X-Git-Url: http://git.bitcoin.ninja/index.cgi?a=blobdiff_plain;f=youtube_dl%2Fextractor%2Farte.py;h=e10c74c112a0b7bbb9335e76cdff88524c4cbafb;hb=3e56add7c9acf659d984be5aef1b8fb6e42599c6;hp=48f7494605f7cb29018248dcad3aab97029fc80c;hpb=d5822b96b00fce48e04a14953c4cb25cef1cdbaf;p=youtube-dl

diff --git a/youtube_dl/extractor/arte.py b/youtube_dl/extractor/arte.py
index 48f749460..e10c74c11 100644
--- a/youtube_dl/extractor/arte.py
+++ b/youtube_dl/extractor/arte.py
@@ -1,134 +1,251 @@
+# encoding: utf-8
 import re
-import socket
+import json
+import xml.etree.ElementTree
 
 from .common import InfoExtractor
 from ..utils import (
-    compat_http_client,
-    compat_str,
-    compat_urllib_error,
-    compat_urllib_parse,
-    compat_urllib_request,
-
     ExtractorError,
+    find_xpath_attr,
     unified_strdate,
+    determine_ext,
+    get_element_by_id,
 )
 
-class ArteTvIE(InfoExtractor):
-    """arte.tv information extractor."""
+# 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.
 
-    _VALID_URL = r'(?:http://)?videos\.arte\.tv/(?:fr|de)/videos/.*'
+class ArteTvIE(InfoExtractor):
+    _VIDEOS_URL = r'(?:http://)?videos.arte.tv/(?P<lang>fr|de)/.*-(?P<id>.*?).html'
+    _LIVEWEB_URL = r'(?:http://)?liveweb.arte.tv/(?P<lang>fr|de)/(?P<subpage>.+?)/(?P<name>.+)'
     _LIVE_URL = r'index-[0-9]+\.html$'
 
     IE_NAME = u'arte.tv'
 
-    def fetch_webpage(self, url):
-        request = compat_urllib_request.Request(url)
-        try:
-            self.report_download_webpage(url)
-            webpage = compat_urllib_request.urlopen(request).read()
-        except (compat_urllib_error.URLError, compat_http_client.HTTPException, socket.error) as err:
-            raise ExtractorError(u'Unable to retrieve video webpage: %s' % compat_str(err))
-        except ValueError as err:
-            raise ExtractorError(u'Invalid URL: %s' % url)
-        return webpage
-
-    def grep_webpage(self, url, regex, regexFlags, matchTuples):
-        page = self.fetch_webpage(url)
-        mobj = re.search(regex, page, regexFlags)
-        info = {}
-
-        if mobj is None:
-            raise ExtractorError(u'Invalid URL: %s' % url)
-
-        for (i, key, err) in matchTuples:
-            if mobj.group(i) is None:
-                raise ExtractorError(err)
+    @classmethod
+    def suitable(cls, url):
+        return any(re.match(regex, url) for regex in (cls._VIDEOS_URL, cls._LIVEWEB_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 _real_extract(self, url):
+        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)
+
+        mobj = re.match(self._LIVEWEB_URL, url)
+        if mobj is not None:
+            name = mobj.group('name')
+            lang = mobj.group('lang')
+            return self._extract_liveweb(url, name, 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_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:
-                info[key] = mobj.group(i)
-
-        return info
-
-    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 extractPlus7Stream(self, url):
-        video_lang = url.split('/')[-3]
-        info = self.grep_webpage(
-            url,
-            r'param name="movie".*?videorefFileUrl=(http[^\'"&]*)',
-            0,
-            [
-                (1, 'url', u'Invalid URL: %s' % url)
-            ]
-        )
-        next_url = compat_urllib_parse.unquote(info.get('url'))
-        info = self.grep_webpage(
-            next_url,
-            r'<video lang="%s" ref="(http[^\'"&]*)' % video_lang,
-            0,
-            [
-                (1, 'url', u'Could not find <video> tag: %s' % url)
-            ]
-        )
-        next_url = compat_urllib_parse.unquote(info.get('url'))
-
-        info = self.grep_webpage(
-            next_url,
-            r'<video id="(.*?)".*?>.*?' +
-                '<name>(.*?)</name>.*?' +
-                '<dateVideo>(.*?)</dateVideo>.*?' +
-                '<url quality="hd">(.*?)</url>',
-            re.DOTALL,
-            [
-                (1, 'id',    u'could not extract video id: %s' % url),
-                (2, 'title', u'could not extract video title: %s' % url),
-                (3, 'date',  u'could not extract video date: %s' % url),
-                (4, 'url',   u'could not extract video url: %s' % url)
-            ]
-        )
-
-        return {
-            'id':           info.get('id'),
-            'url':          compat_urllib_parse.unquote(info.get('url')),
-            'uploader':     u'arte.tv',
-            'upload_date':  unified_strdate(info.get('date')),
-            'title':        info.get('title').decode('utf-8'),
-            'ext':          u'mp4',
-            'format':       u'NA',
-            'player_url':   None,
-        }
+                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',
+                }
+
+    def _extract_liveweb(self, url, name, lang):
+        """Extract form http://liveweb.arte.tv/"""
+        webpage = self._download_webpage(url, name)
+        video_id = self._search_regex(r'eventId=(\d+?)("|&)', webpage, u'event id')
+        config_xml = self._download_webpage('http://download.liveweb.arte.tv/o21/liveweb/events/event-%s.xml' % video_id,
+                                            video_id, u'Downloading information')
+        config_doc = xml.etree.ElementTree.fromstring(config_xml.encode('utf-8'))
+        event_doc = config_doc.find('event')
+        url_node = event_doc.find('video').find('urlHd')
+        if url_node is None:
+            url_node = video_doc.find('urlSd')
+
+        return {'id': video_id,
+                'title': event_doc.find('name%s' % lang.capitalize()).text,
+                'url': url_node.text.replace('MP4', 'mp4'),
+                'ext': 'flv',
+                'thumbnail': self._og_search_thumbnail(webpage),
+                }
+
+
+class ArteTVPlus7IE(InfoExtractor):
+    IE_NAME = u'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 = url.split('/')[-1]
+        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 url')
+
+        json_info = self._download_webpage(json_url, video_id, 'Downloading info json')
         self.report_extraction(video_id)
+        info = json.loads(json_info)
+        player_info = info['videoJsonPlayer']
+
+        info_dict = {
+            'id': player_info['VID'],
+            'title': player_info['VTI'],
+            'description': player_info.get('VDE'),
+            'upload_date': unified_strdate(player_info.get('VDA', '').split(' ')[0]),
+            'thumbnail': player_info.get('programImage') or player_info.get('VTU', {}).get('IUR'),
+        }
 
