release 2014.10.18

Also, improve detection (Addresses #3662)

youtube_dl/extractor/__init__.py

@@ -173,7 +173,6 @@ from .jadorecettepub import JadoreCettePubIE
 from .jeuxvideo import JeuxVideoIE
 from .jove import JoveIE
 from .jukebox import JukeboxIE
-from .justintv import JustinTVIE
 from .jpopsukitv import JpopsukiIE
 from .kankan import KankanIE
 from .keezmovies import KeezMoviesIE
@@ -316,6 +315,7 @@ from .sbs import SBSIE
 from .scivee import SciVeeIE
 from .screencast import ScreencastIE
 from .servingsys import ServingSysIE
+from .sexykarma import SexyKarmaIE
 from .shared import SharedIE
 from .sharesix import ShareSixIE
 from .sina import SinaIE
@@ -395,6 +395,7 @@ from .tutv import TutvIE
 from .tvigle import TvigleIE
 from .tvp import TvpIE
 from .tvplay import TVPlayIE
+from .twitch import TwitchIE
 from .ubu import UbuIE
 from .udemy import (
     UdemyIE,

youtube_dl/extractor/arte.py

@@ -160,6 +160,9 @@ class ArteTVPlus7IE(InfoExtractor):
                 format_id = '%s-%s' % (quality, format_info['versionCode'])
             else:
                 format_id = quality
+            media_type = format_info.get('mediaType')
+            if media_type is not None:
+                format_id += '-%s' % media_type
             info = {
                 'format_id': format_id,
                 'format_note': format_info.get('versionLibelle'),

youtube_dl/extractor/generic.py

@@ -609,13 +609,13 @@ class GenericIE(InfoExtractor):
         if mobj:
             player_url = unescapeHTML(mobj.group('url'))
             surl = smuggle_url(player_url, {'Referer': url})
-            return self.url_result(surl, 'Vimeo')
+            return self.url_result(surl)
 
         # Look for embedded (swf embed) Vimeo player
         mobj = re.search(
-            r'<embed[^>]+?src="(https?://(?:www\.)?vimeo\.com/moogaloop\.swf.+?)"', webpage)
+            r'<embed[^>]+?src="((?:https?:)?//(?:www\.)?vimeo\.com/moogaloop\.swf.+?)"', webpage)
         if mobj:
-            return self.url_result(mobj.group(1), 'Vimeo')
+            return self.url_result(mobj.group(1))
 
         # Look for embedded YouTube player
         matches = re.findall(r'''(?x)
@@ -654,15 +654,17 @@ class GenericIE(InfoExtractor):
         match = re.search(
             r'<iframe[^>]+?src=(["\'])(?P<url>(?:https?:)?//(?:fast\.)?wistia\.net/embed/iframe/.+?)\1', webpage)
         if match:
+            embed_url = self._proto_relative_url(
+                unescapeHTML(match.group('url')))
             return {
                 '_type': 'url_transparent',
-                'url': unescapeHTML(match.group('url')),
+                'url': embed_url,
                 'ie_key': 'Wistia',
                 'uploader': video_uploader,
                 'title': video_title,
                 'id': video_id,
             }
-        match = re.search(r'(?:id=["\']wistia_|data-wistiaid=["\']|Wistia\.embed\(["\'])(?P<id>[^"\']+)', webpage)
+        match = re.search(r'(?:id=["\']wistia_|data-wistia-?id=["\']|Wistia\.embed\(["\'])(?P<id>[^"\']+)', webpage)
         if match:
             return {
                 '_type': 'url_transparent',

