# coding: utf-8
from __future__ import unicode_literals
-import datetime
-import functools
+import re
import json
-import math
+import datetime
from .common import InfoExtractor
+from ..postprocessor.ffmpeg import FFmpegPostProcessor
from ..compat import (
compat_parse_qs,
compat_urllib_parse_urlparse,
)
from ..utils import (
- determine_ext,
dict_get,
ExtractorError,
- float_or_none,
- InAdvancePagedList,
int_or_none,
+ float_or_none,
+ OnDemandPagedList,
parse_duration,
parse_iso8601,
+ PostProcessingError,
remove_start,
try_get,
unified_timestamp,
self._downloader.report_warning('unable to log in: bad username or password')
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'
+ def _get_heartbeat_info(self, info_dict):
- session_api_data = api_data['video']['dmcInfo']['session_api']
- session_api_endpoint = session_api_data['urls'][0]
+ video_id, video_src_id, audio_src_id = info_dict['url'].split(':')[1].split('/')
- format_id = '-'.join(map(lambda s: remove_start(s['id'], 'archive_'), [video_quality, audio_quality]))
+ # Get video webpage for API data.
+ webpage, handle = self._download_webpage_handle(
+ 'http://www.nicovideo.jp/watch/' + video_id, video_id)
+
+ api_data = self._parse_json(self._html_search_regex(
+ 'data-api-data="([^"]+)"', webpage,
+ 'API data', default='{}'), video_id)
+
+ session_api_data = try_get(api_data, lambda x: x['video']['dmcInfo']['session_api'])
+ session_api_endpoint = try_get(session_api_data, lambda x: x['urls'][0])
+
+ # ping
+ self._download_json(
+ 'https://nvapi.nicovideo.jp/v1/2ab0cbaa/watch', video_id,
+ query={'t': try_get(api_data, lambda x: x['video']['dmcInfo']['tracking_id'])},
+ headers={
+ 'Origin': 'https://www.nicovideo.jp',
+ 'Referer': 'https://www.nicovideo.jp/watch/' + video_id,
+ 'X-Frontend-Id': '6',
+ 'X-Frontend-Version': '0'
+ })
+
+ yesno = lambda x: 'yes' if x else 'no'
+
+ # m3u8 (encryption)
+ if 'encryption' in try_get(api_data, lambda x: x['video']['dmcInfo']) or {}:
+ protocol = 'm3u8'
+ session_api_http_parameters = {
+ 'parameters': {
+ 'hls_parameters': {
+ 'encryption': {
+ 'hls_encryption_v1': {
+ 'encrypted_key': try_get(api_data, lambda x: x['video']['dmcInfo']['encryption']['hls_encryption_v1']['encrypted_key']),
+ 'key_uri': try_get(api_data, lambda x: x['video']['dmcInfo']['encryption']['hls_encryption_v1']['key_uri'])
+ }
+ },
+ 'transfer_preset': '',
+ 'use_ssl': yesno(session_api_endpoint['is_ssl']),
+ 'use_well_known_port': yesno(session_api_endpoint['is_well_known_port']),
+ 'segment_duration': 6000
+ }
+ }
+ }
+ # http
+ else:
+ protocol = 'http'
+ session_api_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']),
+ }
+ }
+ }
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,
+ note='Downloading JSON metadata for %s' % info_dict['format_id'],
data=json.dumps({
'session': {
'client_info': {
- 'player_id': session_api_data['player_id'],
+ 'player_id': session_api_data.get('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'],
+ 'auth_type': try_get(session_api_data, lambda x: x['auth_types'][session_api_data['protocols'][0]]),
+ 'content_key_timeout': session_api_data.get('content_key_timeout'),
'service_id': 'nicovideo',
- 'service_user_id': session_api_data['service_user_id']
+ 'service_user_id': session_api_data.get('service_user_id')
},
- 'content_id': session_api_data['content_id'],
+ 'content_id': session_api_data.get('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']],
+ 'audio_src_ids': [audio_src_id],
+ 'video_src_ids': [video_src_id],
}
}]
}],
'content_uri': '',
'keep_method': {
'heartbeat': {
- 'lifetime': session_api_data['heartbeat_lifetime']
+ 'lifetime': session_api_data.get('heartbeat_lifetime')
}
},
- 'priority': session_api_data['priority'],
+ 'priority': session_api_data.get('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']),
- }
- }
- }
+ 'http_parameters': session_api_http_parameters
}
},
- 'recipe_id': session_api_data['recipe_id'],
+ 'recipe_id': session_api_data.get('recipe_id'),
'session_operation_auth': {
'session_operation_auth_by_signature': {
- 'signature': session_api_data['signature'],
- 'token': session_api_data['token'],
+ 'signature': session_api_data.get('signature'),
+ 'token': session_api_data.get('token'),
}
},
'timing_constraint': 'unlimited'
}
}).encode())
- resolution = video_quality.get('resolution', {})
+ info_dict['url'] = session_response['data']['session']['content_uri']
+ info_dict['protocol'] = protocol
+
+ # get heartbeat info
+ heartbeat_info_dict = {
+ 'url': session_api_endpoint['url'] + '/' + session_response['data']['session']['id'] + '?_format=json&_method=PUT',
+ 'data': json.dumps(session_response['data']),
+ # interval, convert milliseconds to seconds, then halve to make a buffer.
