aboutsummaryrefslogtreecommitdiff
path: root/youtube_dl
diff options
context:
space:
mode:
Diffstat (limited to 'youtube_dl')
-rw-r--r--youtube_dl/extractor/rtp.py83
1 files changed, 30 insertions, 53 deletions
diff --git a/youtube_dl/extractor/rtp.py b/youtube_dl/extractor/rtp.py
index 533ee27cb..02986f442 100644
--- a/youtube_dl/extractor/rtp.py
+++ b/youtube_dl/extractor/rtp.py
@@ -1,9 +1,11 @@
# coding: utf-8
from __future__ import unicode_literals
-import re
-
from .common import InfoExtractor
+from ..utils import (
+ determine_ext,
+ js_to_json,
+)
class RTPIE(InfoExtractor):
@@ -18,10 +20,6 @@ class RTPIE(InfoExtractor):
'description': 'As paixões musicais de António Cartaxo e António Macedo',
'thumbnail': r're:^https?://.*\.jpg',
},
- 'params': {
- # rtmp download
- 'skip_download': True,
- },
}, {
'url': 'http://www.rtp.pt/play/p831/a-quimica-das-coisas',
'only_matching': True,
@@ -33,57 +31,36 @@ class RTPIE(InfoExtractor):
webpage = self._download_webpage(url, video_id)
title = self._html_search_meta(
'twitter:title', webpage, display_name='title', fatal=True)
- description = self._html_search_meta('description', webpage)
- thumbnail = self._og_search_thumbnail(webpage)
-
- player_config = self._search_regex(
- r'(?s)RTPPLAY\.player\.newPlayer\(\s*(\{.*?\})\s*\)', webpage, 'player config')
- config = self._parse_json(player_config, video_id)
-
- path, ext = config.get('file').rsplit('.', 1)
- formats = [{
- 'format_id': 'rtmp',
- 'ext': ext,
- 'vcodec': config.get('type') == 'audio' and 'none' or None,
- 'preference': -2,
- 'url': 'rtmp://{streamer:s}/{application:s}'.format(**config),
- 'app': config.get('application'),
- 'play_path': '{ext:s}:{path:s}'.format(ext=ext, path=path),
- 'page_url': url,
- 'rtmp_live': config.get('live', False),
- 'player_url': 'http://programas.rtp.pt/play/player.swf?v3',
- 'rtmp_real_time': True,
- }]
-
- # Construct regular HTTP download URLs
- replacements = {
- 'audio': {
- 'format_id': 'mp3',
- 'pattern': r'^nas2\.share/wavrss/',
- 'repl': 'http://rsspod.rtp.pt/podcasts/',
- 'vcodec': 'none',
- },
- 'video': {
- 'format_id': 'mp4_h264',
- 'pattern': r'^nas2\.share/h264/',
- 'repl': 'http://rsspod.rtp.pt/videocasts/',
- 'vcodec': 'h264',
- },
- }
- r = replacements[config['type']]
- if re.match(r['pattern'], config['file']) is not None:
- formats.append({
- 'format_id': r['format_id'],
- 'url': re.sub(r['pattern'], r['repl'], config['file']),
- 'vcodec': r['vcodec'],
- })
- self._sort_formats(formats)
+ config = self._parse_json(self._search_regex(
+ r'(?s)RTPPlayer\(({.+?})\);', webpage,
+ 'player config'), video_id, js_to_json)
+ file_url = config['file']
+ ext = determine_ext(file_url)
+ if ext == 'm3u8':
+ file_key = config.get('fileKey')
+ formats = self._extract_m3u8_formats(
+ file_url, video_id, 'mp4', 'm3u8_native',
+ m3u8_id='hls', fatal=file_key)
+ if file_key:
+ formats.append({
+ 'url': 'https://cdn-ondemand.rtp.pt' + file_key,
+ 'preference': 1,
+ })
+ self._sort_formats(formats)
+ else:
+ formats = [{
+ 'url': file_url,
+ 'ext': ext,
+ }]
+ if config.get('mediaType') == 'audio':
+ for f in formats:
+ f['vcodec'] = 'none'
return {
'id': video_id,
'title': title,
'formats': formats,
- 'description': description,
- 'thumbnail': thumbnail,
+ 'description': self._html_search_meta(['description', 'twitter:description'], webpage),
+ 'thumbnail': config.get('poster') or self._og_search_thumbnail(webpage),
}