1
0
Fork 0
mirror of https://github.com/blackjack4494/yt-dlc.git synced 2024-12-22 15:57:23 +00:00

[nbc:nbcnews] improve extraction(fixes #6922)

- extract more metadata and formats
- relax regex
This commit is contained in:
remitamine 2016-03-17 16:07:36 +01:00
parent 9f02ff537c
commit 574b2a7393

View file

@ -3,13 +3,16 @@ from __future__ import unicode_literals
import re
from .common import InfoExtractor
from ..compat import compat_HTTPError
from .theplatform import ThePlatformIE
from ..utils import (
ExtractorError,
find_xpath_attr,
lowercase_escape,
smuggle_url,
unescapeHTML,
update_url_query,
int_or_none,
HEADRequest,
parse_iso8601,
)
@ -131,10 +134,10 @@ class NBCSportsIE(InfoExtractor):
NBCSportsVPlayerIE._extract_url(webpage), 'NBCSportsVPlayer')
class NBCNewsIE(InfoExtractor):
class NBCNewsIE(ThePlatformIE):
_VALID_URL = r'''(?x)https?://(?:www\.)?nbcnews\.com/
(?:video/.+?/(?P<id>\d+)|
(?:watch|feature|nightly-news)/[^/]+/(?P<title>.+))
([^/]+/)*(?P<display_id>[^/?]+))
'''
_TESTS = [
@ -149,15 +152,14 @@ class NBCNewsIE(InfoExtractor):
},
},
{
'url': 'http://www.nbcnews.com/feature/edward-snowden-interview/how-twitter-reacted-snowden-interview-n117236',
'md5': 'b2421750c9f260783721d898f4c42063',
'url': 'http://www.nbcnews.com/watch/nbcnews-com/how-twitter-reacted-to-the-snowden-interview-269389891880',
'md5': 'af1adfa51312291a017720403826bb64',
'info_dict': {
'id': 'I1wpAI_zmhsQ',
'id': '269389891880',
'ext': 'mp4',
'title': 'How Twitter Reacted To The Snowden Interview',
'description': 'md5:65a0bd5d76fe114f3c2727aa3a81fe64',
},
'add_ie': ['ThePlatform'],
},
{
'url': 'http://www.nbcnews.com/feature/dateline-full-episodes/full-episode-family-business-n285156',
@ -168,17 +170,29 @@ class NBCNewsIE(InfoExtractor):
'title': 'FULL EPISODE: Family Business',
'description': 'md5:757988edbaae9d7be1d585eb5d55cc04',
},
'skip': 'This page is unavailable.',
},
{
'url': 'http://www.nbcnews.com/nightly-news/video/nightly-news-with-brian-williams-full-broadcast-february-4-394064451844',
'md5': 'b5dda8cddd8650baa0dcb616dd2cf60d',
'md5': '73135a2e0ef819107bbb55a5a9b2a802',
'info_dict': {
'id': 'sekXqyTVnmN3',
'id': '394064451844',
'ext': 'mp4',
'title': 'Nightly News with Brian Williams Full Broadcast (February 4)',
'description': 'md5:1c10c1eccbe84a26e5debb4381e2d3c5',
},
},
{
'url': 'http://www.nbcnews.com/business/autos/volkswagen-11-million-vehicles-could-have-suspect-software-emissions-scandal-n431456',
'md5': 'a49e173825e5fcd15c13fc297fced39d',
'info_dict': {
'id': '529953347624',
'ext': 'mp4',
'title': 'Volkswagen U.S. Chief: We \'Totally Screwed Up\'',
'description': 'md5:d22d1281a24f22ea0880741bb4dd6301',
},
'expected_warnings': ['http-6000 is not available']
},
{
'url': 'http://www.nbcnews.com/watch/dateline/full-episode--deadly-betrayal-386250819952',
'only_matching': True,
@ -202,49 +216,80 @@ class NBCNewsIE(InfoExtractor):
}
else:
# "feature" and "nightly-news" pages use theplatform.com
title = mobj.group('title')
webpage = self._download_webpage(url, title)
display_id = mobj.group('display_id')
webpage = self._download_webpage(url, display_id)
info = None
bootstrap_json = self._search_regex(
r'var\s+(?:bootstrapJson|playlistData)\s*=\s*({.+});?\s*$',
webpage, 'bootstrap json', flags=re.MULTILINE)
bootstrap = self._parse_json(bootstrap_json, video_id)
info = bootstrap['results'][0]['video']
mpxid = info['mpxId']
r'(?m)var\s+(?:bootstrapJson|playlistData)\s*=\s*({.+});?\s*$',
webpage, 'bootstrap json', default=None)
if bootstrap_json:
bootstrap = self._parse_json(bootstrap_json, display_id)
info = bootstrap['results'][0]['video']
else:
player_instance_json = self._search_regex(
r'videoObj\s*:\s*({.+})', webpage, 'player instance')
info = self._parse_json(player_instance_json, display_id)
video_id = info['mpxId']
title = info['title']
base_urls = [
info['fallbackPlaylistUrl'],
info['associatedPlaylistUrl'],
]
subtitles = {}
caption_links = info.get('captionLinks')
if caption_links:
for (sub_key, sub_ext) in (('smpte-tt', 'ttml'), ('web-vtt', 'vtt'), ('srt', 'srt')):
sub_url = caption_links.get(sub_key)
if sub_url:
subtitles.setdefault('en', []).append({
'url': sub_url,
'ext': sub_ext,
})
for base_url in base_urls:
if not base_url:
formats = []
for video_asset in info['videoAssets']:
video_url = video_asset.get('publicUrl')
if not video_url:
continue
playlist_url = base_url + '?form=MPXNBCNewsAPI'
try:
all_videos = self._download_json(playlist_url, title)
except ExtractorError as ee:
if isinstance(ee.cause, compat_HTTPError):
continue
raise
if not all_videos or 'videos' not in all_videos:
container = video_asset.get('format')
asset_type = video_asset.get('assetType') or ''
if container == 'ISM' or asset_type == 'FireTV-Once':
continue
try:
info = next(v for v in all_videos['videos'] if v['mpxId'] == mpxid)
break
except StopIteration:
continue
if info is None:
raise ExtractorError('Could not find video in playlists')
elif asset_type == 'OnceURL':
tp_formats, tp_subtitles = self._extract_theplatform_smil(
video_url, video_id)
formats.extend(tp_formats)
subtitles = self._merge_subtitles(subtitles, tp_subtitles)
else:
tbr = int_or_none(video_asset.get('bitRate'), 1000)
format_id = 'http%s' % ('-%d' % tbr if tbr else '')
video_url = update_url_query(
video_url, {'format': 'redirect'})
# resolve the url so that we can check availability and detect the correct extension
head = self._request_webpage(
HEADRequest(video_url), video_id,
'Checking %s url' % format_id,
'%s is not available' % format_id,
fatal=False)
if head:
video_url = head.geturl()
formats.append({
'format_id': format_id,
'url': video_url,
'width': int_or_none(video_asset.get('width')),
'height': int_or_none(video_asset.get('height')),
'tbr': tbr,
'container': video_asset.get('format'),
})
self._sort_formats(formats)
return {
'_type': 'url',
# We get the best quality video
'url': info['videoAssets'][-1]['publicUrl'],
'ie_key': 'ThePlatform',
'id': video_id,
'title': title,
'description': info.get('description'),
'thumbnail': info.get('description'),
'thumbnail': info.get('thumbnail'),
'duration': int_or_none(info.get('duration')),
'timestamp': parse_iso8601(info.get('pubDate')),
'formats': formats,
'subtitles': subtitles,
}