4oD extractor, with a testcase

mostly unusable, since FlashAccess DRM is not implemented, here or in ffmpeg.

youtube-dl can at least download subtitles and metadata, and make the stream URL available through -j for external decryption.

youtube_dl/extractor/__init__.py

@@ -63,6 +63,7 @@ from .cbsnews import CBSNewsIE
 from .cbssports import CBSSportsIE
 from .ccc import CCCIE
 from .ceskatelevize import CeskaTelevizeIE
+from .channel4 import Channel4IE
 from .channel9 import Channel9IE
 from .chilloutzone import ChilloutzoneIE
 from .chirbit import (

youtube_dl/extractor/channel4.py

@@ -0,0 +1,78 @@
+from __future__ import unicode_literals
+
+import json
+
+from .common import InfoExtractor
+from ..compat import (
+    compat_urllib_request,
+    compat_urlparse,
+)
+from ..utils import (
+    unified_strdate,
+    ExtractorError
+)
+
+class Channel4IE(InfoExtractor):
+    _VALID_URL = r'http://(?:www\.)?channel4\.com/programmes/(?P<pid>.*?)/on-demand/(?P<id>.*)'
+
+    _TESTS = [{
+        'url': 'http://www.channel4.com/programmes/black-mirror/on-demand/49114-002',
+        'info_dict': {
+            'id': '49114-002',
+            '_programme_title': "Black Mirror",
+            'title': "15 Million Merits",
+            'description': "In the near future, everyone is confined to a life of strange physical drudgery. The only way to escape is to enter the 'Hot Shot' talent show and pray you can impress the judges.",
+            'duration': 222780,
+        },
+        'params': {
+            # unimplemented DRM
+            'skip_download': True,
+        }
+    }]
+
+    def _real_extract(self, url):
+        video_id = self._match_id(url)
+        webpage = self._download_webpage(url, video_id)
+
+        video_data = json.loads(self._search_regex(
+            r'onDemand.selectedEpisode = (?P<json>{.+?});\n',
+            webpage, 'video data json'))
+
+        thumbnails = [{
+            'url': video_data["pictureComponent"]["url"]
+        }]
+
+        request_id = video_data["requestId"]
+        # XXX: the Flash player also puts the Unix timestamp in the query string. the download works without it just fine, though
+        asset_url = 'http://ais.channel4.com/asset/%s' % (request_id)
+        stream_info = self._download_xml(asset_url, video_id)
+
+        service_report = stream_info.find('./serviceReport')
+        if service_report.attrib.get('returnCode') != "200":
+            raise ExtractorError(service_report.findtext('./description'), expected=True)
+
+        subtitles = stream_info.findtext('./assetInfo/subtitlesFileUri')
+        if subtitles:
+            subtitles = {
+                'en': [{
+                    'ext': 'sami',
+                    'url': compat_urlparse.urljoin(asset_url, subtitles),
+                }]
+            }
+
+        formats = self._extract_f4m_formats(stream_info.findtext('./assetInfo/uriData/streamUri'), video_id)
+
+        return {
+            'id': video_id,
+            '_programme_title': stream_info.findtext('./assetInfo/brandTitle'),
+            'title': stream_info.findtext('./assetInfo/episodeTitle'),
+            'upload_date': unified_strdate(video_data['txDate'] + ' ' + video_data.get('txTime', '')),
+            'description': video_data.get('synopsis'),
+            'thumbnails': thumbnails,
+            'formats': formats,
+            'subtitles': subtitles,
+            'duration': video_data["assetDuration"] * 60,
+            '_drm_token': stream_info.findtext('./assetInfo/uriData/token'),
+            '_programme_series': video_data.get("seriesNumber"),
+            '_programme_episode': video_data.get("episodeNumber"),
+        }