]> Raphaël G. Git Repositories - youtubedl/blobdiff - youtube_dl/extractor/arte.py
Prepare to upload
[youtubedl] / youtube_dl / extractor / arte.py
index b528a9ec50ca6c2dac1a52fe66de2cd66194dd23..2bd3bfe8a852159cf7aefa93e731bab995ca7d23 100644 (file)
-# encoding: utf-8
+# coding: utf-8
 from __future__ import unicode_literals
 
 import re
 
 from .common import InfoExtractor
+from ..compat import compat_str
 from ..utils import (
     ExtractorError,
-    find_xpath_attr,
+    int_or_none,
+    qualities,
+    try_get,
     unified_strdate,
-    determine_ext,
-    get_element_by_id,
-    compat_str,
-    get_element_by_attribute,
 )
 
-# There are different sources of video in arte.tv, the extraction process 
+# 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):
-    _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'],
-            'url': q.text,
-            'ext': 'flv',
-            'quality': 2 if q.attrib['quality'] == 'hd' else 1,
-        } for q in config.findall('./urls/url')]
-        self._sort_formats(formats)
-
-        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
-    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):
-        video_id, lang = self._extract_url_info(url)
-        webpage = self._download_webpage(url, video_id)
-        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="(.*?)"', webpage, 'json vp url')
-        return self._extract_from_json_url(json_url, video_id, lang)
-
-    def _extract_from_json_url(self, json_url, video_id, lang):
+class ArteTVBaseIE(InfoExtractor):
+    def _extract_from_json_url(self, json_url, video_id, lang, title=None):
         info = self._download_json(json_url, video_id)
         player_info = info['videoJsonPlayer']
 
+        vsr = try_get(player_info, lambda x: x['VSR'], dict)
+        if not vsr:
+            error = None
+            if try_get(player_info, lambda x: x['custom_msg']['type']) == 'error':
+                error = try_get(
+                    player_info, lambda x: x['custom_msg']['msg'], compat_str)
+            if not error:
+                error = 'Video %s is not available' % player_info.get('VID') or video_id
+            raise ExtractorError(error, expected=True)
+
+        upload_date_str = player_info.get('shootingDate')
+        if not upload_date_str:
+            upload_date_str = (player_info.get('VRA') or player_info.get('VDA') or '').split(' ')[0]
+
+        title = (player_info.get('VTI') or title or player_info['VID']).strip()
+        subtitle = player_info.get('VSU', '').strip()
+        if subtitle:
+            title += ' - %s' % subtitle
+
         info_dict = {
             'id': player_info['VID'],
-            'title': player_info['VTI'],
+            'title': title,
             'description': player_info.get('VDE'),
-            'upload_date': unified_strdate(player_info.get('VDA', '').split(' ')[0]),
+            'upload_date': unified_strdate(upload_date_str),
             'thumbnail': player_info.get('programImage') or player_info.get('VTU', {}).get('IUR'),
         }
+        qfunc = qualities(['MQ', 'HQ', 'EQ', 'SQ'])
+
+        LANGS = {
+            'fr': 'F',
+            'de': 'A',
+            'en': 'E[ANG]',
+            'es': 'E[ESP]',
+            'it': 'E[ITA]',
+            'pl': 'E[POL]',
+        }
 
