Fix extraction from youtube.

[youtubedl] / youtube_dl / extractor / ntvde.py

diff --git a/youtube_dl/extractor/ntvde.py b/youtube_dl/extractor/ntvde.py
index d2cfe096192f6a44fd0f9dad2ece1474a0d845ab..101a5374ccd9c780d6f34e8ff82ef67148e8a0cb 100644 (file)
--- a/youtube_dl/extractor/ntvde.py
+++ b/youtube_dl/extractor/ntvde.py
@@ -1,7 +1,10 @@
-# encoding: utf-8
+# coding: utf-8
 from __future__ import unicode_literals
 
+import re
+
 from .common import InfoExtractor
+from ..compat import compat_urlparse
 from ..utils import (
     int_or_none,
     js_to_json,
@@ -19,7 +22,7 @@ class NTVDeIE(InfoExtractor):
         'info_dict': {
             'id': '14438086',
             'ext': 'mp4',
-            'thumbnail': 're:^https?://.*\.jpg$',
+            'thumbnail': r're:^https?://.*\.jpg$',
             'title': 'Schnee und Glätte führen zu zahlreichen Unfällen und Staus',
             'alt_title': 'Winterchaos auf deutschen Straßen',
             'description': 'Schnee und Glätte sorgen deutschlandweit für einen chaotischen Start in die Woche: Auf den Straßen kommt es zu kilometerlangen Staus und Dutzenden Glätteunfällen. In Düsseldorf und München wirbelt der Schnee zudem den Flugplan durcheinander. Dutzende Flüge landen zu spät, einige fallen ganz aus.',
@@ -34,26 +37,32 @@ class NTVDeIE(InfoExtractor):
         webpage = self._download_webpage(url, video_id)
 
         info = self._parse_json(self._search_regex(
-            r'(?s)ntv.pageInfo.article =\s(\{.*?\});', webpage, 'info'),
+            r'(?s)ntv\.pageInfo\.article\s*=\s*(\{.*?\});', webpage, 'info'),
             video_id, transform_source=js_to_json)
         timestamp = int_or_none(info.get('publishedDateAsUnixTimeStamp'))
         vdata = self._parse_json(self._search_regex(
             r'(?s)\$\(\s*"\#player"\s*\)\s*\.data\(\s*"player",\s*(\{.*?\})\);',
-            webpage, 'player data'),
-            video_id, transform_source=js_to_json)
+            webpage, 'player data'), video_id,
+            transform_source=lambda s: js_to_json(re.sub(r'advertising:\s*{[^}]+},', '', s)))
         duration = parse_duration(vdata.get('duration'))
-        formats = [{
-            'format_id': 'flash',
-            'url': 'rtmp://fms.n-tv.de/' + vdata['video'],
-        }, {
-            'format_id': 'mobile',
-            'url': 'http://video.n-tv.de' + vdata['videoMp4'],
-            'tbr': 400,  # estimation
-        }]
-        m3u8_url = 'http://video.n-tv.de' + vdata['videoM3u8']
-        formats.extend(self._extract_m3u8_formats(
-            m3u8_url, video_id, ext='mp4',
-            entry_protocol='m3u8_native', preference=0))
+
+        formats = []
+        if vdata.get('video'):
+            formats.append({
+                'format_id': 'flash',
+                'url': 'rtmp://fms.n-tv.de/%s' % vdata['video'],
+            })
+        if vdata.get('videoMp4'):
+            formats.append({
+                'format_id': 'mobile',
+                'url': compat_urlparse.urljoin('http://video.n-tv.de', vdata['videoMp4']),
+                'tbr': 400,  # estimation
+            })
+        if vdata.get('videoM3u8'):
+            m3u8_url = compat_urlparse.urljoin('http://video.n-tv.de', vdata['videoM3u8'])
+            formats.extend(self._extract_m3u8_formats(
+                m3u8_url, video_id, ext='mp4', entry_protocol='m3u8_native',
+                preference=0, m3u8_id='hls', fatal=False))
         self._sort_formats(formats)
 
         return {