New upstream version 2017.12.31

[youtubedl] / youtube_dl / extractor / common.py

diff --git a/youtube_dl/extractor/common.py b/youtube_dl/extractor/common.py
index a692406931d4b63711609c7b8355635235849275..5b6a09c0b7f732ef5e9361775e42719af0deb8b7 100644 (file)
--- a/youtube_dl/extractor/common.py
+++ b/youtube_dl/extractor/common.py
@@ -29,7 +29,10 @@ from ..compat import (
     compat_urlparse,
     compat_xml_parse_error,
 )
     compat_urlparse,
     compat_xml_parse_error,
 )

-from ..downloader.f4m import remove_encrypted_media
+from ..downloader.f4m import (
+    get_base_url,
+    remove_encrypted_media,
+)

 from ..utils import (
     NO_DEFAULT,
     age_restricted,
 from ..utils import (
     NO_DEFAULT,
     age_restricted,


@@ -298,8 +301,9 @@ 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.
 
     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
 
 
     _type "multi_video" indicates that there are multiple videos that


@@ -491,6 +495,16 @@ class InfoExtractor(object):
                 self.to_screen('%s' % (note,))
             else:
                 self.to_screen('%s: %s' % (video_id, note))
                 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)
         if isinstance(url_or_request, compat_urllib_request.Request):
             url_or_request = update_Request(
                 url_or_request, data=data, headers=headers, query=query)


@@ -520,15 +534,6 @@ class InfoExtractor(object):
         if isinstance(url_or_request, (compat_str, str)):
             url_or_request = url_or_request.partition('#')[0]
 
         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
         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


@@ -589,19 +594,11 @@ 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):
         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):
             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
             if len(basen) > 240:
                 h = '___' + hashlib.md5(basen.encode('utf-8')).hexdigest()
                 basen = basen[:240 - len(h)] + h


@@ -1239,11 +1236,8 @@ class InfoExtractor(object):
         media_nodes = remove_encrypted_media(media_nodes)
         if not media_nodes:
             return formats
         media_nodes = remove_encrypted_media(media_nodes)
         if not media_nodes:
             return formats

-        base_url = xpath_text(
-            manifest, ['{http://ns.adobe.com/f4m/1.0}baseURL', '{http://ns.adobe.com/f4m/2.0}baseURL'],
-            'base URL', default=None)
-        if base_url:
-            base_url = base_url.strip()
+
+        manifest_base_url = get_base_url(manifest)

 
         bootstrap_info = xpath_element(
             manifest, ['{http://ns.adobe.com/f4m/1.0}bootstrapInfo', '{http://ns.adobe.com/f4m/2.0}bootstrapInfo'],
 
         bootstrap_info = xpath_element(
             manifest, ['{http://ns.adobe.com/f4m/1.0}bootstrapInfo', '{http://ns.adobe.com/f4m/2.0}bootstrapInfo'],


@@ -1275,7 +1269,7 @@ class InfoExtractor(object):
                     continue
                 manifest_url = (
                     media_url if media_url.startswith('http://') or media_url.startswith('https://')
                     continue
                 manifest_url = (
                     media_url if media_url.startswith('http://') or media_url.startswith('https://')

-                    else ((base_url or '/'.join(manifest_url.split('/')[:-1])) + '/' + media_url))
+                    else ((manifest_base_url or '/'.join(manifest_url.split('/')[:-1])) + '/' + media_url))

                 # If media_url is itself a f4m manifest do the recursive extraction
                 # since bitrates in parent manifest (this one) and media_url manifest
                 # may differ leading to inability to resolve the format by requested
                 # If media_url is itself a f4m manifest do the recursive extraction
                 # since bitrates in parent manifest (this one) and media_url manifest
                 # may differ leading to inability to resolve the format by requested


@@ -1310,6 +1304,7 @@ class InfoExtractor(object):
                 'url': manifest_url,
                 'manifest_url': manifest_url,
                 'ext': 'flv' if bootstrap_info is not None else None,
                 'url': manifest_url,
                 'manifest_url': manifest_url,
                 'ext': 'flv' if bootstrap_info is not None else None,

+                'protocol': 'f4m',

                 'tbr': tbr,
                 'width': width,
                 'height': height,
                 'tbr': tbr,
                 'width': width,
                 'height': height,


@@ -1355,6 +1350,9 @@ class InfoExtractor(object):
         if '#EXT-X-FAXS-CM:' in m3u8_doc:  # Adobe Flash Access
             return []
 