+ 'interval': float_or_none(session_api_data.get('heartbeat_lifetime'), scale=2000),
+ }
+
+ return info_dict, heartbeat_info_dict
+
+ def _extract_format_for_quality(self, api_data, video_id, audio_quality, video_quality):
+ def parse_format_id(id_code):
+ mobj = re.match(r'''(?x)
+ (?:archive_)?
+ (?:(?P<codec>[^_]+)_)?
+ (?:(?P<br>[\d]+)kbps_)?
+ (?:(?P<res>[\d+]+)p_)?
+ ''', '%s_' % id_code)
+ return mobj.groupdict() if mobj else {}
+
+ protocol = 'niconico_dmc'
+ format_id = '-'.join(map(lambda s: remove_start(s['id'], 'archive_'), [video_quality, audio_quality]))
+ vdict = parse_format_id(video_quality['id'])
+ adict = parse_format_id(audio_quality['id'])
+ resolution = video_quality.get('resolution', {'height': vdict.get('res')})
return {
- 'url': session_response['data']['session']['content_uri'],
+ 'url': '%s:%s/%s/%s' % (protocol, video_id, video_quality['id'], audio_quality['id']),
'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'),
+ 'vcodec': vdict.get('codec'),
+ 'acodec': adict.get('codec'),
+ 'vbr': float_or_none(video_quality.get('bitrate'), 1000) or float_or_none(vdict.get('br')),
+ 'abr': float_or_none(audio_quality.get('bitrate'), 1000) or float_or_none(adict.get('br')),
+ 'height': int_or_none(resolution.get('height', vdict.get('res'))),
+ 'width': int_or_none(resolution.get('width')),
+ 'quality': -2 if 'low' in format_id else -1, # Default quality value is -1
+ 'protocol': protocol,
+ 'http_headers': {
+ 'Origin': 'https://www.nicovideo.jp',
+ 'Referer': 'https://www.nicovideo.jp/watch/' + video_id,
+ }
}
def _real_extract(self, url):
video_id = self._match_id(url)
- # Get video webpage. We are not actually interested in it for normal
- # cases, but need the cookies in order to be able to download the
- # info webpage
+ # Get video webpage for API data.
