Raphaël G. Git Repositories - youtubedl/blob - youtube_dl/extractor/rtlnl.py

   1 # coding: utf-8
   2 from __future__ import unicode_literals
   3
   4 from .common import InfoExtractor
   5 from ..utils import (
   6     int_or_none,
   7     parse_duration,
   8 )
   9
  10
  11 class RtlNlIE(InfoExtractor):
  12     IE_NAME = 'rtl.nl'
  13     IE_DESC = 'rtl.nl and rtlxl.nl'
  14     _VALID_URL = r'''(?x)
  15         https?://(?:www\.)?
  16         (?:
  17             rtlxl\.nl/[^\#]*\#!/[^/]+/|
  18             rtl\.nl/(?:system/videoplayer/(?:[^/]+/)+(?:video_)?embed\.html\b.+?\buuid=|video/)
  19         )
  20         (?P<id>[0-9a-f-]+)'''
  21
  22     _TESTS = [{
  23         'url': 'http://www.rtlxl.nl/#!/rtl-nieuws-132237/82b1aad1-4a14-3d7b-b554-b0aed1b2c416',
  24         'md5': '473d1946c1fdd050b2c0161a4b13c373',
  25         'info_dict': {
  26             'id': '82b1aad1-4a14-3d7b-b554-b0aed1b2c416',
  27             'ext': 'mp4',
  28             'title': 'RTL Nieuws',
  29             'description': 'md5:d41d8cd98f00b204e9800998ecf8427e',
  30             'timestamp': 1461951000,
  31             'upload_date': '20160429',
  32             'duration': 1167.96,
  33         },
  34     }, {
  35         # best format avaialble a3t
  36         'url': 'http://www.rtl.nl/system/videoplayer/derden/rtlnieuws/video_embed.html#uuid=84ae5571-ac25-4225-ae0c-ef8d9efb2aed/autoplay=false',
  37         'md5': 'dea7474214af1271d91ef332fb8be7ea',
  38         'info_dict': {
  39             'id': '84ae5571-ac25-4225-ae0c-ef8d9efb2aed',
  40             'ext': 'mp4',
  41             'timestamp': 1424039400,
  42             'title': 'RTL Nieuws - Nieuwe beelden Kopenhagen: chaos direct na aanslag',
  43             'thumbnail': r're:^https?://screenshots\.rtl\.nl/(?:[^/]+/)*sz=[0-9]+x[0-9]+/uuid=84ae5571-ac25-4225-ae0c-ef8d9efb2aed$',
  44             'upload_date': '20150215',
  45             'description': 'Er zijn nieuwe beelden vrijgegeven die vlak na de aanslag in Kopenhagen zijn gemaakt. Op de video is goed te zien hoe omstanders zich bekommeren om één van de slachtoffers, terwijl de eerste agenten ter plaatse komen.',
  46         }
  47     }, {
  48         # empty synopsis and missing episodes (see https://github.com/rg3/youtube-dl/issues/6275)
  49         # best format available nettv
  50         'url': 'http://www.rtl.nl/system/videoplayer/derden/rtlnieuws/video_embed.html#uuid=f536aac0-1dc3-4314-920e-3bd1c5b3811a/autoplay=false',
  51         'info_dict': {
  52             'id': 'f536aac0-1dc3-4314-920e-3bd1c5b3811a',
  53             'ext': 'mp4',
  54             'title': 'RTL Nieuws - Meer beelden van overval juwelier',
  55             'thumbnail': r're:^https?://screenshots\.rtl\.nl/(?:[^/]+/)*sz=[0-9]+x[0-9]+/uuid=f536aac0-1dc3-4314-920e-3bd1c5b3811a$',
  56             'timestamp': 1437233400,
  57             'upload_date': '20150718',
  58             'duration': 30.474,
  59         },
  60         'params': {
  61             'skip_download': True,
  62         },
  63     }, {
  64         # encrypted m3u8 streams, georestricted
  65         'url': 'http://www.rtlxl.nl/#!/afl-2-257632/52a74543-c504-4cde-8aa8-ec66fe8d68a7',
  66         'only_matching': True,
  67     }, {
  68         'url': 'http://www.rtl.nl/system/videoplayer/derden/embed.html#!/uuid=bb0353b0-d6a4-1dad-90e9-18fe75b8d1f0',
  69         'only_matching': True,
  70     }, {
  71         'url': 'http://rtlxl.nl/?_ga=1.204735956.572365465.1466978370#!/rtl-nieuws-132237/3c487912-023b-49ac-903e-2c5d79f8410f',
  72         'only_matching': True,
  73     }, {
  74         'url': 'https://www.rtl.nl/video/c603c9c2-601d-4b5e-8175-64f1e942dc7d/',
  75         'only_matching': True,
  76     }]
  77
  78     def _real_extract(self, url):
  79         uuid = self._match_id(url)
  80         info = self._download_json(
  81             'http://www.rtl.nl/system/s4m/vfd/version=2/uuid=%s/fmt=adaptive/' % uuid,
  82             uuid)
  83
  84         material = info['material'][0]
  85         title = info['abstracts'][0]['name']
  86         subtitle = material.get('title')
  87         if subtitle:
  88             title += ' - %s' % subtitle
  89         description = material.get('synopsis')
  90
  91         meta = info.get('meta', {})
  92
  93         # m3u8 streams are encrypted and may not be handled properly by older ffmpeg/avconv.