         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: (
         formats = []
 
         format_url = lambda u: (


@@ -1401,7 +1399,7 @@ class InfoExtractor(object):
             media_url = media.get('URI')
             if media_url:
                 format_id = []
             media_url = media.get('URI')
             if media_url:
                 format_id = []

-                for v in (group_id, name):
+                for v in (m3u8_id, group_id, name):

                     if v:
                         format_id.append(v)
                 f = {
                     if v:
                         format_id.append(v)
                 f = {


@@ -1882,6 +1880,7 @@ class InfoExtractor(object):
                             'language': lang if lang not in ('mul', 'und', 'zxx', 'mis') else None,
                             'format_note': 'DASH %s' % content_type,
                             'filesize': filesize,
                             '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)
                         }
                         f.update(parse_codecs(representation_attrib.get('codecs')))
                         representation_ms_info = extract_multisegment_info(representation, adaption_set_ms_info)


@@ -1979,6 +1978,22 @@ class InfoExtractor(object):
                                     })
                                     segment_index += 1
                             representation_ms_info['fragments'] = fragments
                                     })
                                     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:
                         # NB: MPD manifest may contain direct URLs to unfragmented media.
                         # No fragments key is present in this case.
                         if 'fragments' in representation_ms_info:


@@ -1993,16 +2008,14 @@ class InfoExtractor(object):
                                     f['url'] = initialization_url
                                 f['fragments'].append({location_key(initialization_url): initialization_url})
                             f['fragments'].extend(representation_ms_info['fragments'])
                                     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
                     else:
                         self.report_warning('Unknown MIME type %s in DASH manifest' % mime_type)
         return formats


@@ -2042,7 +2055,7 @@ class InfoExtractor(object):
             stream_timescale = int_or_none(stream.get('TimeScale')) or timescale
             stream_name = stream.get('Name')
             for track in stream.findall('QualityLevel'):
             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)
                 # TODO: add support for WVC1 and WMAP
                 if fourcc not in ('H264', 'AVC1', 'AACL'):
                     self.report_warning('%s is not a supported codec' % fourcc)


@@ -2233,27 +2246,35 @@ class InfoExtractor(object):
         return formats
 
     def _extract_wowza_formats(self, url, video_id, m3u8_entry_protocol='m3u8_native', skip_protocols=[]):
         return formats
 
     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)
         formats = []
         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)
         formats = []

+
+        def manifest_url(manifest):
+            m_url = '%s/%s' % (http_base_url, manifest)
+            if query:
+                m_url += '?%s' % query
+            return m_url
+

         if 'm3u8' not in skip_protocols:
             formats.extend(self._extract_m3u8_formats(
         if 'm3u8' not in skip_protocols:
             formats.extend(self._extract_m3u8_formats(

-                http_base_url + '/playlist.m3u8', video_id, 'mp4',
+                manifest_url('playlist.m3u8'), video_id, 'mp4',

                 m3u8_entry_protocol, m3u8_id='hls', fatal=False))
         if 'f4m' not in skip_protocols:
             formats.extend(self._extract_f4m_formats(
                 m3u8_entry_protocol, m3u8_id='hls', fatal=False))
         if 'f4m' not in skip_protocols:
             formats.extend(self._extract_f4m_formats(

-                http_base_url + '/manifest.f4m',
+                manifest_url('manifest.f4m'),

                 video_id, f4m_id='hds', fatal=False))
         if 'dash' not in skip_protocols:
             formats.extend(self._extract_mpd_formats(
                 video_id, f4m_id='hds', fatal=False))
         if 'dash' not in skip_protocols:
             formats.extend(self._extract_mpd_formats(

-                http_base_url + '/manifest.mpd',
+                manifest_url('manifest.mpd'),

                 video_id, mpd_id='dash', fatal=False))
         if re.search(r'(?:/smil:|\.smil)', url_base):
             if 'smil' not in skip_protocols:
                 rtmp_formats = self._extract_smil_formats(
                 video_id, mpd_id='dash', fatal=False))
         if re.search(r'(?:/smil:|\.smil)', url_base):
             if 'smil' not in skip_protocols:
                 rtmp_formats = self._extract_smil_formats(

-                    http_base_url + '/jwplayer.smil',
+                    manifest_url('jwplayer.smil'),

                     video_id, fatal=False)
                 for rtmp_format in rtmp_formats:
                     rtsp_format = rtmp_format.copy()
                     video_id, fatal=False)
                 for rtmp_format in rtmp_formats:
                     rtsp_format = rtmp_format.copy()