[arte] Add support for embedded videos (Fixes #2620)
[youtube-dl] / youtube_dl / extractor / arte.py
1 # encoding: utf-8
2 from __future__ import unicode_literals
3
4 import re
5 import json
6
7 from .common import InfoExtractor
8 from ..utils import (
9     ExtractorError,
10     find_xpath_attr,
11     unified_strdate,
12     determine_ext,
13     get_element_by_id,
14     compat_str,
15     get_element_by_attribute,
16 )
17
18 # There are different sources of video in arte.tv, the extraction process 
19 # is different for each one. The videos usually expire in 7 days, so we can't
20 # add tests.
21
22
23 class ArteTvIE(InfoExtractor):
24     _VALID_URL = r'http://videos\.arte\.tv/(?P<lang>fr|de)/.*-(?P<id>.*?)\.html'
25     IE_NAME = 'arte.tv'
26
27     def _real_extract(self, url):
28         ref_xml_url = url.replace('/videos/', '/do_delegate/videos/')
29         ref_xml_url = ref_xml_url.replace('.html', ',view,asPlayerXml.xml')
30         ref_xml_doc = self._download_xml(
31             ref_xml_url, video_id, note='Downloading metadata')
32         config_node = find_xpath_attr(ref_xml_doc, './/video', 'lang', lang)
33         config_xml_url = config_node.attrib['ref']
34         config = self._download_xml(
35             config_xml_url, video_id, note='Downloading configuration')
36
37         formats = [{
38             'forma_id': q.attrib['quality'],
39             'url': q.text,
40             'quality': 2 if q.attrib['quality'] == 'hd' else 1,
41         } for q in config.findall('.//quality')]
42         self._sort_formats(formats)
43
44         title = config.find('.//name').text
45         thumbnail = config.find('.//firstThumbnailUrl').text
46         return {
47             'id': video_id,
48             'title': title,
49             'thumbnail': thumbnail,
50             'url': video_url,
51             'ext': 'flv',
52         }
53
54
55 class ArteTVPlus7IE(InfoExtractor):
56     IE_NAME = 'arte.tv:+7'
57     _VALID_URL = r'https?://(?:www\.)?arte\.tv/guide/(?P<lang>fr|de)/(?:(?:sendungen|emissions)/)?(?P<id>.*?)/(?P<name>.*?)(\?.*)?'
58
59     @classmethod
60     def _extract_url_info(cls, url):
61         mobj = re.match(cls._VALID_URL, url)
62         lang = mobj.group('lang')
63         # This is not a real id, it can be for example AJT for the news
64         # http://www.arte.tv/guide/fr/emissions/AJT/arte-journal
65         video_id = mobj.group('id')
66         return video_id, lang
67
68     def _real_extract(self, url):
69         video_id, lang = self._extract_url_info(url)
70         webpage = self._download_webpage(url, video_id)
71         return self._extract_from_webpage(webpage, video_id, lang)
72
73     def _extract_from_webpage(self, webpage, video_id, lang):
74         json_url = self._html_search_regex(r'arte_vp_url="(.*?)"', webpage, 'json url')
75         return self._extract_from_json_url(json_url, video_id, lang)
76
77     def _extract_from_json_url(self, json_url, video_id, lang):
78         info = self._download_json(json_url, video_id)
79         player_info = info['videoJsonPlayer']
80
81         info_dict = {
82             'id': player_info['VID'],
83             'title': player_info['VTI'],
84             'description': player_info.get('VDE'),
85             'upload_date': unified_strdate(player_info.get('VDA', '').split(' ')[0]),
86             'thumbnail': player_info.get('programImage') or player_info.get('VTU', {}).get('IUR'),
87         }
88
89         all_formats = player_info['VSR'].values()
90         # Some formats use the m3u8 protocol
91         all_formats = list(filter(lambda f: f.get('videoFormat') != 'M3U8', all_formats))
92         def _match_lang(f):
93             if f.get('versionCode') is None:
94                 return True
95             # Return true if that format is in the language of the url
96             if lang == 'fr':
97                 l = 'F'
98             elif lang == 'de':
99                 l = 'A'
100             else:
101                 l = lang
102             regexes = [r'VO?%s' % l, r'VO?.-ST%s' % l]
103             return any(re.match(r, f['versionCode']) for r in regexes)
104         # Some formats may not be in the same language as the url
105         formats = filter(_match_lang, all_formats)
106         formats = list(formats) # in python3 filter returns an iterator
107         if not formats:
108             # Some videos are only available in the 'Originalversion'
109             # they aren't tagged as being in French or German
110             if all(f['versionCode'] == 'VO' for f in all_formats):
111                 formats = all_formats
112             else:
113                 raise ExtractorError(u'The formats list is empty')
114
115         if re.match(r'[A-Z]Q', formats[0]['quality']) is not None:
116             def sort_key(f):
117                 return ['HQ', 'MQ', 'EQ', 'SQ'].index(f['quality'])
118         else:
119             def sort_key(f):
120                 return (
121                     # Sort first by quality
122                     int(f.get('height',-1)),
123                     int(f.get('bitrate',-1)),
124                     # The original version with subtitles has lower relevance
125                     re.match(r'VO-ST(F|A)', f.get('versionCode', '')) is None,