webpage, handle = self._download_webpage_handle(
'http://www.nicovideo.jp/watch/' + video_id, video_id)
if video_id.startswith('so'):
'data-api-data="([^"]+)"', webpage,
'API data', default='{}'), video_id)
- def _format_id_from_url(video_url):
- return 'economy' if video_real_url.endswith('low') else 'normal'
+ def get_video_info_web(items):
+ return dict_get(api_data['video'], items)
+
+ # Get video info
+ video_info_xml = self._download_xml(
+ 'http://ext.nicovideo.jp/api/getthumbinfo/' + video_id,
+ video_id, note='Downloading video info page')
+
+ def get_video_info_xml(items):
+ if not isinstance(items, list):
+ items = [items]
+ for item in items:
+ ret = xpath_text(video_info_xml, './/' + item)
+ if ret:
+ return ret
+
+ if get_video_info_xml('error'):
+ error_code = get_video_info_xml('code')
+
+ if error_code == 'DELETED':
+ raise ExtractorError('The video has been deleted.',
+ expected=True)
+ elif error_code == 'NOT_FOUND':
+ raise ExtractorError('The video is not found.',
+ expected=True)
+ elif error_code == 'COMMUNITY':
+ self.to_screen('%s: The video is community members only.' % video_id)
+ else:
+ raise ExtractorError('%s reports error: %s' % (self.IE_NAME, error_code))
+
+ # Start extracting video formats
+ formats = []
+
+ # Get HTML5 videos info
+ try:
+ dmc_info = api_data['video']['dmcInfo']
+ except KeyError:
+ raise ExtractorError('The video can\'t downloaded.',
+ expected=True)
+
+ quality_info = dmc_info.get('quality')
+ for audio_quality in quality_info.get('audios') or {}:
+ for video_quality in quality_info.get('videos') or {}:
+ if not audio_quality.get('available') or not video_quality.get('available'):
+ continue
+ formats.append(self._extract_format_for_quality(
+ api_data, video_id, audio_quality, video_quality))
+
+ # Get flv/swf info
+ video_real_url = try_get(api_data, lambda x: x['video']['smileInfo']['url'])
+ is_economy = video_real_url.endswith('low')
+
+ if is_economy:
+ self.report_warning('Site is currently in economy mode! You will only have access to lower quality streams')
+
+ # Invoking ffprobe to determine resolution
+ pp = FFmpegPostProcessor(self._downloader)
+ cookies = self._get_cookies('https://nicovideo.jp').output(header='', sep='; path=/; domain=nicovideo.jp;\n')
+
+ self.to_screen('%s: %s' % (video_id, 'Checking smile format with ffprobe'))
try:
- video_real_url = api_data['video']['smileInfo']['url']
- except KeyError: # Flash videos
- # Get flv info
- flv_info_webpage = self._download_webpage(
- 'http://flapi.nicovideo.jp/api/getflv/' + video_id + '?as3=1',
- video_id, 'Downloading flv info')
-
- flv_info = compat_parse_qs(flv_info_webpage)
- if 'url' not in flv_info:
- if 'deleted' in flv_info:
- raise ExtractorError('The video has been deleted.',
- expected=True)
- elif 'closed' in flv_info:
- raise ExtractorError('Niconico videos now require logging in',
- expected=True)
- elif 'error' in flv_info:
- raise ExtractorError('%s reports error: %s' % (
- self.IE_NAME, flv_info['error'][0]), expected=True)
- else:
- raise ExtractorError('Unable to find video URL')
-
- video_info_xml = self._download_xml(
- 'http://ext.nicovideo.jp/api/getthumbinfo/' + video_id,
- video_id, note='Downloading video info page')
-
- def get_video_info(items):
- if not isinstance(items, list):
- items = [items]
- for item in items:
- ret = xpath_text(video_info_xml, './/' + item)
- if 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 = [{
+ metadata = pp.get_metadata_object(video_real_url, ['-cookies', cookies])
+ except PostProcessingError as err:
+ raise ExtractorError(err.msg, expected=True)
+
+ v_stream = a_stream = {}
+
+ # Some complex swf files doesn't have video stream (e.g. nm4809023)
+ for stream in metadata['streams']:
+ if stream['codec_type'] == 'video':
+ v_stream = stream
+ elif stream['codec_type'] == 'audio':
+ a_stream = stream
+
+ # Community restricted videos seem to have issues with the thumb API not returning anything at all
+ filesize = int(
+ (get_video_info_xml('size_high') if not is_economy else get_video_info_xml('size_low'))
+ or metadata['format']['size']
+ )
+ extension = (
+ get_video_info_xml('movie_type')
+ or 'mp4' if 'mp4' in metadata['format']['format_name'] else metadata['format']['format_name']
+ )
+
+ # 'creation_time' tag on video stream of re-encoded SMILEVIDEO mp4 files are '1970-01-01T00:00:00.000000Z'.
