debian/changelog: Annotate bugs being closed.

[youtubedl] / youtube_dl / extractor / arte.py

diff --git a/youtube_dl/extractor/arte.py b/youtube_dl/extractor/arte.py
index 18d5916589b239c9cf12a7caa629efeea784ce86..b9a9440c09b85365a2997bd5feddbae017601c2d 100644 (file)
--- a/youtube_dl/extractor/arte.py
+++ b/youtube_dl/extractor/arte.py
@@ -1,146 +1,278 @@
+# encoding: utf-8
+from __future__ import unicode_literals
+

 import re
 import re

-import json
-import xml.etree.ElementTree

 
 from .common import InfoExtractor
 from ..utils import (
     ExtractorError,
     find_xpath_attr,
     unified_strdate,
 
 from .common import InfoExtractor
 from ..utils import (
     ExtractorError,
     find_xpath_attr,
     unified_strdate,

+    determine_ext,
+    get_element_by_id,
+    get_element_by_attribute,
+    int_or_none,

 )
 
 )
 

+# There are different sources of video in arte.tv, the extraction process 
+# is different for each one. The videos usually expire in 7 days, so we can't
+# add tests.
+
+

 class ArteTvIE(InfoExtractor):
 class ArteTvIE(InfoExtractor):

-    """
-    There are two sources of video in arte.tv: videos.arte.tv and
-    www.arte.tv/guide, the extraction process is different for each one.
-    The videos expire in 7 days, so we can't add tests.
-    """
-    _EMISSION_URL = r'(?:http://)?www\.arte.tv/guide/(?P<lang>fr|de)/(?:(?:sendungen|emissions)/)?(?P<id>.*?)/(?P<name>.*?)(\?.*)?'
-    _VIDEOS_URL = r'(?:http://)?videos.arte.tv/(?P<lang>fr|de)/.*-(?P<id>.*?).html'
-    _LIVE_URL = r'index-[0-9]+\.html$'
+    _VALID_URL = r'http://videos\.arte\.tv/(?P<lang>fr|de)/.*-(?P<id>.*?)\.html'
+    IE_NAME = 'arte.tv'
+
+    def _real_extract(self, url):
+        mobj = re.match(self._VALID_URL, url)
+        lang = mobj.group('lang')
+        video_id = mobj.group('id')
+
+        ref_xml_url = url.replace('/videos/', '/do_delegate/videos/')
+        ref_xml_url = ref_xml_url.replace('.html', ',view,asPlayerXml.xml')
+        ref_xml_doc = self._download_xml(
+            ref_xml_url, video_id, note='Downloading metadata')
+        config_node = find_xpath_attr(ref_xml_doc, './/video', 'lang', lang)
+        config_xml_url = config_node.attrib['ref']
+        config = self._download_xml(
+            config_xml_url, video_id, note='Downloading configuration')
+
+        formats = [{
+            'forma_id': q.attrib['quality'],
+            # The playpath starts at 'mp4:', if we don't manually
+            # split the url, rtmpdump will incorrectly parse them
+            'url': q.text.split('mp4:', 1)[0],
+            'play_path': 'mp4:' + q.text.split('mp4:', 1)[1],
+            'ext': 'flv',
+            'quality': 2 if q.attrib['quality'] == 'hd' else 1,
+        } for q in config.findall('./urls/url')]
+        self._sort_formats(formats)

 
 

-    IE_NAME = u'arte.tv'
+        title = config.find('.//name').text
+        thumbnail = config.find('.//firstThumbnailUrl').text
+        return {
+            'id': video_id,
+            'title': title,
+            'thumbnail': thumbnail,
+            'formats': formats,
+        }
+
+
+class ArteTVPlus7IE(InfoExtractor):
+    IE_NAME = 'arte.tv:+7'
+    _VALID_URL = r'https?://(?:www\.)?arte\.tv/guide/(?P<lang>fr|de)/(?:(?:sendungen|emissions)/)?(?P<id>.*?)/(?P<name>.*?)(\?.*)?'

