]> Raphaƫl G. Git Repositories - youtubedl/blobdiff - youtube_dl/extractor/common.py
Prepate to upload
[youtubedl] / youtube_dl / extractor / common.py
index e2d9f52b018c25abc5a58a93785473ff88d90b74..59b9d373951331ef5c02bb6ed80d0f28bfaf4c8f 100644 (file)
@@ -174,6 +174,8 @@ class InfoExtractor(object):
                                  width : height ratio as float.
                     * no_resume  The server does not support resuming the
                                  (HTTP or RTMP) download. Boolean.
+                    * downloader_options  A dictionary of downloader options as
+                                 described in FileDownloader
 
     url:            Final video URL.
     ext:            Video filename extension.
@@ -301,8 +303,9 @@ class InfoExtractor(object):
     There must be a key "entries", which is a list, an iterable, or a PagedList
     object, each element of which is a valid dictionary by this specification.
 
-    Additionally, playlists can have "title", "description" and "id" attributes
-    with the same semantics as videos (see above).
+    Additionally, playlists can have "id", "title", "description", "uploader",
+    "uploader_id", "uploader_url" attributes with the same semantics as videos
+    (see above).
 
 
     _type "multi_video" indicates that there are multiple videos that
@@ -494,6 +497,16 @@ class InfoExtractor(object):
                 self.to_screen('%s' % (note,))
             else:
                 self.to_screen('%s: %s' % (video_id, note))
+
+        # Some sites check X-Forwarded-For HTTP header in order to figure out
+        # the origin of the client behind proxy. This allows bypassing geo
+        # restriction by faking this header's value to IP that belongs to some
+        # geo unrestricted country. We will do so once we encounter any
+        # geo restriction error.
+        if self._x_forwarded_for_ip:
+            if 'X-Forwarded-For' not in headers:
+                headers['X-Forwarded-For'] = self._x_forwarded_for_ip
+
         if isinstance(url_or_request, compat_urllib_request.Request):
             url_or_request = update_Request(
                 url_or_request, data=data, headers=headers, query=query)
@@ -523,15 +536,6 @@ class InfoExtractor(object):
         if isinstance(url_or_request, (compat_str, str)):
             url_or_request = url_or_request.partition('#')[0]
 
-        # Some sites check X-Forwarded-For HTTP header in order to figure out
-        # the origin of the client behind proxy. This allows bypassing geo
-        # restriction by faking this header's value to IP that belongs to some
-        # geo unrestricted country. We will do so once we encounter any
-        # geo restriction error.
-        if self._x_forwarded_for_ip:
-            if 'X-Forwarded-For' not in headers:
-                headers['X-Forwarded-For'] = self._x_forwarded_for_ip
-
         urlh = self._request_webpage(url_or_request, video_id, note, errnote, fatal, data=data, headers=headers, query=query)
         if urlh is False:
             assert not fatal
@@ -592,19 +596,11 @@ class InfoExtractor(object):
         if not encoding:
             encoding = self._guess_encoding_from_content(content_type, webpage_bytes)
         if self._downloader.params.get('dump_intermediate_pages', False):
-            try:
-                url = url_or_request.get_full_url()
-            except AttributeError:
-                url = url_or_request
-            self.to_screen('Dumping request to ' + url)
+            self.to_screen('Dumping request to ' + urlh.geturl())
             dump = base64.b64encode(webpage_bytes).decode('ascii')
             self._downloader.to_screen(dump)
         if self._downloader.params.get('write_pages', False):
-            try:
-                url = url_or_request.get_full_url()
-            except AttributeError:
-                url = url_or_request
-            basen = '%s_%s' % (video_id, url)
+            basen = '%s_%s' % (video_id, urlh.geturl())
             if len(basen) > 240:
                 h = '___' + hashlib.md5(basen.encode('utf-8')).hexdigest()
                 basen = basen[:240 - len(h)] + h
@@ -648,19 +644,31 @@ class InfoExtractor(object):
             content, _ = res
             return content
 
