aboutsummaryrefslogtreecommitdiff
diff options
context:
space:
mode:
-rw-r--r--youtube_dl/extractor/__init__.py1
-rw-r--r--youtube_dl/extractor/eagleplatform.py99
2 files changed, 100 insertions, 0 deletions
diff --git a/youtube_dl/extractor/__init__.py b/youtube_dl/extractor/__init__.py
index f72d06fc4..b489d5770 100644
--- a/youtube_dl/extractor/__init__.py
+++ b/youtube_dl/extractor/__init__.py
@@ -116,6 +116,7 @@ from .defense import DefenseGouvFrIE
from .discovery import DiscoveryIE
from .divxstage import DivxStageIE
from .dropbox import DropboxIE
+from .eagleplatform import EaglePlatformIE
from .ebaumsworld import EbaumsWorldIE
from .echomsk import EchoMskIE
from .ehow import EHowIE
diff --git a/youtube_dl/extractor/eagleplatform.py b/youtube_dl/extractor/eagleplatform.py
new file mode 100644
index 000000000..654f746fe
--- /dev/null
+++ b/youtube_dl/extractor/eagleplatform.py
@@ -0,0 +1,99 @@
+# coding: utf-8
+from __future__ import unicode_literals
+
+import re
+
+from .common import InfoExtractor
+from ..utils import (
+ ExtractorError,
+ int_or_none,
+)
+
+
+class EaglePlatformIE(InfoExtractor):
+ _VALID_URL = r'''(?x)
+ (?:
+ eagleplatform:(?P<custom_host>[^/]+):|
+ https?://(?P<host>.+?\.media\.eagleplatform\.com)/index/player\?.*\brecord_id=
+ )
+ (?P<id>\d+)
+ '''
+ _TESTS = [{
+ # http://lenta.ru/news/2015/03/06/navalny/
+ 'url': 'http://lentaru.media.eagleplatform.com/index/player?player=new&record_id=227304&player_template_id=5201',
+ 'md5': '0b7994faa2bd5c0f69a3db6db28d078d',
+ 'info_dict': {
+ 'id': '227304',
+ 'ext': 'mp4',
+ 'title': 'Навальный вышел на свободу',
+ 'description': 'md5:d97861ac9ae77377f3f20eaf9d04b4f5',
+ 'thumbnail': 're:^https?://.*\.jpg$',
+ 'duration': 87,
+ 'view_count': int,
+ 'age_limit': 0,
+ },
+ }, {
+ # http://muz-tv.ru/play/7129/
+ # http://media.clipyou.ru/index/player?record_id=12820&width=730&height=415&autoplay=true
+ 'url': 'eagleplatform:media.clipyou.ru:12820',
+ 'md5': '6c2ebeab03b739597ce8d86339d5a905',
+ 'info_dict': {
+ 'id': '12820',
+ 'ext': 'mp4',
+ 'title': "'O Sole Mio",
+ 'thumbnail': 're:^https?://.*\.jpg$',
+ 'duration': 216,
+ 'view_count': int,
+ },
+ }]
+
+ def _handle_error(self, response):
+ print(response['status'])
+ status = int_or_none(response.get('status', 200))
+ if status != 200:
+ raise ExtractorError(' '.join(response['errors']), expected=True)
+
+ def _download_json(self, url_or_request, video_id, note='Downloading JSON metadata'):
+ response = super(EaglePlatformIE, self)._download_json(url_or_request, video_id, note)
+ self._handle_error(response)
+ return response
+
+ def _real_extract(self, url):
+ mobj = re.match(self._VALID_URL, url)
+ host, video_id = mobj.group('custom_host') or mobj.group('host'), mobj.group('id')
+
+ player_data = self._download_json(
+ 'http://%s/api/player_data?id=%s' % (host, video_id), video_id)
+
+ media = player_data['data']['playlist']['viewports'][0]['medialist'][0]
+
+ title = media['title']
+ description = media.get('description')
+ thumbnail = media.get('snapshot')
+ duration = int_or_none(media.get('duration'))
+ view_count = int_or_none(media.get('views'))
+
+ age_restriction = media.get('age_restriction')
+ age_limit = None
+ if age_restriction:
+ age_limit = 0 if age_restriction == 'allow_all' else 18
+
+ m3u8_data = self._download_json(
+ media['sources']['secure_m3u8']['auto'],
+ video_id, 'Downloading m3u8 JSON')
+
+ formats = self._extract_m3u8_formats(
+ m3u8_data['data'][0], video_id,
+ 'mp4', entry_protocol='m3u8_native')
+ self._sort_formats(formats)
+
+ return {
+ 'id': video_id,
+ 'title': title,
+ 'description': description,
+ 'thumbnail': thumbnail,
+ 'duration': duration,
+ 'view_count': view_count,
+ 'age_limit': age_limit,
+ 'formats': formats,
+ }