 
     @classmethod
 
     @classmethod

-    def suitable(cls, url):
-        return any(re.match(regex, url) for regex in (cls._EMISSION_URL, cls._VIDEOS_URL))
-
-    # TODO implement Live Stream
-    # from ..utils import compat_urllib_parse
-    # def extractLiveStream(self, url):
-    #     video_lang = url.split('/')[-4]
-    #     info = self.grep_webpage(
-    #         url,
-    #         r'src="(.*?/videothek_js.*?\.js)',
-    #         0,
-    #         [
-    #             (1, 'url', u'Invalid URL: %s' % url)
-    #         ]
-    #     )
-    #     http_host = url.split('/')[2]
-    #     next_url = 'http://%s%s' % (http_host, compat_urllib_parse.unquote(info.get('url')))
-    #     info = self.grep_webpage(
-    #         next_url,
-    #         r'(s_artestras_scst_geoFRDE_' + video_lang + '.*?)\'.*?' +
-    #             '(http://.*?\.swf).*?' +
-    #             '(rtmp://.*?)\'',
-    #         re.DOTALL,
-    #         [
-    #             (1, 'path',   u'could not extract video path: %s' % url),
-    #             (2, 'player', u'could not extract video player: %s' % url),
-    #             (3, 'url',    u'could not extract video url: %s' % url)
-    #         ]
-    #     )
-    #     video_url = u'%s/%s' % (info.get('url'), info.get('path'))
+    def _extract_url_info(cls, url):
+        mobj = re.match(cls._VALID_URL, url)
+        lang = mobj.group('lang')
+        # This is not a real id, it can be for example AJT for the news
+        # http://www.arte.tv/guide/fr/emissions/AJT/arte-journal
+        video_id = mobj.group('id')
+        return video_id, lang

 
     def _real_extract(self, url):
 
     def _real_extract(self, url):

-        mobj = re.match(self._EMISSION_URL, url)
-        if mobj is not None:
-            lang = mobj.group('lang')
-            # This is not a real id, it can be for example AJT for the news
-            # http://www.arte.tv/guide/fr/emissions/AJT/arte-journal
-            video_id = mobj.group('id')
-            return self._extract_emission(url, video_id, lang)
-
-        mobj = re.match(self._VIDEOS_URL, url)
-        if mobj is not None:
-            id = mobj.group('id')
-            lang = mobj.group('lang')
-            return self._extract_video(url, id, lang)
-
-        if re.search(self._LIVE_URL, video_id) is not None:
-            raise ExtractorError(u'Arte live streams are not yet supported, sorry')
-            # self.extractLiveStream(url)
-            # return
-
-    def _extract_emission(self, url, video_id, lang):
-        """Extract from www.arte.tv/guide"""
+        video_id, lang = self._extract_url_info(url)

         webpage = self._download_webpage(url, video_id)
         webpage = self._download_webpage(url, video_id)

-        json_url = self._html_search_regex(r'arte_vp_url="(.*?)"', webpage, 'json url')
+        return self._extract_from_webpage(webpage, video_id, lang)
+
+    def _extract_from_webpage(self, webpage, video_id, lang):
+        json_url = self._html_search_regex(
+            [r'arte_vp_url=["\'](.*?)["\']', r'data-url=["\']([^"]+)["\']'],
+            webpage, 'json vp url')
+        return self._extract_from_json_url(json_url, video_id, lang)

 
 

-        json_info = self._download_webpage(json_url, video_id, 'Downloading info json')
-        self.report_extraction(video_id)
-        info = json.loads(json_info)
+    def _extract_from_json_url(self, json_url, video_id, lang):
+        info = self._download_json(json_url, video_id)

         player_info = info['videoJsonPlayer']
 
         player_info = info['videoJsonPlayer']
 

-        info_dict = {'id': player_info['VID'],
-                     'title': player_info['VTI'],
-                     'description': player_info['VDE'],
-                     'upload_date': unified_strdate(player_info['VDA'].split(' ')[0]),
-                     'thumbnail': player_info['programImage'],
-                     'ext': 'flv',
-                     }
+        upload_date_str = player_info.get('shootingDate')
+        if not upload_date_str:
+            upload_date_str = player_info.get('VDA', '').split(' ')[0]
+
+        title = player_info['VTI'].strip()
+        subtitle = player_info.get('VSU', '').strip()
+        if subtitle:
+            title += ' - %s' % subtitle
+
+        info_dict = {
+            'id': player_info['VID'],
+            'title': title,
+            'description': player_info.get('VDE'),
+            'upload_date': unified_strdate(upload_date_str),
+            'thumbnail': player_info.get('programImage') or player_info.get('VTU', {}).get('IUR'),
+        }

 
 

-        formats = player_info['VSR'].values()
+        all_formats = []
+        for format_id, format_dict in player_info['VSR'].items():
+            fmt = dict(format_dict)
+            fmt['format_id'] = format_id
+            all_formats.append(fmt)
+        # Some formats use the m3u8 protocol
+        all_formats = list(filter(lambda f: f.get('videoFormat') != 'M3U8', all_formats))

         def _match_lang(f):
         def _match_lang(f):

+            if f.get('versionCode') is None:
+                return True

             # Return true if that format is in the language of the url
             if lang == 'fr':
                 l = 'F'
             elif lang == 'de':
                 l = 'A'
             # Return true if that format is in the language of the url
             if lang == 'fr':
                 l = 'F'
             elif lang == 'de':
                 l = 'A'