+ timestamp = (
+ parse_iso8601(get_video_info_web('first_retrieve'))
+ or unified_timestamp(get_video_info_web('postedDateTime'))
+ )
+ metadata_timestamp = (
+ parse_iso8601(try_get(v_stream, lambda x: x['tags']['creation_time']))
+ or timestamp if extension != 'mp4' else 0
+ )
+
+ # According to compconf, smile videos from pre-2017 are always better quality than their DMC counterparts
+ smile_threshold_timestamp = parse_iso8601('2016-12-08T00:00:00+09:00')
+
+ is_source = timestamp < smile_threshold_timestamp or metadata_timestamp > 0
+
+ # If movie file size is unstable, old server movie is not source movie.
+ if filesize > 1:
+ formats.append({
'url': video_real_url,
+ 'format_id': 'smile' if not is_economy else 'smile_low',
+ 'format_note': 'SMILEVIDEO source' if not is_economy else 'SMILEVIDEO low quality',
'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)
+ 'container': extension,
+ 'vcodec': v_stream.get('codec_name'),
+ 'acodec': a_stream.get('codec_name'),
+ # Some complex swf files doesn't have total bit rate metadata (e.g. nm6049209)
+ 'tbr': int_or_none(metadata['format'].get('bit_rate'), scale=1000),
+ 'vbr': int_or_none(v_stream.get('bit_rate'), scale=1000),
+ 'abr': int_or_none(a_stream.get('bit_rate'), scale=1000),
+ 'height': int_or_none(v_stream.get('height')),
+ 'width': int_or_none(v_stream.get('width')),
+ 'source_preference': 5 if not is_economy else -2,
+ 'quality': 5 if is_source and not is_economy else None,
+ 'filesize': filesize
+ })
+
+ if len(formats) == 0:
+ raise ExtractorError('Unable to find video info.')
+
+ self._sort_formats(formats)
# Start extracting information
- title = get_video_info('title')
+ title = get_video_info_web('originalTitle')
if not title:
title = self._og_search_title(webpage, default=None)
if not title:
video_detail = watch_api_data.get('videoDetail', {})
thumbnail = (
- get_video_info(['thumbnail_url', 'thumbnailURL'])
+ self._html_search_regex(r'<meta property="og:image" content="([^"]+)">', webpage, 'thumbnail data', default=None)
+ or get_video_info_web(['thumbnail_url', 'largeThumbnailURL', 'thumbnailURL'])
or self._html_search_meta('image', webpage, 'thumbnail', default=None)
or video_detail.get('thumbnail'))
- description = get_video_info('description')
+ description = get_video_info_web('description')
- timestamp = (parse_iso8601(get_video_info('first_retrieve'))
- or unified_timestamp(get_video_info('postedDateTime')))
if not timestamp:
match = self._html_search_meta('datePublished', webpage, 'date published', default=None)
if match:
video_detail['postedAt'].replace('/', '-'),
delimiter=' ', timezone=datetime.timedelta(hours=9))
- view_count = int_or_none(get_video_info(['view_counter', 'viewCount']))
+ view_count = int_or_none(get_video_info_web(['view_counter', 'viewCount']))
if not view_count:
match = self._html_search_regex(
r'>Views: <strong[^>]*>([^<]+)</strong>',
view_count = int_or_none(match.replace(',', ''))
view_count = view_count or video_detail.get('viewCount')
- comment_count = (int_or_none(get_video_info('comment_num'))
+ comment_count = (int_or_none(get_video_info_web('comment_num'))
or video_detail.get('commentCount')
or try_get(api_data, lambda x: x['thread']['commentCount']))
if not comment_count:
comment_count = int_or_none(match.replace(',', ''))
duration = (parse_duration(
- get_video_info('length')
+ get_video_info_web('length')
or self._html_search_meta(
'video:duration', webpage, 'video duration', default=None))
or video_detail.get('length')
- or get_video_info('duration'))
+ or get_video_info_web('duration'))