youtube_dl/extractor/justintv.py

@@ -1,155 +0,0 @@
-from __future__ import unicode_literals
-
-import itertools
-import json
-import os
-import re
-
-from .common import InfoExtractor
-from ..utils import (
-    compat_str,
-    ExtractorError,
-    formatSeconds,
-)
-
-
-class JustinTVIE(InfoExtractor):
-    """Information extractor for justin.tv and twitch.tv"""
-    # TODO: One broadcast may be split into multiple videos. The key
-    # 'broadcast_id' is the same for all parts, and 'broadcast_part'
-    # starts at 1 and increases. Can we treat all parts as one video?
-
-    _VALID_URL = r"""(?x)^(?:http://)?(?:www\.)?(?:twitch|justin)\.tv/
-        (?:
-            (?P<channelid>[^/]+)|
-            (?:(?:[^/]+)/b/(?P<videoid>[^/]+))|
-            (?:(?:[^/]+)/c/(?P<chapterid>[^/]+))
-        )
-        /?(?:\#.*)?$
-        """
-    _JUSTIN_PAGE_LIMIT = 100
-    IE_NAME = 'justin.tv'
-    IE_DESC = 'justin.tv and twitch.tv'
-    _TEST = {
-        'url': 'http://www.twitch.tv/thegamedevhub/b/296128360',
-        'md5': 'ecaa8a790c22a40770901460af191c9a',
-        'info_dict': {
-            'id': '296128360',
-            'ext': 'flv',
-            'upload_date': '20110927',
-            'uploader_id': 25114803,
-            'uploader': 'thegamedevhub',
-            'title': 'Beginner Series - Scripting With Python Pt.1'
-        }
-    }
-
-    # Return count of items, list of *valid* items
-    def _parse_page(self, url, video_id, counter):
-        info_json = self._download_webpage(
-            url, video_id,
-            'Downloading video info JSON on page %d' % counter,
-            'Unable to download video info JSON %d' % counter)
-
-        response = json.loads(info_json)
-        if type(response) != list:
-            error_text = response.get('error', 'unknown error')
-            raise ExtractorError('Justin.tv API: %s' % error_text)
-        info = []
-        for clip in response:
-            video_url = clip['video_file_url']
-            if video_url:
-                video_extension = os.path.splitext(video_url)[1][1:]
-                video_date = re.sub('-', '', clip['start_time'][:10])
-                video_uploader_id = clip.get('user_id', clip.get('channel_id'))
-                video_id = clip['id']
-                video_title = clip.get('title', video_id)
-                info.append({
-                    'id': compat_str(video_id),
-                    'url': video_url,
-                    'title': video_title,
-                    'uploader': clip.get('channel_name', video_uploader_id),
-                    'uploader_id': video_uploader_id,
-                    'upload_date': video_date,
-                    'ext': video_extension,
-                })
-        return (len(response), info)
-
-    def _real_extract(self, url):
-        mobj = re.match(self._VALID_URL, url)
-
-        api_base = 'http://api.justin.tv'
-        paged = False
-        if mobj.group('channelid'):
-            paged = True
-            video_id = mobj.group('channelid')
-            api = api_base + '/channel/archives/%s.json' % video_id
-        elif mobj.group('chapterid'):
-            chapter_id = mobj.group('chapterid')
-
-            webpage = self._download_webpage(url, chapter_id)
-            m = re.search(r'PP\.archive_id = "([0-9]+)";', webpage)
-            if not m:
-                raise ExtractorError('Cannot find archive of a chapter')
-            archive_id = m.group(1)
-
-            api = api_base + '/broadcast/by_chapter/%s.xml' % chapter_id
-            doc = self._download_xml(
-                api, chapter_id,
-                note='Downloading chapter information',
-                errnote='Chapter information download failed')
-            for a in doc.findall('.//archive'):
-                if archive_id == a.find('./id').text:
-                    break
-            else:
-                raise ExtractorError('Could not find chapter in chapter information')
-
-            video_url = a.find('./video_file_url').text
-            video_ext = video_url.rpartition('.')[2] or 'flv'
-
-            chapter_api_url = 'https://api.twitch.tv/kraken/videos/c' + chapter_id
-            chapter_info = self._download_json(
-                chapter_api_url, 'c' + chapter_id,
-                note='Downloading chapter metadata',
-                errnote='Download of chapter metadata failed')
-
-            bracket_start = int(doc.find('.//bracket_start').text)
-            bracket_end = int(doc.find('.//bracket_end').text)
-
-            # TODO determine start (and probably fix up file)
-            #  youtube-dl -v http://www.twitch.tv/firmbelief/c/1757457
-            #video_url += '?start=' + TODO:start_timestamp
-            # bracket_start is 13290, but we want 51670615
-            self._downloader.report_warning('Chapter detected, but we can just download the whole file. '
-                                            'Chapter starts at %s and ends at %s' % (formatSeconds(bracket_start), formatSeconds(bracket_end)))
-
-            info = {
-                'id': 'c' + chapter_id,
-                'url': video_url,
-                'ext': video_ext,
-                'title': chapter_info['title'],
-                'thumbnail': chapter_info['preview'],
-                'description': chapter_info['description'],
-                'uploader': chapter_info['channel']['display_name'],
-                'uploader_id': chapter_info['channel']['name'],
-            }
-            return info
-        else:
-            video_id = mobj.group('videoid')
-            api = api_base + '/broadcast/by_archive/%s.json' % video_id
-
-        entries = []
-        offset = 0
-        limit = self._JUSTIN_PAGE_LIMIT
-        for counter in itertools.count(1):
-            page_url = api + ('?offset=%d&limit=%d' % (offset, limit))
-            page_count, page_info = self._parse_page(
-                page_url, video_id, counter)
-            entries.extend(page_info)
-            if not paged or page_count != limit:
-                break
-            offset += limit
-        return {
-            '_type': 'playlist',
-            'id': video_id,
-            'entries': entries,
-        }