+            else:
+                l = lang

             regexes = [r'VO?%s' % l, r'VO?.-ST%s' % l]
             return any(re.match(r, f['versionCode']) for r in regexes)
         # Some formats may not be in the same language as the url
             regexes = [r'VO?%s' % l, r'VO?.-ST%s' % l]
             return any(re.match(r, f['versionCode']) for r in regexes)
         # Some formats may not be in the same language as the url

-        formats = filter(_match_lang, formats)
-        # We order the formats by quality
-        formats = sorted(formats, key=lambda f: int(f['height']))
-        # Pick the best quality
-        format_info = formats[-1]
-        if format_info['mediaType'] == u'rtmp':
-            info_dict['url'] = format_info['streamer']
-            info_dict['play_path'] = 'mp4:' + format_info['url']
+        # TODO: Might want not to drop videos that does not match requested language
+        # but to process those formats with lower precedence
+        formats = filter(_match_lang, all_formats)
+        formats = list(formats)  # in python3 filter returns an iterator
+        if not formats:
+            # Some videos are only available in the 'Originalversion'
+            # they aren't tagged as being in French or German
+            # Sometimes there are neither videos of requested lang code
+            # nor original version videos available
+            # For such cases we just take all_formats as is
+            formats = all_formats
+            if not formats:
+                raise ExtractorError('The formats list is empty')
+
+        if re.match(r'[A-Z]Q', formats[0]['quality']) is not None:
+            def sort_key(f):
+                return ['HQ', 'MQ', 'EQ', 'SQ'].index(f['quality'])

         else:
         else:

-            info_dict['url'] = format_info['url']
+            def sort_key(f):
+                versionCode = f.get('versionCode')
+                if versionCode is None:
+                    versionCode = ''
+                return (
+                    # Sort first by quality
+                    int(f.get('height', -1)),
+                    int(f.get('bitrate', -1)),
+                    # The original version with subtitles has lower relevance
+                    re.match(r'VO-ST(F|A)', versionCode) is None,
+                    # The version with sourds/mal subtitles has also lower relevance
+                    re.match(r'VO?(F|A)-STM\1', versionCode) is None,
+                    # Prefer http downloads over m3u8
+                    0 if f['url'].endswith('m3u8') else 1,
+                )
+        formats = sorted(formats, key=sort_key)
+        def _format(format_info):
+            info = {
+                'format_id': format_info['format_id'],
+                'format_note': '%s, %s' % (format_info.get('versionCode'), format_info.get('versionLibelle')),
+                'width': int_or_none(format_info.get('width')),
+                'height': int_or_none(format_info.get('height')),
+                'tbr': int_or_none(format_info.get('bitrate')),
+            }
+            if format_info['mediaType'] == 'rtmp':
+                info['url'] = format_info['streamer']
+                info['play_path'] = 'mp4:' + format_info['url']
+                info['ext'] = 'flv'
+            else:
+                info['url'] = format_info['url']
+                info['ext'] = determine_ext(info['url'])
+            return info
+        info_dict['formats'] = [_format(f) for f in formats]

 
         return info_dict
 
 
         return info_dict
 

-    def _extract_video(self, url, video_id, lang):
-        """Extract from videos.arte.tv"""
-        ref_xml_url = url.replace('/videos/', '/do_delegate/videos/')
-        ref_xml_url = ref_xml_url.replace('.html', ',view,asPlayerXml.xml')
-        ref_xml = self._download_webpage(ref_xml_url, video_id, note=u'Downloading metadata')
-        ref_xml_doc = xml.etree.ElementTree.fromstring(ref_xml)
-        config_node = find_xpath_attr(ref_xml_doc, './/video', 'lang', lang)
-        config_xml_url = config_node.attrib['ref']
-        config_xml = self._download_webpage(config_xml_url, video_id, note=u'Downloading configuration')

 
 