+    def _download_xml_handle(
+            self, url_or_request, video_id, note='Downloading XML',
+            errnote='Unable to download XML', transform_source=None,
+            fatal=True, encoding=None, data=None, headers={}, query={}):
+        """Return a tuple (xml as an xml.etree.ElementTree.Element, URL handle)"""
+        res = self._download_webpage_handle(
+            url_or_request, video_id, note, errnote, fatal=fatal,
+            encoding=encoding, data=data, headers=headers, query=query)
+        if res is False:
+            return res
+        xml_string, urlh = res
+        return self._parse_xml(
+            xml_string, video_id, transform_source=transform_source,
+            fatal=fatal), urlh
+
     def _download_xml(self, url_or_request, video_id,
                       note='Downloading XML', errnote='Unable to download XML',
                       transform_source=None, fatal=True, encoding=None,
                       data=None, headers={}, query={}):
         """Return the xml as an xml.etree.ElementTree.Element"""
-        xml_string = self._download_webpage(
-            url_or_request, video_id, note, errnote, fatal=fatal,
-            encoding=encoding, data=data, headers=headers, query=query)
-        if xml_string is False:
-            return xml_string
-        return self._parse_xml(
-            xml_string, video_id, transform_source=transform_source,
-            fatal=fatal)
+        res = self._download_xml_handle(
+            url_or_request, video_id, note=note, errnote=errnote,
+            transform_source=transform_source, fatal=fatal, encoding=encoding,
+            data=data, headers=headers, query=query)
+        return res if res is False else res[0]
 
     def _parse_xml(self, xml_string, video_id, transform_source=None, fatal=True):
         if transform_source:
@@ -1017,7 +1025,7 @@ class InfoExtractor(object):
             })
 
         for e in json_ld:
-            if e.get('@context') == 'http://schema.org':
+            if isinstance(e.get('@context'), compat_str) and re.match(r'^https?://schema.org/?$', e.get('@context')):
                 item_type = e.get('@type')
                 if expected_type is not None and expected_type != item_type:
                     return info
@@ -1033,7 +1041,7 @@ class InfoExtractor(object):
                     part_of_series = e.get('partOfSeries') or e.get('partOfTVSeries')
                     if isinstance(part_of_series, dict) and part_of_series.get('@type') in ('TVSeries', 'Series', 'CreativeWorkSeries'):
                         info['series'] = unescapeHTML(part_of_series.get('name'))
-                elif item_type == 'Article':
+                elif item_type in ('Article', 'NewsArticle'):
                     info.update({
                         'timestamp': parse_iso8601(e.get('datePublished')),
                         'title': unescapeHTML(e.get('headline')),
@@ -1356,6 +1364,9 @@ class InfoExtractor(object):
         if '#EXT-X-FAXS-CM:' in m3u8_doc:  # Adobe Flash Access
             return []
 
+        if re.search(r'#EXT-X-SESSION-KEY:.*?URI="skd://', m3u8_doc):  # Apple FairPlay
+            return []
+
         formats = []
 
         format_url = lambda u: (
@@ -1695,22 +1706,24 @@ class InfoExtractor(object):
             })
         return subtitles
 
-    def _extract_xspf_playlist(self, playlist_url, playlist_id, fatal=True):
+    def _extract_xspf_playlist(self, xspf_url, playlist_id, fatal=True):
         xspf = self._download_xml(
-            playlist_url, playlist_id, 'Downloading xpsf playlist',
+            xspf_url, playlist_id, 'Downloading xpsf playlist',
             'Unable to download xspf manifest', fatal=fatal)
         if xspf is False:
             return []
-        return self._parse_xspf(xspf, playlist_id)
+        return self._parse_xspf(
+            xspf, playlist_id, xspf_url=xspf_url,
+            xspf_base_url=base_url(xspf_url))
 
-    def _parse_xspf(self, playlist, playlist_id):
+    def _parse_xspf(self, xspf_doc, playlist_id, xspf_url=None, xspf_base_url=None):
         NS_MAP = {
             'xspf': 'http://xspf.org/ns/0/',
             's1': 'http://static.streamone.nl/player/ns/0',
         }
 
         entries = []
-        for track in playlist.findall(xpath_with_ns('./xspf:trackList/xspf:track', NS_MAP)):
+        for track in xspf_doc.findall(xpath_with_ns('./xspf:trackList/xspf:track', NS_MAP)):
             title = xpath_text(
                 track, xpath_with_ns('./xspf:title', NS_MAP), 'title', default=playlist_id)
             description = xpath_text(
@@ -1720,12 +1733,18 @@ class InfoExtractor(object):
             duration = float_or_none(
                 xpath_text(track, xpath_with_ns('./xspf:duration', NS_MAP), 'duration'), 1000)
 
