mirror of https://github.com/blackjack4494/yt-dlc
[niconico] Support videos with multiple formats (closes #13522)
This commit is contained in:
parent
463e7216c8
commit
ee6a611665
|
@ -1,6 +1,7 @@
|
||||||
version <unreleased>
|
version <unreleased>
|
||||||
|
|
||||||
Extractors
|
Extractors
|
||||||
|
+ [niconico] Support videos with multiple formats (#13522)
|
||||||
+ [niconico] Support HTML5-only videos (#13806)
|
+ [niconico] Support HTML5-only videos (#13806)
|
||||||
|
|
||||||
|
|
||||||
|
|
|
@ -14,8 +14,10 @@ from ..utils import (
|
||||||
dict_get,
|
dict_get,
|
||||||
ExtractorError,
|
ExtractorError,
|
||||||
int_or_none,
|
int_or_none,
|
||||||
|
float_or_none,
|
||||||
parse_duration,
|
parse_duration,
|
||||||
parse_iso8601,
|
parse_iso8601,
|
||||||
|
remove_start,
|
||||||
try_get,
|
try_get,
|
||||||
unified_timestamp,
|
unified_timestamp,
|
||||||
urlencode_postdata,
|
urlencode_postdata,
|
||||||
|
@ -93,11 +95,12 @@ class NiconicoIE(InfoExtractor):
|
||||||
},
|
},
|
||||||
'skip': 'The viewing period of the video you were searching for has expired.',
|
'skip': 'The viewing period of the video you were searching for has expired.',
|
||||||
}, {
|
}, {
|
||||||
# video not available via `getflv`
|
# video not available via `getflv`; "old" HTML5 video
|
||||||
'url': 'http://www.nicovideo.jp/watch/sm1151009',
|
'url': 'http://www.nicovideo.jp/watch/sm1151009',
|
||||||
|
'md5': '8fa81c364eb619d4085354eab075598a',
|
||||||
'info_dict': {
|
'info_dict': {
|
||||||
'id': 'sm1151009',
|
'id': 'sm1151009',
|
||||||
'ext': 'flv',
|
'ext': 'mp4',
|
||||||
'title': 'マスターシステム本体内蔵のスペハリのメインテーマ(PSG版)',
|
'title': 'マスターシステム本体内蔵のスペハリのメインテーマ(PSG版)',
|
||||||
'description': 'md5:6ee077e0581ff5019773e2e714cdd0b7',
|
'description': 'md5:6ee077e0581ff5019773e2e714cdd0b7',
|
||||||
'thumbnail': r're:https?://.*',
|
'thumbnail': r're:https?://.*',
|
||||||
|
@ -110,6 +113,25 @@ class NiconicoIE(InfoExtractor):
|
||||||
'comment_count': int,
|
'comment_count': int,
|
||||||
},
|
},
|
||||||
'skip': 'Requires an account',
|
'skip': 'Requires an account',
|
||||||
|
}, {
|
||||||
|
# "New" HTML5 video
|
||||||
|
'url': 'http://www.nicovideo.jp/watch/sm31464864',
|
||||||
|
'md5': '351647b4917660986dc0fa8864085135',
|
||||||
|
'info_dict': {
|
||||||
|
'id': 'sm31464864',
|
||||||
|
'ext': 'mp4',
|
||||||
|
'title': '新作TVアニメ「戦姫絶唱シンフォギアAXZ」PV 最高画質',
|
||||||
|
'description': 'md5:e52974af9a96e739196b2c1ca72b5feb',
|
||||||
|
'timestamp': 1498514060,
|
||||||
|
'upload_date': '20170626',
|
||||||
|
'uploader': 'ゲス',
|
||||||
|
'uploader_id': '40826363',
|
||||||
|
'thumbnail': r're:https?://.*',
|
||||||
|
'duration': 198,
|
||||||
|
'view_count': int,
|
||||||
|
'comment_count': int,
|
||||||
|
},
|
||||||
|
'skip': 'Requires an account',
|
||||||
}, {
|
}, {
|
||||||
'url': 'http://sp.nicovideo.jp/watch/sm28964488?ss_pos=1&cp_in=wt_tg',
|
'url': 'http://sp.nicovideo.jp/watch/sm28964488?ss_pos=1&cp_in=wt_tg',
|
||||||
'only_matching': True,
|
'only_matching': True,
|
||||||
|
@ -147,6 +169,84 @@ class NiconicoIE(InfoExtractor):
|
||||||
self._downloader.report_warning('unable to log in: bad username or password')
|
self._downloader.report_warning('unable to log in: bad username or password')
|
||||||
return login_ok
|
return login_ok
|
||||||
|
|
||||||
|
def _extract_format_for_quality(self, api_data, video_id, audio_quality, video_quality):
|
||||||
|
def yesno(boolean):
|
||||||
|
return 'yes' if boolean else 'no'
|
||||||
|
|
||||||
|
session_api_data = api_data['video']['dmcInfo']['session_api']
|
||||||
|
session_api_endpoint = session_api_data['urls'][0]
|
||||||
|
|
||||||
|
format_id = '-'.join(map(lambda s: remove_start(s['id'], 'archive_'), [video_quality, audio_quality]))