youtube_dl/extractor/sexykarma.py

@@ -0,0 +1,106 @@
+# coding: utf-8
+from __future__ import unicode_literals
+
+import re
+
+from .common import InfoExtractor
+from ..utils import (
+    unified_strdate,
+    parse_duration,
+    int_or_none,
+)
+
+
+class SexyKarmaIE(InfoExtractor):
+    _VALID_URL = r'https?://(?:www\.)?sexykarma\.com/gonewild/video/(?P<display_id>[^/]+)-(?P<id>[a-zA-Z0-9]+)\.html'
+    _TESTS = [{
+        'url': 'http://www.sexykarma.com/gonewild/video/taking-a-quick-pee-yHI70cOyIHt.html',
+        'md5': 'b9798e7d1ef1765116a8f516c8091dbd',
+        'info_dict': {
+            'id': 'yHI70cOyIHt',
+            'display_id': 'taking-a-quick-pee',
+            'ext': 'mp4',
+            'title': 'Taking a quick pee.',
+            'description': '',
+            'thumbnail': 're:^https?://.*\.jpg$',
+            'uploader': 'wildginger7',
+            'upload_date': '20141007',
+            'duration': 81,
+            'view_count': int,
+            'comment_count': int,
+            'categories': list,
+        }
+    }, {
+        'url': 'http://www.sexykarma.com/gonewild/video/pot-pixie-tribute-8Id6EZPbuHf.html',
+        'md5': 'dd216c68d29b49b12842b9babe762a5d',
+        'info_dict': {
+            'id': '8Id6EZPbuHf',
+            'display_id': 'pot-pixie-tribute',
+            'ext': 'mp4',
+            'title': 'pot_pixie tribute',
+            'description': 'tribute',
+            'thumbnail': 're:^https?://.*\.jpg$',
+            'uploader': 'banffite',
+            'upload_date': '20141013',
+            'duration': 16,
+            'view_count': int,
+            'comment_count': int,
+            'categories': list,
+        }
+    }]
+
+    def _real_extract(self, url):
+        mobj = re.match(self._VALID_URL, url)
+        video_id = mobj.group('id')
+        display_id = mobj.group('display_id')
+
+        webpage = self._download_webpage(url, display_id)
+
+        video_url = self._html_search_regex(
+            r'<p>Save this video to your computer: </p><p><a href="([^"]+)"',
+            webpage, 'url')
+
+        title = self._html_search_regex(
+            r'<h2 class="he2"><span>(.*?)</span>',
+            webpage, 'title')
+        description = self._html_search_meta(
+            'description', webpage, 'description', fatal=False, default='')
+        thumbnail = self._html_search_regex(
+            r'<span id="container"><img\s+src="([^"]+)"',
+            webpage, 'thumbnail', fatal=False)
+
+        uploader = self._html_search_regex(
+            r'class="aupa">\s*(.*?)</a>',
+            webpage, 'uploader')
+        upload_date = unified_strdate(self._html_search_regex(
+            r'Added: <strong>(.+?)</strong>', webpage, 'upload date', fatal=False))
+
+        duration = parse_duration(self._search_regex(
+            r'<td>Time:\s*</td>\s*<td align="right"><span>\s*(.+?)\s*</span>',
+            webpage, 'duration', fatal=False))
+
+        view_count = int_or_none(self._search_regex(
+            r'<td>Views:\s*</td>\s*<td align="right"><span>\s*(\d+)\s*</span>',
+            webpage, 'view count', fatal=False))
+        comment_count = int_or_none(self._search_regex(
+            r'<td>Comments:\s*</td>\s*<td align="right"><span>\s*(\d+)\s*</span>',
+            webpage, 'comment count', fatal=False))
+
+        categories = self._html_search_meta(
+            'keywords', webpage, 'categories',
+            fatal=False, default='').split(',')
+
+        return {
+            'id': video_id,
+            'display_id': display_id,
+            'url': video_url,
+            'title': title,
+            'description': description,
+            'thumbnail': thumbnail,
+            'uploader': uploader,
+            'upload_date': upload_date,
+            'duration': duration,
+            'view_count': view_count,
+            'comment_count': comment_count,
+            'categories': categories,
+        }