  94         # To workaround this previously adaptive -> flash trick was used to obtain
  95         # unencrypted m3u8 streams (see https://github.com/rg3/youtube-dl/issues/4118)
  96         # and bypass georestrictions as well.
  97         # Currently, unencrypted m3u8 playlists are (intentionally?) invalid and therefore
  98         # unusable albeit can be fixed by simple string replacement (see
  99         # https://github.com/rg3/youtube-dl/pull/6337)
 100         # Since recent ffmpeg and avconv handle encrypted streams just fine encrypted
 101         # streams are used now.
 102         videopath = material['videopath']
 103         m3u8_url = meta.get('videohost', 'http://manifest.us.rtl.nl') + videopath
 104
 105         formats = self._extract_m3u8_formats(
 106             m3u8_url, uuid, 'mp4', m3u8_id='hls', fatal=False)
 107
 108         video_urlpart = videopath.split('/adaptive/')[1][:-5]
 109         PG_URL_TEMPLATE = 'http://pg.us.rtl.nl/rtlxl/network/%s/progressive/%s.mp4'
 110
 111         PG_FORMATS = (
 112             ('a2t', 512, 288),
 113             ('a3t', 704, 400),
 114             ('nettv', 1280, 720),
 115         )
 116
 117         def pg_format(format_id, width, height):
 118             return {
 119                 'url': PG_URL_TEMPLATE % (format_id, video_urlpart),
 120                 'format_id': 'pg-%s' % format_id,
 121                 'protocol': 'http',
 122                 'width': width,
 123                 'height': height,
 124             }
 125
 126         if not formats:
 127             formats = [pg_format(*pg_tuple) for pg_tuple in PG_FORMATS]
 128         else:
 129             pg_formats = []
 130             for format_id, width, height in PG_FORMATS:
 131                 try:
 132                     # Find hls format with the same width and height corresponding
 133                     # to progressive format and copy metadata from it.
 134                     f = next(f for f in formats if f.get('height') == height)
 135                     # hls formats may have invalid width
 136                     f['width'] = width
 137                     f_copy = f.copy()
 138                     f_copy.update(pg_format(format_id, width, height))
 139                     pg_formats.append(f_copy)
 140                 except StopIteration:
 141                     # Missing hls format does mean that no progressive format with
 142                     # such width and height exists either.
 143                     pass
 144             formats.extend(pg_formats)
 145         self._sort_formats(formats)
 146
 147         thumbnails = []
 148
 149         for p in ('poster_base_url', '"thumb_base_url"'):
 150             if not meta.get(p):
 151                 continue
 152
 153             thumbnails.append({
 154                 'url': self._proto_relative_url(meta[p] + uuid),
 155                 'width': int_or_none(self._search_regex(
 156                     r'/sz=([0-9]+)', meta[p], 'thumbnail width', fatal=False)),
 157                 'height': int_or_none(self._search_regex(
 158                     r'/sz=[0-9]+x([0-9]+)',
 159                     meta[p], 'thumbnail height', fatal=False))
 160             })
 161
 162         return {
 163             'id': uuid,
 164             'title': title,
 165             'formats': formats,
 166             'timestamp': material['original_date'],
 167             'description': description,
 168             'duration': parse_duration(material.get('duration')),
 169             'thumbnails': thumbnails,
 170         }