-        all_formats = player_info['VSR'].values()
-        # Some formats use the m3u8 protocol
-        all_formats = list(filter(lambda f: f.get('videoFormat') != 'M3U8', all_formats))
-        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'
-            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
-        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
-            if all(f['versionCode'] == 'VO' for f in all_formats):
-                formats = all_formats
+        langcode = LANGS.get(lang, lang)
+
+        formats = []
+        for format_id, format_dict in vsr.items():
+            f = dict(format_dict)
+            versionCode = f.get('versionCode')
+            l = re.escape(langcode)
+
+            # Language preference from most to least priority
+            # Reference: section 6.8 of
+            # https://www.arte.tv/sites/en/corporate/files/complete-technical-guidelines-arte-geie-v1-07-1.pdf
+            PREFERENCES = (
+                # original version in requested language, without subtitles
+                r'VO{0}$'.format(l),
+                # original version in requested language, with partial subtitles in requested language
+                r'VO{0}-ST{0}$'.format(l),
+                # original version in requested language, with subtitles for the deaf and hard-of-hearing in requested language
+                r'VO{0}-STM{0}$'.format(l),
+                # non-original (dubbed) version in requested language, without subtitles
+                r'V{0}$'.format(l),
+                # non-original (dubbed) version in requested language, with subtitles partial subtitles in requested language
+                r'V{0}-ST{0}$'.format(l),
+                # non-original (dubbed) version in requested language, with subtitles for the deaf and hard-of-hearing in requested language
+                r'V{0}-STM{0}$'.format(l),
+                # original version in requested language, with partial subtitles in different language
+                r'VO{0}-ST(?!{0}).+?$'.format(l),
+                # original version in requested language, with subtitles for the deaf and hard-of-hearing in different language
+                r'VO{0}-STM(?!{0}).+?$'.format(l),
+                # original version in different language, with partial subtitles in requested language
+                r'VO(?:(?!{0}).+?)?-ST{0}$'.format(l),
+                # original version in different language, with subtitles for the deaf and hard-of-hearing in requested language
+                r'VO(?:(?!{0}).+?)?-STM{0}$'.format(l),
+                # original version in different language, without subtitles
+                r'VO(?:(?!{0}))?$'.format(l),
+                # original version in different language, with partial subtitles in different language
+                r'VO(?:(?!{0}).+?)?-ST(?!{0}).+?$'.format(l),
+                # original version in different language, with subtitles for the deaf and hard-of-hearing in different language
+                r'VO(?:(?!{0}).+?)?-STM(?!{0}).+?$'.format(l),
+            )
+
+            for pref, p in enumerate(PREFERENCES):
+                if re.match(p, versionCode):
+                    lang_pref = len(PREFERENCES) - pref
+                    break
             else:
-                raise ExtractorError(u'The formats list is empty')
+                lang_pref = -1
 
-        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:
-            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):
-            quality = ''
-            height = format_info.get('height')
-            if height is not None:
-                quality = compat_str(height)
-            bitrate = format_info.get('bitrate')
-            if bitrate is not None:
-                quality += '-%d' % bitrate
-            if format_info.get('versionCode') is not None:
-                format_id = '%s-%s' % (quality, format_info['versionCode'])
-            else:
-                format_id = quality
-            info = {
+            format = {
                 'format_id': format_id,
-                'format_note': format_info.get('versionLibelle'),
-                'width': format_info.get('width'),
-                'height': height,
+                'preference': -10 if f.get('videoFormat') == 'M3U8' else None,
+                'language_preference': lang_pref,
+                'format_note': '%s, %s' % (f.get('versionCode'), f.get('versionLibelle')),
+                'width': int_or_none(f.get('width')),
+                'height': int_or_none(f.get('height')),
+                'tbr': int_or_none(f.get('bitrate')),
+                'quality': qfunc(f.get('quality')),
             }
-            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
+            if f.get('mediaType') == 'rtmp':
+                format['url'] = f['streamer']
+                format['play_path'] = 'mp4:' + f['url']
+                format['ext'] = 'flv'
+            else:
+                format['url'] = f['url']
 
+            formats.append(format)
 
-# 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>.+)'
+        self._check_formats(formats, video_id)
+        self._sort_formats(formats)
 
-    _TEST = {
-        'url': 'http://creative.arte.tv/de/magazin/agentur-amateur-corporate-design',
-        'info_dict': {
-            'id': '050489-002',
-            'ext': 'mp4',
-            'title': 'Agentur Amateur / Agence Amateur #2 : Corporate Design',
-        },
-    }
+        info_dict['formats'] = formats
+        return info_dict
 