|
||||||
|
|
||||||
|
session_response = self._download_json(
|
||||||
|
session_api_endpoint['url'], video_id,
|
||||||
|
query={'_format': 'json'},
|
||||||
|
headers={'Content-Type': 'application/json'},
|
||||||
|
note='Downloading JSON metadata for %s' % format_id,
|
||||||
|
data=json.dumps({
|
||||||
|
'session': {
|
||||||
|
'client_info': {
|
||||||
|
'player_id': session_api_data['player_id'],
|
||||||
|
},
|
||||||
|
'content_auth': {
|
||||||
|
'auth_type': session_api_data['auth_types'][session_api_data['protocols'][0]],
|
||||||
|
'content_key_timeout': session_api_data['content_key_timeout'],
|
||||||
|
'service_id': 'nicovideo',
|
||||||
|
'service_user_id': session_api_data['service_user_id']
|
||||||
|
},
|
||||||
|
'content_id': session_api_data['content_id'],
|
||||||
|
'content_src_id_sets': [{
|
||||||
|
'content_src_ids': [{
|
||||||
|
'src_id_to_mux': {
|
||||||
|
'audio_src_ids': [audio_quality['id']],
|
||||||
|
'video_src_ids': [video_quality['id']],
|
||||||
|
}
|
||||||
|
}]
|
||||||
|
}],
|
||||||
|
'content_type': 'movie',
|
||||||
|
'content_uri': '',
|
||||||
|
'keep_method': {
|
||||||
|
'heartbeat': {
|
||||||
|
'lifetime': session_api_data['heartbeat_lifetime']
|
||||||
|
}
|
||||||
|
},
|
||||||
|
'priority': session_api_data['priority'],
|
||||||
|
'protocol': {
|
||||||
|
'name': 'http',
|
||||||
|
'parameters': {
|
||||||
|
'http_parameters': {
|
||||||
|
'parameters': {
|
||||||
|
'http_output_download_parameters': {
|
||||||
|
'use_ssl': yesno(session_api_endpoint['is_ssl']),
|
||||||
|
'use_well_known_port': yesno(session_api_endpoint['is_well_known_port']),
|
||||||
|
}
|
||||||
|
}
|
||||||
|
}
|
||||||
|
}
|
||||||
|
},
|
||||||
|
'recipe_id': session_api_data['recipe_id'],
|
||||||
|
'session_operation_auth': {
|
||||||
|
'session_operation_auth_by_signature': {
|
||||||
|
'signature': session_api_data['signature'],
|
||||||
|
'token': session_api_data['token'],
|
||||||
|
}
|
||||||
|
},
|
||||||
|
'timing_constraint': 'unlimited'
|
||||||
|
}
|
||||||
|
}))
|
||||||
|
|
||||||
|
resolution = video_quality.get('resolution', {})
|
||||||
|
|
||||||
|
return {
|
||||||
|
'url': session_response['data']['session']['content_uri'],
|
||||||
|
'format_id': format_id,
|
||||||
|
'ext': 'mp4', # Session API are used in HTML5, which always serves mp4
|
||||||
|
'abr': float_or_none(audio_quality.get('bitrate'), 1000),
|
||||||
|
'vbr': float_or_none(video_quality.get('bitrate'), 1000),
|
||||||
|
'height': resolution.get('height'),
|
||||||
|
'width': resolution.get('width'),
|
||||||
|
}
|
||||||
|
|
||||||
def _real_extract(self, url):
|
def _real_extract(self, url):
|
||||||
video_id = self._match_id(url)
|
video_id = self._match_id(url)
|
||||||
|
|
||||||
|
@ -161,13 +261,13 @@ class NiconicoIE(InfoExtractor):
|
||||||
api_data = self._parse_json(self._html_search_regex(
|
api_data = self._parse_json(self._html_search_regex(
|
||||||
'data-api-data="([^"]+)"', webpage,
|
'data-api-data="([^"]+)"', webpage,
|
||||||
'API data', default='{}'), video_id)
|
'API data', default='{}'), video_id)
|
||||||
video_real_url = try_get(
|
|
||||||
api_data, lambda x: x['video']['smileInfo']['url'])
|
|
||||||
|
|
||||||
if video_real_url:
|
def _format_id_from_url(video_url):
|
||||||
def get_video_info(items):
|
return 'economy' if video_real_url.endswith('low') else 'normal'
|
||||||
return dict_get(api_data['video'], items)
|
|
||||||
else:
|
try:
|
||||||
|
video_real_url = api_data['video']['smileInfo']['url']
|
||||||
|
except KeyError: # Flash videos
|
||||||
# Get flv info
|
# Get flv info
|
||||||
flv_info_webpage = self._download_webpage(
|
flv_info_webpage = self._download_webpage(
|
||||||
