Merge remote-tracking branch 'upstream/master'

youtube_dl/extractor/__init__.py

@@ -389,7 +389,8 @@ from .npo import (
     NPOLiveIE,
     NPORadioIE,
     NPORadioFragmentIE,
-    TegenlichtVproIE,
+    VPROIE,
+    WNLIE
 )
 from .nrk import (
     NRKIE,

youtube_dl/extractor/common.py

@@ -27,7 +27,9 @@ from ..utils import (
     bug_reports_message,
     clean_html,
     compiled_regex_type,
+    determine_ext,
     ExtractorError,
+    fix_xml_ampersands,
     float_or_none,
     int_or_none,
     RegexNotFoundError,
@@ -837,7 +839,10 @@ class InfoExtractor(object):
     def _extract_f4m_formats(self, manifest_url, video_id, preference=None, f4m_id=None):
         manifest = self._download_xml(
             manifest_url, video_id, 'Downloading f4m manifest',
-            'Unable to download f4m manifest')
+            'Unable to download f4m manifest',
+            # Some manifests may be malformed, e.g. prosiebensat1 generated manifests
+            # (see https://github.com/rg3/youtube-dl/issues/6215#issuecomment-121704244)
+            transform_source=lambda s: fix_xml_ampersands(s).strip())
 
         formats = []
         manifest_version = '1.0'
@@ -847,8 +852,19 @@ class InfoExtractor(object):
             media_nodes = manifest.findall('{http://ns.adobe.com/f4m/2.0}media')
         for i, media_el in enumerate(media_nodes):
             if manifest_version == '2.0':
-                manifest_url = ('/'.join(manifest_url.split('/')[:-1]) + '/' +
+                media_url = media_el.attrib.get('href') or media_el.attrib.get('url')
-                                (media_el.attrib.get('href') or media_el.attrib.get('url')))
+                if not media_url:
+                    continue
+                manifest_url = (
+                    media_url if media_url.startswith('http://') or media_url.startswith('https://')
+                    else ('/'.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
+                # bitrate in f4m downloader
+                if determine_ext(manifest_url) == 'f4m':
+                    formats.extend(self._extract_f4m_formats(manifest_url, video_id, preference, f4m_id))
+                    continue
             tbr = int_or_none(media_el.attrib.get('bitrate'))
             formats.append({
                 'format_id': '-'.join(filter(None, [f4m_id, compat_str(i if tbr is None else tbr)])),

youtube_dl/extractor/npo.py

@@ -1,5 +1,7 @@
 from __future__ import unicode_literals
 
+import re
+
 from .common import InfoExtractor
 from ..utils import (
     fix_xml_ampersands,
@@ -7,7 +9,6 @@ from ..utils import (
     qualities,
     strip_jsonp,
     unified_strdate,
-    url_basename,
 )
 
 
@@ -37,8 +38,21 @@ class NPOBaseIE(InfoExtractor):
 
 
 class NPOIE(NPOBaseIE):
-    IE_NAME = 'npo.nl'
+    IE_NAME = 'npo'
-    _VALID_URL = r'https?://(?:www\.)?npo\.nl/(?!live|radio)[^/]+/[^/]+/(?P<id>[^/?]+)'
+    IE_DESC = 'npo.nl and ntr.nl'
+    _VALID_URL = r'''(?x)
+                    (?:
+                        npo:|
+                        https?://
+                            (?:www\.)?
+                            (?:
+                                npo\.nl/(?!live|radio)(?:[^/]+/){2}|
+                                ntr\.nl/(?:[^/]+/){2,}|
+                                omroepwnl\.nl/video/fragment/[^/]+__
+                            )
+                        )
+                        (?P<id>[^/?#]+)
+                '''
 
