thumbnail': video_info.get('thumbnail_url'),
+        }