-        if re.search(self._LIVE_URL, video_id) is not None:
-            self.extractLiveStream(url)
-            return
+        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'
+            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
+            else:
+                raise ExtractorError(u'The formats list is empty')
+        # We order the formats by quality
+        if re.match(r'[A-Z]Q', formats[0]['quality']) is not None:
+            sort_key = lambda f: ['HQ', 'MQ', 'EQ', 'SQ'].index(f['quality'])
         else:
-            info = self.extractPlus7Stream(url)
+            sort_key = lambda f: int(f.get('height',-1))
+        formats = sorted(formats, key=sort_key)
+        # Prefer videos without subtitles in the same language
+        formats = sorted(formats, key=lambda f: re.match(r'VO(F|A)-STM\1', f.get('versionCode', '')) is None)
+        # Pick the best quality
+        def _format(format_info):
+            quality = format_info['quality']
+            m_quality = re.match(r'\w*? - (\d*)p', quality)
+            if m_quality is not None:
+                quality = m_quality.group(1)
+            if format_info.get('versionCode') is not None:
+                format_id = u'%s-%s' % (quality, format_info['versionCode'])
+            else:
+                format_id = quality
+            info = {
+                'format_id': format_id,
+                'format_note': format_info.get('versionLibelle'),
+                'width': format_info.get('width'),
+                'height': format_info.get('height'),
+            }
+            if format_info['mediaType'] == u'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
+
+
+# It also uses the arte_vp_url url from the webpage to extract the information
+class ArteTVCreativeIE(ArteTVPlus7IE):
+    IE_NAME = u'arte.tv:creative'
+    _VALID_URL = r'https?://creative\.arte\.tv/(?P<lang>fr|de)/magazine?/(?P<id>.+)'
 
-        return [info]
+    _TEST = {
+        u'url': u'http://creative.arte.tv/de/magazin/agentur-amateur-corporate-design',
+        u'file': u'050489-002.mp4',
+        u'info_dict': {
+            u'title': u'Agentur Amateur / Agence Amateur #2 : Corporate Design',
+        },
+    }
+
+
+class ArteTVFutureIE(ArteTVPlus7IE):
+    IE_NAME = u'arte.tv:future'
+    _VALID_URL = r'https?://future\.arte\.tv/(?P<lang>fr|de)/(thema|sujet)/.*?#article-anchor-(?P<id>\d+)'
+
+    _TEST = {
+        u'url': u'http://future.arte.tv/fr/sujet/info-sciences#article-anchor-7081',
+        u'file': u'050940-003.mp4',
+        u'info_dict': {
+            u'title': u'Les champignons au secours de la planète',
+        },
+    }
+
+    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)