 
-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+)'
+class ArteTVPlus7IE(ArteTVBaseIE):
+    IE_NAME = 'arte.tv:+7'
+    _VALID_URL = r'https?://(?:www\.)?arte\.tv/(?P<lang>fr|de|en|es|it|pl)/videos/(?P<id>\d{6}-\d{3}-[AF])'
 
-    _TEST = {
-        'url': 'http://future.arte.tv/fr/sujet/info-sciences#article-anchor-7081',
+    _TESTS = [{
+        'url': 'https://www.arte.tv/en/videos/088501-000-A/mexico-stealing-petrol-to-survive/',
         'info_dict': {
-            'id': '050940-003',
+            'id': '088501-000-A',
             'ext': 'mp4',
-            'title': 'Les champignons au secours de la planète',
+            'title': 'Mexico: Stealing Petrol to Survive',
+            'upload_date': '20190628',
         },
-    }
+    }]
 
     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)
+        lang, video_id = re.match(self._VALID_URL, url).groups()
+        return self._extract_from_json_url(
+            'https://api.arte.tv/api/player/v1/config/%s/%s' % (lang, video_id),
+            video_id, lang)
+
 
+class ArteTVEmbedIE(ArteTVPlus7IE):
+    IE_NAME = 'arte.tv:embed'
+    _VALID_URL = r'''(?x)
+        https://www\.arte\.tv
+        /player/v3/index\.php\?json_url=
+        (?P<json_url>
+            https?://api\.arte\.tv/api/player/v1/config/
+            (?P<lang>[^/]+)/(?P<id>\d{6}-\d{3}-[AF])
+        )
+    '''
 
-class ArteTVDDCIE(ArteTVPlus7IE):
-    IE_NAME = 'arte.tv:ddc'
-    _VALID_URL = r'https?://ddc\.arte\.tv/(?P<lang>emission|folge)/(?P<id>.+)'
+    _TESTS = []
 
     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')
+        json_url, lang, video_id = re.match(self._VALID_URL, url).groups()
         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>.+)'
+class ArteTVPlaylistIE(ArteTVBaseIE):
+    IE_NAME = 'arte.tv:playlist'
+    _VALID_URL = r'https?://(?:www\.)?arte\.tv/(?P<lang>fr|de|en|es|it|pl)/videos/(?P<id>RC-\d{6})'
 
-    _TEST = {
-        'url': 'http://concert.arte.tv/de/notwist-im-pariser-konzertclub-divan-du-monde',
-        'md5': '9ea035b7bd69696b67aa2ccaaa218161',
+    _TESTS = [{
+        'url': 'https://www.arte.tv/en/videos/RC-016954/earn-a-living/',
         'info_dict': {
-            'id': '186',
-            'ext': 'mp4',
-            'title': 'The Notwist im Pariser Konzertclub "Divan du Monde"',
-            'upload_date': '20140128',
-            'description': 'md5:486eb08f991552ade77439fe6d82c305',
+            'id': 'RC-016954',
+            'title': 'Earn a Living',
+            'description': 'md5:d322c55011514b3a7241f7fb80d494c2',
         },
-    }
-
-
-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>[^/]+)[^&]*
-        )
-    '''
+        'playlist_mincount': 6,
+    }]
 
     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)
+        lang, playlist_id = re.match(self._VALID_URL, url).groups()
+        collection = self._download_json(
+            'https://api.arte.tv/api/player/v1/collectionData/%s/%s?source=videos'
+            % (lang, playlist_id), playlist_id)
+        title = collection.get('title')
+        description = collection.get('shortDescription') or collection.get('teaserText')
+        entries = [
+            self._extract_from_json_url(
+                video['jsonUrl'], video.get('programId') or playlist_id, lang)
+            for video in collection['videos'] if video.get('jsonUrl')]
+        return self.playlist_result(entries, playlist_id, title, description)