- webpage_url = get_video_info('watch_url') or url
+ webpage_url = get_video_info_web('watch_url') or url
# Note: cannot use api_data.get('owner', {}) because owner may be set to "null"
# in the JSON, which will cause None to be returned instead of {}.
owner = try_get(api_data, lambda x: x.get('owner'), dict) or {}
- uploader_id = get_video_info(['ch_id', 'user_id']) or owner.get('id')
- uploader = get_video_info(['ch_name', 'user_nickname']) or owner.get('nickname')
+ uploader_id = get_video_info_web(['ch_id', 'user_id']) or owner.get('id')
+ uploader = get_video_info_web(['ch_name', 'user_nickname']) or owner.get('nickname')
return {
'id': video_id,
'url': 'https://www.nicovideo.jp/user/805442/mylist/27411728',
'only_matching': True,
}]
- _PAGE_SIZE = 100
-
- def _call_api(self, list_id, resource, query):
- return self._download_json(
- 'https://nvapi.nicovideo.jp/v2/mylists/' + list_id, list_id,
- 'Downloading %s JSON metatdata' % resource, query=query,
- headers={'X-Frontend-Id': 6})['data']['mylist']
-
- def _parse_owner(self, item):
- owner = item.get('owner') or {}
- if owner:
- return {
- 'uploader': owner.get('name'),
- 'uploader_id': owner.get('id'),
- }
- return {}
-
- def _fetch_page(self, list_id, page):
- page += 1
- items = self._call_api(list_id, 'page %d' % page, {
- 'page': page,
- 'pageSize': self._PAGE_SIZE,
- })['items']
- for item in items:
- video = item.get('video') or {}
- video_id = video.get('id')
- if not video_id:
- continue
- count = video.get('count') or {}
- get_count = lambda x: int_or_none(count.get(x))
- info = {
- '_type': 'url',
- 'id': video_id,
- 'title': video.get('title'),
- 'url': 'https://www.nicovideo.jp/watch/' + video_id,
- 'description': video.get('shortDescription'),
- 'duration': int_or_none(video.get('duration')),
- 'view_count': get_count('view'),
- 'comment_count': get_count('comment'),
- 'ie_key': NiconicoIE.ie_key(),
- }
- info.update(self._parse_owner(video))
- yield info
def _real_extract(self, url):
list_id = self._match_id(url)
- mylist = self._call_api(list_id, 'list', {
- 'pageSize': 1,
- })
- entries = InAdvancePagedList(
- functools.partial(self._fetch_page, list_id),
- math.ceil(mylist['totalItemCount'] / self._PAGE_SIZE),
- self._PAGE_SIZE)
- result = self.playlist_result(
- entries, list_id, mylist.get('name'), mylist.get('description'))
- result.update(self._parse_owner(mylist))
- return result
+ webpage = self._download_webpage(url, list_id)
+
+ header = self._parse_json(self._html_search_regex(
+ r'data-common-header="([^"]+)"', webpage,
+ 'webpage header'), list_id)
+ frontendId = header.get('initConfig').get('frontendId')
+ frontendVersion = header.get('initConfig').get('frontendVersion')
+
+ def get_page_data(pagenum, pagesize):
+ return self._download_json(
+ 'http://nvapi.nicovideo.jp/v2/mylists/' + list_id, list_id,
+ query={'page': 1 + pagenum, 'pageSize': pagesize},
+ headers={
+ 'X-Frontend-Id': frontendId,
+ 'X-Frontend-Version': frontendVersion,
+ }).get('data').get('mylist')
+
+ data = get_page_data(0, 1)
+ title = data.get('name')
+ description = data.get('description')
+ uploader = data.get('owner').get('name')
+ uploader_id = data.get('owner').get('id')
+
+ def pagefunc(pagenum):
+ data = get_page_data(pagenum, 25)
+ return ({
+ '_type': 'url',
+ 'url': 'http://www.nicovideo.jp/watch/' + item.get('watchId'),
+ } for item in data.get('items'))
+
+ return {
+ '_type': 'playlist',
+ 'id': list_id,
+ 'title': title,
+ 'description': description,
+ 'uploader': uploader,
+ 'uploader_id': uploader_id,
+ 'entries': OnDemandPagedList(pagefunc, 25),
+ }
projects / yt-dlp.git / commitdiff