     _TESTS = [
         {
@@ -58,7 +72,7 @@ class NPOIE(NPOBaseIE):
             'info_dict': {
                 'id': 'VARA_101191800',
                 'ext': 'm4v',
-                'title': 'De Mega Mike & Mega Thomas show',
+                'title': 'De Mega Mike & Mega Thomas show: The best of.',
                 'description': 'md5:3b74c97fc9d6901d5a665aac0e5400f4',
                 'upload_date': '20090227',
                 'duration': 2400,
@@ -70,8 +84,8 @@ class NPOIE(NPOBaseIE):
             'info_dict': {
                 'id': 'VPWON_1169289',
                 'ext': 'm4v',
-                'title': 'Tegenlicht',
+                'title': 'Tegenlicht: De toekomst komt uit Afrika',
-                'description': 'md5:d6476bceb17a8c103c76c3b708f05dd1',
+                'description': 'md5:52cf4eefbc96fffcbdc06d024147abea',
                 'upload_date': '20130225',
                 'duration': 3000,
             },
@@ -100,6 +114,30 @@ class NPOIE(NPOBaseIE):
                 'title': 'Hoe gaat Europa verder na Parijs?',
             },
         },
+        {
+            'url': 'http://www.ntr.nl/Aap-Poot-Pies/27/detail/Aap-poot-pies/VPWON_1233944#content',
+            'md5': '01c6a2841675995da1f0cf776f03a9c3',
+            'info_dict': {
+                'id': 'VPWON_1233944',
+                'ext': 'm4v',
+                'title': 'Aap, poot, pies',
+                'description': 'md5:c9c8005d1869ae65b858e82c01a91fde',
+                'upload_date': '20150508',
+                'duration': 599,
+            },
+        },
+        {
+            'url': 'http://www.omroepwnl.nl/video/fragment/vandaag-de-dag-verkiezingen__POMS_WNL_853698',
+            'md5': 'd30cd8417b8b9bca1fdff27428860d08',
+            'info_dict': {
+                'id': 'POW_00996502',
+                'ext': 'm4v',
+                'title': '''"Dit is wel een 'landslide'..."''',
+                'description': 'md5:f8d66d537dfb641380226e31ca57b8e8',
+                'upload_date': '20150508',
+                'duration': 462,
+            },
+        }
     ]
 
     def _real_extract(self, url):
@@ -114,6 +152,18 @@ class NPOIE(NPOBaseIE):
             transform_source=strip_jsonp,
         )
 
+        # For some videos actual video id (prid) is different (e.g. for
+        # http://www.omroepwnl.nl/video/fragment/vandaag-de-dag-verkiezingen__POMS_WNL_853698
+        # video id is POMS_WNL_853698 but prid is POW_00996502)
+        video_id = metadata.get('prid') or video_id
+
+        # titel is too generic in some cases so utilize aflevering_titel as well
+        # when available (e.g. http://tegenlicht.vpro.nl/afleveringen/2014-2015/access-to-africa.html)
+        title = metadata['titel']
+        sub_title = metadata.get('aflevering_titel')
+        if sub_title and sub_title != title:
+            title += ': %s' % sub_title
+
         token = self._get_token(video_id)
 
         formats = []
@@ -186,8 +236,8 @@ class NPOIE(NPOBaseIE):
 
         return {
             'id': video_id,
-            'title': metadata['titel'],
+            'title': title,
-            'description': metadata['info'],
+            'description': metadata.get('info'),
             'thumbnail': metadata.get('images', [{'url': None}])[-1]['url'],
             'upload_date': unified_strdate(metadata.get('gidsdatum')),
             'duration': parse_duration(metadata.get('tijdsduur')),
@@ -356,9 +406,8 @@ class NPORadioFragmentIE(InfoExtractor):
         }
 
 
-class TegenlichtVproIE(NPOIE):
+class VPROIE(NPOIE):
-    IE_NAME = 'tegenlicht.vpro.nl'
+    _VALID_URL = r'https?://(?:www\.)?(?:tegenlicht\.)?vpro\.nl/(?:[^/]+/){2,}(?P<id>[^/]+)\.html'
-    _VALID_URL = r'https?://tegenlicht\.vpro\.nl/afleveringen/.*?'
 
     _TESTS = [
         {
@@ -367,17 +416,72 @@ class TegenlichtVproIE(NPOIE):
             'info_dict': {
                 'id': 'VPWON_1169289',
                 'ext': 'm4v',
-                'title': 'Tegenlicht',
+                'title': 'De toekomst komt uit Afrika',
-                'description': 'md5:d6476bceb17a8c103c76c3b708f05dd1',
+                'description': 'md5:52cf4eefbc96fffcbdc06d024147abea',
                 'upload_date': '20130225',
             },
         },
+        {
+            'url': 'http://www.vpro.nl/programmas/2doc/2015/sergio-herman.html',
+            'info_dict': {
+                'id': 'sergio-herman',
+                'title': 'Sergio Herman: Fucking perfect',
+            },
+            'playlist_count': 2,
+        },
+        {
+            # playlist with youtube embed
+            'url': 'http://www.vpro.nl/programmas/2doc/2015/education-education.html',
+            'info_dict': {
+                'id': 'education-education',
+                'title': '2Doc',
+            },
+            'playlist_count': 2,
+        }
     ]
 