126                     # The version with sourds/mal subtitles has also lower relevance
127                     re.match(r'VO?(F|A)-STM\1', f.get('versionCode', '')) is None,
128                     # Prefer http downloads over m3u8
129                     0 if f['url'].endswith('m3u8') else 1,
130                 )
131         formats = sorted(formats, key=sort_key)
132         def _format(format_info):
133             quality = ''
134             height = format_info.get('height')
135             if height is not None:
136                 quality = compat_str(height)
137             bitrate = format_info.get('bitrate')
138             if bitrate is not None:
139                 quality += '-%d' % bitrate
140             if format_info.get('versionCode') is not None:
141                 format_id = '%s-%s' % (quality, format_info['versionCode'])
142             else:
143                 format_id = quality
144             info = {
145                 'format_id': format_id,
146                 'format_note': format_info.get('versionLibelle'),
147                 'width': format_info.get('width'),
148                 'height': height,
149             }
150             if format_info['mediaType'] == 'rtmp':
151                 info['url'] = format_info['streamer']
152                 info['play_path'] = 'mp4:' + format_info['url']
153                 info['ext'] = 'flv'
154             else:
155                 info['url'] = format_info['url']
156                 info['ext'] = determine_ext(info['url'])
157             return info
158         info_dict['formats'] = [_format(f) for f in formats]
159
160         return info_dict
161
162
163 # It also uses the arte_vp_url url from the webpage to extract the information
164 class ArteTVCreativeIE(ArteTVPlus7IE):
165     IE_NAME = 'arte.tv:creative'
166     _VALID_URL = r'https?://creative\.arte\.tv/(?P<lang>fr|de)/magazine?/(?P<id>.+)'
167
168     _TEST = {
169         'url': 'http://creative.arte.tv/de/magazin/agentur-amateur-corporate-design',
170         'info_dict': {
171             'id': '050489-002',
172             'ext': 'mp4',
173             'title': 'Agentur Amateur / Agence Amateur #2 : Corporate Design',
174         },
175     }
176
177
178 class ArteTVFutureIE(ArteTVPlus7IE):
179     IE_NAME = 'arte.tv:future'
180     _VALID_URL = r'https?://future\.arte\.tv/(?P<lang>fr|de)/(thema|sujet)/.*?#article-anchor-(?P<id>\d+)'
181
182     _TEST = {
183         'url': 'http://future.arte.tv/fr/sujet/info-sciences#article-anchor-7081',
184         'info_dict': {
185             'id': '050940-003',
186             'ext': 'mp4',
187             'title': 'Les champignons au secours de la planète',
188         },
189     }
190
191     def _real_extract(self, url):
192         anchor_id, lang = self._extract_url_info(url)
193         webpage = self._download_webpage(url, anchor_id)
194         row = get_element_by_id(anchor_id, webpage)
195         return self._extract_from_webpage(row, anchor_id, lang)
196
197
198 class ArteTVDDCIE(ArteTVPlus7IE):
199     IE_NAME = 'arte.tv:ddc'
200     _VALID_URL = r'https?://ddc\.arte\.tv/(?P<lang>emission|folge)/(?P<id>.+)'
201
202     def _real_extract(self, url):
203         video_id, lang = self._extract_url_info(url)
204         if lang == 'folge':
205             lang = 'de'
206         elif lang == 'emission':
207             lang = 'fr'
208         webpage = self._download_webpage(url, video_id)
209         scriptElement = get_element_by_attribute('class', 'visu_video_block', webpage)
210         script_url = self._html_search_regex(r'src="(.*?)"', scriptElement, 'script url')
211         javascriptPlayerGenerator = self._download_webpage(script_url, video_id, 'Download javascript player generator')
212         json_url = self._search_regex(r"json_url=(.*)&rendering_place.*", javascriptPlayerGenerator, 'json url')
213         return self._extract_from_json_url(json_url, video_id, lang)
214
215
216 class ArteTVConcertIE(ArteTVPlus7IE):
217     IE_NAME = 'arte.tv:concert'
218     _VALID_URL = r'https?://concert\.arte\.tv/(?P<lang>de|fr)/(?P<id>.+)'
219
220     _TEST = {
221         'url': 'http://concert.arte.tv/de/notwist-im-pariser-konzertclub-divan-du-monde',
222         'md5': '9ea035b7bd69696b67aa2ccaaa218161',
223         'info_dict': {
224             'id': '186',
225             'ext': 'mp4',
226             'title': 'The Notwist im Pariser Konzertclub "Divan du Monde"',
227             'upload_date': '20140128',
228             'description': 'md5:486eb08f991552ade77439fe6d82c305',
229         },
230     }
231
232
233 class ArteTVEmbedIE(ArteTVPlus7IE):
234     IE_NAME = 'arte.tv:embed'
235     _VALID_URL = r'''(?x)
236         http://www\.arte\.tv
237         /playerv2/embed\.php\?json_url=
238         (?P<json_url>
239             http://arte\.tv/papi/tvguide/videos/stream/player/
240             (?P<lang>[^/]+)/(?P<id>[^/]+)[^&]*
241         )
242     '''
243
244     def _real_extract(self, url):
245         mobj = re.match(self._VALID_URL, url)
246         video_id = mobj.group('id')
247         lang = mobj.group('lang')
248         json_url = mobj.group('json_url')
249         return self._extract_from_json_url(json_url, video_id, lang)