-            formats = [{
-                'url': location.text,
-                'format_id': location.get(xpath_with_ns('s1:label', NS_MAP)),
-                'width': int_or_none(location.get(xpath_with_ns('s1:width', NS_MAP))),
-                'height': int_or_none(location.get(xpath_with_ns('s1:height', NS_MAP))),
-            } for location in track.findall(xpath_with_ns('./xspf:location', NS_MAP))]
+            formats = []
+            for location in track.findall(xpath_with_ns('./xspf:location', NS_MAP)):
+                format_url = urljoin(xspf_base_url, location.text)
+                if not format_url:
+                    continue
+                formats.append({
+                    'url': format_url,
+                    'manifest_url': xspf_url,
+                    'format_id': location.get(xpath_with_ns('s1:label', NS_MAP)),
+                    'width': int_or_none(location.get(xpath_with_ns('s1:width', NS_MAP))),
+                    'height': int_or_none(location.get(xpath_with_ns('s1:height', NS_MAP))),
+                })
             self._sort_formats(formats)
 
             entries.append({
@@ -1739,18 +1758,18 @@ class InfoExtractor(object):
         return entries
 
     def _extract_mpd_formats(self, mpd_url, video_id, mpd_id=None, note=None, errnote=None, fatal=True, formats_dict={}):
-        res = self._download_webpage_handle(
+        res = self._download_xml_handle(
             mpd_url, video_id,
             note=note or 'Downloading MPD manifest',
             errnote=errnote or 'Failed to download MPD manifest',
             fatal=fatal)
         if res is False:
             return []
-        mpd, urlh = res
+        mpd_doc, urlh = res
         mpd_base_url = base_url(urlh.geturl())
 
         return self._parse_mpd_formats(
-            compat_etree_fromstring(mpd.encode('utf-8')), mpd_id, mpd_base_url,
+            mpd_doc, mpd_id=mpd_id, mpd_base_url=mpd_base_url,
             formats_dict=formats_dict, mpd_url=mpd_url)
 
     def _parse_mpd_formats(self, mpd_doc, mpd_id=None, mpd_base_url='', formats_dict={}, mpd_url=None):
@@ -1883,6 +1902,7 @@ class InfoExtractor(object):
                             'language': lang if lang not in ('mul', 'und', 'zxx', 'mis') else None,
                             'format_note': 'DASH %s' % content_type,
                             'filesize': filesize,
+                            'container': mimetype2ext(mime_type) + '_dash',
                         }
                         f.update(parse_codecs(representation_attrib.get('codecs')))
                         representation_ms_info = extract_multisegment_info(representation, adaption_set_ms_info)
@@ -1980,6 +2000,22 @@ class InfoExtractor(object):
                                     })
                                     segment_index += 1
                             representation_ms_info['fragments'] = fragments
+                        elif 'segment_urls' in representation_ms_info:
+                            # Segment URLs with no SegmentTimeline
+                            # Example: https://www.seznam.cz/zpravy/clanek/cesko-zasahne-vitr-o-sile-vichrice-muze-byt-i-zivotu-nebezpecny-39091
+                            # https://github.com/rg3/youtube-dl/pull/14844
+                            fragments = []
+                            segment_duration = float_or_none(
+                                representation_ms_info['segment_duration'],
+                                representation_ms_info['timescale']) if 'segment_duration' in representation_ms_info else None
+                            for segment_url in representation_ms_info['segment_urls']:
+                                fragment = {
+                                    location_key(segment_url): segment_url,
+                                }
+                                if segment_duration:
+                                    fragment['duration'] = segment_duration
+                                fragments.append(fragment)
+                            representation_ms_info['fragments'] = fragments
                         # NB: MPD manifest may contain direct URLs to unfragmented media.
                         # No fragments key is present in this case.
                         if 'fragments' in representation_ms_info:
@@ -1994,32 +2030,29 @@ class InfoExtractor(object):
                                     f['url'] = initialization_url
                                 f['fragments'].append({location_key(initialization_url): initialization_url})
                             f['fragments'].extend(representation_ms_info['fragments'])
-                        try:
-                            existing_format = next(
-                                fo for fo in formats
-                                if fo['format_id'] == representation_id)
-                        except StopIteration:
-                            full_info = formats_dict.get(representation_id, {}).copy()
-                            full_info.update(f)
-                            formats.append(full_info)
-                        else:
-                            existing_format.update(f)
+                        # According to [1, 5.3.5.2, Table 7, page 35] @id of Representation
+                        # is not necessarily unique within a Period thus formats with
+                        # the same `format_id` are quite possible. There are numerous examples
+                        # of such manifests (see https://github.com/rg3/youtube-dl/issues/15111,
+                        # https://github.com/rg3/youtube-dl/issues/13919)
+                        full_info = formats_dict.get(representation_id, {}).copy()
+                        full_info.update(f)
+                        formats.append(full_info)
                     else:
                         self.report_warning('Unknown MIME type %s in DASH manifest' % mime_type)
         return formats
 