-        video_urls = list(re.finditer(r'<url quality="(?P<quality>.*?)">(?P<url>.*?)</url>', config_xml))
-        def _key(m):
-            quality = m.group('quality')
-            if quality == 'hd':
-                return 2
-            else:
-                return 1
-        # We pick the best quality
-        video_urls = sorted(video_urls, key=_key)
-        video_url = list(video_urls)[-1].group('url')
-        
-        title = self._html_search_regex(r'<name>(.*?)</name>', config_xml, 'title')
-        thumbnail = self._html_search_regex(r'<firstThumbnailUrl>(.*?)</firstThumbnailUrl>',
-                                            config_xml, 'thumbnail')
-        return {'id': video_id,
-                'title': title,
-                'thumbnail': thumbnail,
-                'url': video_url,
-                'ext': 'flv',
-                }
+# It also uses the arte_vp_url url from the webpage to extract the information
+class ArteTVCreativeIE(ArteTVPlus7IE):
+    IE_NAME = 'arte.tv:creative'
+    _VALID_URL = r'https?://creative\.arte\.tv/(?P<lang>fr|de)/(?:magazine?/)?(?P<id>[^?#]+)'
+
+    _TESTS = [{
+        'url': 'http://creative.arte.tv/de/magazin/agentur-amateur-corporate-design',
+        'info_dict': {
+            'id': '72176',
+            'ext': 'mp4',
+            'title': 'Folge 2 - Corporate Design',
+            'upload_date': '20131004',
+        },
+    }, {
+        'url': 'http://creative.arte.tv/fr/Monty-Python-Reunion',
+        'info_dict': {
+            'id': '160676',
+            'ext': 'mp4',
+            'title': 'Monty Python live (mostly)',
+            'description': 'Événement ! Quarante-cinq ans après leurs premiers succès, les légendaires Monty Python remontent sur scène.\n',
+            'upload_date': '20140805',
+        }
+    }]
+
+
+class ArteTVFutureIE(ArteTVPlus7IE):
+    IE_NAME = 'arte.tv:future'
+    _VALID_URL = r'https?://future\.arte\.tv/(?P<lang>fr|de)/(thema|sujet)/.*?#article-anchor-(?P<id>\d+)'
+
+    _TEST = {
+        'url': 'http://future.arte.tv/fr/sujet/info-sciences#article-anchor-7081',
+        'info_dict': {
+            'id': '5201',
+            'ext': 'mp4',
+            'title': 'Les champignons au secours de la planète',
+            'upload_date': '20131101',
+        },
+    }
+
+    def _real_extract(self, url):
+        anchor_id, lang = self._extract_url_info(url)
+        webpage = self._download_webpage(url, anchor_id)
+        row = get_element_by_id(anchor_id, webpage)
+        return self._extract_from_webpage(row, anchor_id, lang)
+
+
+class ArteTVDDCIE(ArteTVPlus7IE):
+    IE_NAME = 'arte.tv:ddc'
+    _VALID_URL = r'https?://ddc\.arte\.tv/(?P<lang>emission|folge)/(?P<id>.+)'
+
+    def _real_extract(self, url):
+        video_id, lang = self._extract_url_info(url)
+        if lang == 'folge':
+            lang = 'de'
+        elif lang == 'emission':
+            lang = 'fr'
+        webpage = self._download_webpage(url, video_id)
+        scriptElement = get_element_by_attribute('class', 'visu_video_block', webpage)
+        script_url = self._html_search_regex(r'src="(.*?)"', scriptElement, 'script url')
+        javascriptPlayerGenerator = self._download_webpage(script_url, video_id, 'Download javascript player generator')
+        json_url = self._search_regex(r"json_url=(.*)&rendering_place.*", javascriptPlayerGenerator, 'json url')
+        return self._extract_from_json_url(json_url, video_id, lang)
+
+
+class ArteTVConcertIE(ArteTVPlus7IE):
+    IE_NAME = 'arte.tv:concert'
+    _VALID_URL = r'https?://concert\.arte\.tv/(?P<lang>de|fr)/(?P<id>.+)'
+
+    _TEST = {
+        'url': 'http://concert.arte.tv/de/notwist-im-pariser-konzertclub-divan-du-monde',
+        'md5': '9ea035b7bd69696b67aa2ccaaa218161',
+        'info_dict': {
+            'id': '186',
+            'ext': 'mp4',
+            'title': 'The Notwist im Pariser Konzertclub "Divan du Monde"',
+            'upload_date': '20140128',
+            'description': 'md5:486eb08f991552ade77439fe6d82c305',
+        },
+    }
+
+
+class ArteTVEmbedIE(ArteTVPlus7IE):
+    IE_NAME = 'arte.tv:embed'
+    _VALID_URL = r'''(?x)
+        http://www\.arte\.tv
+        /playerv2/embed\.php\?json_url=
+        (?P<json_url>
+            http://arte\.tv/papi/tvguide/videos/stream/player/
+            (?P<lang>[^/]+)/(?P<id>[^/]+)[^&]*
+        )
+    '''
+
+    def _real_extract(self, url):
+        mobj = re.match(self._VALID_URL, url)
+        video_id = mobj.group('id')
+        lang = mobj.group('lang')
+        json_url = mobj.group('json_url')
+        return self._extract_from_json_url(json_url, video_id, lang)