'http://flapi.nicovideo.jp/api/getflv/' + video_id + '?as3=1',
|
'http://flapi.nicovideo.jp/api/getflv/' + video_id + '?as3=1',
|
||||||
|
@ -187,8 +287,6 @@ class NiconicoIE(InfoExtractor):
|
||||||
else:
|
else:
|
||||||
raise ExtractorError('Unable to find video URL')
|
raise ExtractorError('Unable to find video URL')
|
||||||
|
|
||||||
video_real_url = flv_info['url'][0]
|
|
||||||
|
|
||||||
video_info_xml = self._download_xml(
|
video_info_xml = self._download_xml(
|
||||||
'http://ext.nicovideo.jp/api/getthumbinfo/' + video_id,
|
'http://ext.nicovideo.jp/api/getthumbinfo/' + video_id,
|
||||||
video_id, note='Downloading video info page')
|
video_id, note='Downloading video info page')
|
||||||
|
@ -201,6 +299,41 @@ class NiconicoIE(InfoExtractor):
|
||||||
if ret:
|
if ret:
|
||||||
return ret
|
return ret
|
||||||
|
|
||||||
|
video_real_url = flv_info['url'][0]
|
||||||
|
|
||||||
|
extension = get_video_info('movie_type')
|
||||||
|
if not extension:
|
||||||
|
extension = determine_ext(video_real_url)
|
||||||
|
|
||||||
|
formats = [{
|
||||||
|
'url': video_real_url,
|
||||||
|
'ext': extension,
|
||||||
|
'format_id': _format_id_from_url(video_real_url),
|
||||||
|
}]
|
||||||
|
else:
|
||||||
|
formats = []
|
||||||
|
|
||||||
|
dmc_info = api_data['video'].get('dmcInfo')
|
||||||
|
if dmc_info: # "New" HTML5 videos
|
||||||
|
quality_info = dmc_info['quality']
|
||||||
|
for audio_quality in quality_info['audios']:
|
||||||
|
for video_quality in quality_info['videos']:
|
||||||
|
if not audio_quality['available'] or not video_quality['available']:
|
||||||
|
continue
|
||||||
|
formats.append(self._extract_format_for_quality(
|
||||||
|
api_data, video_id, audio_quality, video_quality))
|
||||||
|
|
||||||
|
self._sort_formats(formats)
|
||||||
|
else: # "Old" HTML5 videos
|
||||||
|
formats = [{
|
||||||
|
'url': video_real_url,
|
||||||
|
'ext': 'mp4',
|
||||||
|
'format_id': _format_id_from_url(video_real_url),
|
||||||
|
}]
|
||||||
|
|
||||||
|
def get_video_info(items):
|
||||||
|
return dict_get(api_data['video'], items)
|
||||||
|
|
||||||
# Start extracting information
|
# Start extracting information
|
||||||
title = get_video_info('title')
|
title = get_video_info('title')
|
||||||
if not title:
|
if not title:
|
||||||
|
@ -216,10 +349,6 @@ class NiconicoIE(InfoExtractor):
|
||||||
watch_api_data = self._parse_json(watch_api_data_string, video_id) if watch_api_data_string else {}
|
watch_api_data = self._parse_json(watch_api_data_string, video_id) if watch_api_data_string else {}
|
||||||
video_detail = watch_api_data.get('videoDetail', {})
|
video_detail = watch_api_data.get('videoDetail', {})
|
||||||
|
|
||||||
extension = get_video_info(['movie_type', 'movieType'])
|
|
||||||
if not extension:
|
|
||||||
extension = determine_ext(video_real_url)
|
|
||||||
|
|
||||||
thumbnail = (
|
thumbnail = (
|
||||||
get_video_info(['thumbnail_url', 'thumbnailURL']) or
|
get_video_info(['thumbnail_url', 'thumbnailURL']) or
|
||||||
self._html_search_meta('image', webpage, 'thumbnail', default=None) or
|
self._html_search_meta('image', webpage, 'thumbnail', default=None) or
|
||||||
|
@ -272,10 +401,8 @@ class NiconicoIE(InfoExtractor):
|
||||||
|
|
||||||
return {
|
return {
|
||||||
'id': video_id,
|
'id': video_id,
|
||||||
'url': video_real_url,
|
|
||||||
'title': title,
|
'title': title,
|
||||||
'ext': extension,
|
'formats': formats,
|
||||||
'format_id': 'economy' if video_real_url.endswith('low') else 'normal',
|
|
||||||
'thumbnail': thumbnail,
|
'thumbnail': thumbnail,
|
||||||
'description': description,
|
'description': description,
|
||||||
'uploader': uploader,
|
'uploader': uploader,
|
||||||
|
|
Loading…
Reference in New Issue