     def _extract_ism_formats(self, ism_url, video_id, ism_id=None, note=None, errnote=None, fatal=True):
-        res = self._download_webpage_handle(
+        res = self._download_xml_handle(
             ism_url, video_id,
             note=note or 'Downloading ISM manifest',
             errnote=errnote or 'Failed to download ISM manifest',
             fatal=fatal)
         if res is False:
             return []
-        ism, urlh = res
+        ism_doc, urlh = res
 
-        return self._parse_ism_formats(
-            compat_etree_fromstring(ism.encode('utf-8')), urlh.geturl(), ism_id)
+        return self._parse_ism_formats(ism_doc, urlh.geturl(), ism_id)
 
     def _parse_ism_formats(self, ism_doc, ism_url, ism_id=None):
         """
@@ -2043,7 +2076,7 @@ class InfoExtractor(object):
             stream_timescale = int_or_none(stream.get('TimeScale')) or timescale
             stream_name = stream.get('Name')
             for track in stream.findall('QualityLevel'):
-                fourcc = track.get('FourCC')
+                fourcc = track.get('FourCC', 'AACL' if track.get('AudioTag') == '255' else None)
                 # TODO: add support for WVC1 and WMAP
                 if fourcc not in ('H264', 'AVC1', 'AACL'):
                     self.report_warning('%s is not a supported codec' % fourcc)
@@ -2117,8 +2150,8 @@ class InfoExtractor(object):
         return formats
 
     def _parse_html5_media_entries(self, base_url, webpage, video_id, m3u8_id=None, m3u8_entry_protocol='m3u8', mpd_id=None, preference=None):
-        def absolute_url(video_url):
-            return compat_urlparse.urljoin(base_url, video_url)
+        def absolute_url(item_url):
+            return urljoin(base_url, item_url)
 
         def parse_content_type(content_type):
             if not content_type:
@@ -2175,7 +2208,7 @@ class InfoExtractor(object):
             if src:
                 _, formats = _media_formats(src, media_type)
                 media_info['formats'].extend(formats)
-            media_info['thumbnail'] = media_attributes.get('poster')
+            media_info['thumbnail'] = absolute_url(media_attributes.get('poster'))
             if media_content:
                 for source_tag in re.findall(r'<source[^>]+>', media_content):
                     source_attributes = extract_attributes(source_tag)
@@ -2236,9 +2269,10 @@ class InfoExtractor(object):
     def _extract_wowza_formats(self, url, video_id, m3u8_entry_protocol='m3u8_native', skip_protocols=[]):
         query = compat_urlparse.urlparse(url).query
         url = re.sub(r'/(?:manifest|playlist|jwplayer)\.(?:m3u8|f4m|mpd|smil)', '', url)
-        url_base = self._search_regex(
-            r'(?:(?:https?|rtmp|rtsp):)?(//[^?]+)', url, 'format url')
-        http_base_url = '%s:%s' % ('http', url_base)
+        mobj = re.search(
+            r'(?:(?:http|rtmp|rtsp)(?P<s>s)?:)?(?P<url>//[^?]+)', url)
+        url_base = mobj.group('url')
+        http_base_url = '%s%s:%s' % ('http', mobj.group('s') or '', url_base)
         formats = []
 
         def manifest_url(manifest):
@@ -2338,7 +2372,10 @@ class InfoExtractor(object):
                 for track in tracks:
                     if not isinstance(track, dict):
                         continue
-                    if track.get('kind') != 'captions':
+                    track_kind = track.get('kind')
+                    if not track_kind or not isinstance(track_kind, compat_str):
+                        continue
+                    if track_kind.lower() not in ('captions', 'subtitles'):
                         continue
                     track_url = urljoin(base_url, track.get('file'))
                     if not track_url:
@@ -2392,7 +2429,7 @@ class InfoExtractor(object):
                 formats.extend(self._extract_m3u8_formats(
                     source_url, video_id, 'mp4', entry_protocol='m3u8_native',
                     m3u8_id=m3u8_id, fatal=False))
-            elif ext == 'mpd':
+            elif source_type == 'dash' or ext == 'mpd':
                 formats.extend(self._extract_mpd_formats(
                     source_url, video_id, mpd_id=mpd_id, fatal=False))
             elif ext == 'smil':