     def _real_extract(self, url):
-        name = url_basename(url)
+        playlist_id = self._match_id(url)
-        webpage = self._download_webpage(url, name)
+
-        urn = self._html_search_meta('mediaurn', webpage)
+        webpage = self._download_webpage(url, playlist_id)
-        info_page = self._download_json(
+
-            'http://rs.vpro.nl/v2/api/media/%s.json' % urn, name)
+        entries = [
-        return self._get_info(info_page['mid'])
+            self.url_result('npo:%s' % video_id if not video_id.startswith('http') else video_id)
+            for video_id in re.findall(r'data-media-id="([^"]+)"', webpage)
+        ]
+
+        playlist_title = self._search_regex(
+            r'<title>\s*([^>]+?)\s*-\s*Teledoc\s*-\s*VPRO\s*</title>',
+            webpage, 'playlist title', default=None) or self._og_search_title(webpage)
+
+        return self.playlist_result(entries, playlist_id, playlist_title)
+
+
+class WNLIE(InfoExtractor):
+    _VALID_URL = r'https?://(?:www\.)?omroepwnl\.nl/video/detail/(?P<id>[^/]+)__\d+'
+
+    _TEST = {
+        'url': 'http://www.omroepwnl.nl/video/detail/vandaag-de-dag-6-mei__060515',
+        'info_dict': {
+            'id': 'vandaag-de-dag-6-mei',
+            'title': 'Vandaag de Dag 6 mei',
+        },
+        'playlist_count': 4,
+    }
+
+    def _real_extract(self, url):
+        playlist_id = self._match_id(url)
+
+        webpage = self._download_webpage(url, playlist_id)
+
+        entries = [
+            self.url_result('npo:%s' % video_id, 'NPO')
+            for video_id, part in re.findall(
+                r'<a[^>]+href="([^"]+)"[^>]+class="js-mid"[^>]*>(Deel \d+)', webpage)
+        ]
+
+        playlist_title = self._html_search_regex(
+            r'(?s)<h1[^>]+class="subject"[^>]*>(.+?)</h1>',
+            webpage, 'playlist title')
+
+        return self.playlist_result(entries, playlist_id, playlist_title)

youtube_dl/extractor/prosiebensat1.py

@@ -9,8 +9,9 @@ from ..compat import (
     compat_urllib_parse,
 )
 from ..utils import (
-    unified_strdate,
+    determine_ext,
     int_or_none,
+    unified_strdate,
 )
 
 
@@ -21,6 +22,11 @@ class ProSiebenSat1IE(InfoExtractor):
 
     _TESTS = [
         {
+            # Tests changes introduced in https://github.com/rg3/youtube-dl/pull/6242
+            # in response to fixing https://github.com/rg3/youtube-dl/issues/6215:
+            # - malformed f4m manifest support
+            # - proper handling of URLs starting with `https?://` in 2.0 manifests
+            # - recursive child f4m manifests extraction
             'url': 'http://www.prosieben.de/tv/circus-halligalli/videos/218-staffel-2-episode-18-jahresrueckblick-ganze-folge',
             'info_dict': {
                 'id': '2104602',
@@ -208,7 +214,7 @@ class ProSiebenSat1IE(InfoExtractor):
         clip_id = self._html_search_regex(self._CLIPID_REGEXES, webpage, 'clip id')
 
         access_token = 'prosieben'
-        client_name = 'kolibri-1.12.6'
+        client_name = 'kolibri-2.0.19-splec4'
         client_location = url
 
         videos_api_url = 'http://vas.sim-technik.de/vas/live/v2/videos?%s' % compat_urllib_parse.urlencode({
@@ -275,8 +281,9 @@ class ProSiebenSat1IE(InfoExtractor):
 
         for source in urls_sources:
             protocol = source['protocol']
+            source_url = source['url']
             if protocol == 'rtmp' or protocol == 'rtmpe':
-                mobj = re.search(r'^(?P<url>rtmpe?://[^/]+)/(?P<path>.+)$', source['url'])
+                mobj = re.search(r'^(?P<url>rtmpe?://[^/]+)/(?P<path>.+)$', source_url)
                 if not mobj:
                     continue
                 path = mobj.group('path')
@@ -293,9 +300,11 @@ class ProSiebenSat1IE(InfoExtractor):
                     'ext': 'mp4',
                     'format_id': '%s_%s' % (source['cdn'], source['bitrate']),
                 })
+            elif 'f4mgenerator' in source_url or determine_ext(source_url) == 'f4m':
+                formats.extend(self._extract_f4m_formats(source_url, clip_id))
             else:
                 formats.append({
-                    'url': source['url'],
+                    'url': source_url,
                     'vbr': fix_bitrate(source['bitrate']),
                 })