youtube_dl/extractor/twitch.py

@@ -0,0 +1,187 @@
+from __future__ import unicode_literals
+
+import itertools
+import re
+
+from .common import InfoExtractor
+from ..utils import (
+    ExtractorError,
+    parse_iso8601,
+)
+
+
+class TwitchIE(InfoExtractor):
+    # TODO: One broadcast may be split into multiple videos. The key
+    # 'broadcast_id' is the same for all parts, and 'broadcast_part'
+    # starts at 1 and increases. Can we treat all parts as one video?
+    _VALID_URL = r"""(?x)^(?:http://)?(?:www\.)?twitch\.tv/
+        (?:
+            (?P<channelid>[^/]+)|
+            (?:(?:[^/]+)/b/(?P<videoid>[^/]+))|
+            (?:(?:[^/]+)/c/(?P<chapterid>[^/]+))
+        )
+        /?(?:\#.*)?$
+        """
+    _PAGE_LIMIT = 100
+    _API_BASE = 'https://api.twitch.tv'
+    _TESTS = [{
+        'url': 'http://www.twitch.tv/riotgames/b/577357806',
+        'info_dict': {
+            'id': 'a577357806',
+            'title': 'Worlds Semifinals - Star Horn Royal Club vs. OMG',
+        },
+        'playlist_mincount': 12,
+    }, {
+        'url': 'http://www.twitch.tv/acracingleague/c/5285812',
+        'info_dict': {
+            'id': 'c5285812',
+            'title': 'ACRL Off Season - Sports Cars @ Nordschleife',
+        },
+        'playlist_mincount': 3,
+    }, {
+        'url': 'http://www.twitch.tv/vanillatv',
+        'info_dict': {
+            'id': 'vanillatv',
+            'title': 'VanillaTV',
+        },
+        'playlist_mincount': 412,
+    }]
+
+    def _handle_error(self, response):
+        if not isinstance(response, dict):
+            return
+        error = response.get('error')
+        if error:
+            raise ExtractorError(
+                '%s returned error: %s - %s' % (self.IE_NAME, error, response.get('message')),
+                expected=True)
+
+    def _download_json(self, url, video_id, note='Downloading JSON metadata'):
+        response = super(TwitchIE, self)._download_json(url, video_id, note)
+        self._handle_error(response)
+        return response
+
+    def _extract_media(self, item, item_id):
+        ITEMS = {
+            'a': 'video',
+            'c': 'chapter',
+        }
+        info = self._extract_info(self._download_json(
+            '%s/kraken/videos/%s%s' % (self._API_BASE, item, item_id), item_id,
+            'Downloading %s info JSON' % ITEMS[item]))
+        response = self._download_json(
+            '%s/api/videos/%s%s' % (self._API_BASE, item, item_id), item_id,
+            'Downloading %s playlist JSON' % ITEMS[item])
+        entries = []
+        chunks = response['chunks']
+        qualities = list(chunks.keys())
+        for num, fragment in enumerate(zip(*chunks.values()), start=1):
+            formats = []
+            for fmt_num, fragment_fmt in enumerate(fragment):
+                format_id = qualities[fmt_num]
+                fmt = {
+                    'url': fragment_fmt['url'],
+                    'format_id': format_id,
+                    'quality': 1 if format_id == 'live' else 0,
+                }
+                m = re.search(r'^(?P<height>\d+)[Pp]', format_id)
+                if m:
+                    fmt['height'] = int(m.group('height'))
+                formats.append(fmt)
+            self._sort_formats(formats)
+            entry = dict(info)
+            entry['id'] = '%s_%d' % (entry['id'], num)
+            entry['title'] = '%s part %d' % (entry['title'], num)
+            entry['formats'] = formats
+            entries.append(entry)
+        return self.playlist_result(entries, info['id'], info['title'])
+
+    def _extract_info(self, info):
+        return {
+            'id': info['_id'],
+            'title': info['title'],
+            'description': info['description'],
+            'duration': info['length'],
+            'thumbnail': info['preview'],
+            'uploader': info['channel']['display_name'],
+            'uploader_id': info['channel']['name'],
+            'timestamp': parse_iso8601(info['recorded_at']),
+            'view_count': info['views'],
+        }
+
+    def _real_extract(self, url):
+        mobj = re.match(self._VALID_URL, url)
+        if mobj.group('chapterid'):
+            return self._extract_media('c', mobj.group('chapterid'))
+
+            """
+            webpage = self._download_webpage(url, chapter_id)
+            m = re.search(r'PP\.archive_id = "([0-9]+)";', webpage)
+            if not m:
+                raise ExtractorError('Cannot find archive of a chapter')
+            archive_id = m.group(1)
+
+            api = api_base + '/broadcast/by_chapter/%s.xml' % chapter_id
+            doc = self._download_xml(
+                api, chapter_id,
+                note='Downloading chapter information',
+                errnote='Chapter information download failed')
+            for a in doc.findall('.//archive'):
+                if archive_id == a.find('./id').text:
+                    break
+            else:
+                raise ExtractorError('Could not find chapter in chapter information')
+
+            video_url = a.find('./video_file_url').text
+            video_ext = video_url.rpartition('.')[2] or 'flv'
+
+            chapter_api_url = 'https://api.twitch.tv/kraken/videos/c' + chapter_id
+            chapter_info = self._download_json(
+                chapter_api_url, 'c' + chapter_id,
+                note='Downloading chapter metadata',
+                errnote='Download of chapter metadata failed')
+
+            bracket_start = int(doc.find('.//bracket_start').text)
+            bracket_end = int(doc.find('.//bracket_end').text)
+
+            # TODO determine start (and probably fix up file)
+            #  youtube-dl -v http://www.twitch.tv/firmbelief/c/1757457
+            #video_url += '?start=' + TODO:start_timestamp
+            # bracket_start is 13290, but we want 51670615
+            self._downloader.report_warning('Chapter detected, but we can just download the whole file. '
+                                            'Chapter starts at %s and ends at %s' % (formatSeconds(bracket_start), formatSeconds(bracket_end)))
+
+            info = {
+                'id': 'c' + chapter_id,
+                'url': video_url,
+                'ext': video_ext,
+                'title': chapter_info['title'],
+                'thumbnail': chapter_info['preview'],
+                'description': chapter_info['description'],
+                'uploader': chapter_info['channel']['display_name'],
+                'uploader_id': chapter_info['channel']['name'],
+            }
+            return info
+            """
+        elif mobj.group('videoid'):
+            return self._extract_media('a', mobj.group('videoid'))
+        elif mobj.group('channelid'):
+            channel_id = mobj.group('channelid')
+            info = self._download_json(
+                '%s/kraken/channels/%s' % (self._API_BASE, channel_id),
+                channel_id, 'Downloading channel info JSON')
+            channel_name = info.get('display_name') or info.get('name')
+            entries = []
+            offset = 0
+            limit = self._PAGE_LIMIT
+            for counter in itertools.count(1):
+                response = self._download_json(
+                    '%s/kraken/channels/%s/videos/?offset=%d&limit=%d'
+                    % (self._API_BASE, channel_id, offset, limit),
+                    channel_id, 'Downloading channel videos JSON page %d' % counter)
+                videos = response['videos']
+                if not videos:
+                    break
+                entries.extend([self.url_result(video['url'], 'Twitch') for video in videos])
+                offset += limit
+            return self.playlist_result(entries, channel_id, channel_name)

youtube_dl/extractor/vimeo.py

@@ -157,6 +157,18 @@ class VimeoIE(VimeoBaseInfoExtractor, SubtitlesInfoExtractor):
                 'duration': 62,
             }
         },
+        {
+            # from https://www.ouya.tv/game/Pier-Solar-and-the-Great-Architects/
+            'url': 'https://player.vimeo.com/video/98044508',
+            'note': 'The js code contains assignments to the same variable as the config',
+            'info_dict': {
+                'id': '98044508',
+                'ext': 'mp4',
+                'title': 'Pier Solar OUYA Official Trailer',
+                'uploader': 'Tulio Gonçalves',
+                'uploader_id': 'user28849593',
+            },
+        },
     ]
 
     def _verify_video_password(self, url, video_id, webpage):
@@ -244,7 +256,7 @@ class VimeoIE(VimeoBaseInfoExtractor, SubtitlesInfoExtractor):
                 # We try to find out to which variable is assigned the config dic
                 m_variable_name = re.search('(\w)\.video\.id', webpage)
                 if m_variable_name is not None:
-                    config_re = r'%s=({.+?});' % re.escape(m_variable_name.group(1))
+                    config_re = r'%s=({[^}].+?});' % re.escape(m_variable_name.group(1))
                 else:
                     config_re = [r' = {config:({.+?}),assets:', r'(?:[abc])=({.+?});']
                 config = self._search_regex(config_re, webpage, 'info section',

youtube_dl/version.py

@@ -1,2 +1,2 @@
 
-__version__ = '2014.10.15'
+__version__ = '2014.10.18'