paradizelost/youtube-dl
1
0
Fork 0
Code Issues Pull Requests Projects Releases Wiki Activity

Merge 43fe4b597f into 4d05f84325

Browse Source
This commit is contained in:
L0wbyte 2024-06-27 06:34:31 +08:00 committed by GitHub
commit aa584e202c
No known key found for this signature in database
GPG Key ID: B5690EEEBB952194
1 changed files with 55 additions and 44 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

99
youtube_dl/extractor/ciscolive.py
View File

@ -1,8 +1,6 @@
# coding: utf-8 # coding: utf-8
from __future__ import unicode_literals from __future__ import unicode_literals
import itertools
from .common import InfoExtractor from .common import InfoExtractor
from ..compat import ( from ..compat import (
compat_parse_qs, compat_parse_qs,
@ -10,6 +8,7 @@ from ..compat import (
) )
from ..utils import ( from ..utils import (
clean_html, clean_html,
ExtractorError,
float_or_none, float_or_none,
int_or_none, int_or_none,
try_get, try_get,
@ -21,12 +20,13 @@ class CiscoLiveBaseIE(InfoExtractor):
# These appear to be constant across all Cisco Live presentations # These appear to be constant across all Cisco Live presentations
# and are not tied to any user session or event # and are not tied to any user session or event
RAINFOCUS_API_URL = 'https://events.rainfocus.com/api/%s' RAINFOCUS_API_URL = 'https://events.rainfocus.com/api/%s'
RAINFOCUS_API_PROFILE_ID = 'Na3vqYdAlJFSxhYTYQGuMbpafMqftalz' RAINFOCUS_API_PROFILE_ID = '' # if blank will be fetched at runtime from site javascript
RAINFOCUS_WIDGET_ID = 'n6l4Lo05R8fiy3RpUBm447dZN8uNWoye' RAINFOCUS_WIDGET_ID = '' # if blank will be fetched at runtime from site javascript
RAINFOCUS_TOKENS_URL = 'https://cdn-events.rainfocus.com/pages/cisco/clondemand/catalog.js'
BRIGHTCOVE_URL_TEMPLATE = 'http://players.brightcove.net/5647924234001/SyK2FdqjM_default/index.html?videoId=%s' BRIGHTCOVE_URL_TEMPLATE = 'http://players.brightcove.net/5647924234001/SyK2FdqjM_default/index.html?videoId=%s'
HEADERS = { HEADERS = {
'Origin': 'https://ciscolive.cisco.com', 'Origin': 'https://ciscolive.com',
'rfApiProfileId': RAINFOCUS_API_PROFILE_ID, 'rfApiProfileId': RAINFOCUS_API_PROFILE_ID,
'rfWidgetId': RAINFOCUS_WIDGET_ID, 'rfWidgetId': RAINFOCUS_WIDGET_ID,
} }
@ -34,9 +34,25 @@ class CiscoLiveBaseIE(InfoExtractor):
def _call_api(self, ep, rf_id, query, referrer, note=None): def _call_api(self, ep, rf_id, query, referrer, note=None):
headers = self.HEADERS.copy() headers = self.HEADERS.copy()
headers['Referer'] = referrer headers['Referer'] = referrer
return self._download_json( if not self.RAINFOCUS_API_PROFILE_ID or not self.RAINFOCUS_WIDGET_ID:
rf_token_js = self._download_webpage(self.RAINFOCUS_TOKENS_URL, 'catalog.js', headers=headers)
for token in ('apiToken', 'widgetId'):
if token not in rf_token_js:
raise ExtractorError(
'Unable to fetch ' + token, expected=True)
api_token = self._html_search_regex(
r'''apiToken:\s+["'](\w+)''', rf_token_js, 'apiToken')
widget_id = self._html_search_regex(
r'''widgetId:\s+["'](\w+)''', rf_token_js, 'widgetId')
headers['rfApiProfileId'] = api_token
headers['rfWidgetId'] = widget_id
rf_result = self._download_json(
self.RAINFOCUS_API_URL % ep, rf_id, note=note, self.RAINFOCUS_API_URL % ep, rf_id, note=note,
data=urlencode_postdata(query), headers=headers) data=urlencode_postdata(query), headers=headers)
if int(rf_result['responseCode']) != 0:
raise ExtractorError(
'Rainfocus %s api returned a non success responseCode: %s. api keys might be invalid' % (ep, rf_result['responseCode']), expected=True)
return rf_result
def _parse_rf_item(self, rf_item): def _parse_rf_item(self, rf_item):
event_name = rf_item.get('eventName') event_name = rf_item.get('eventName')
@ -67,17 +83,16 @@ class CiscoLiveBaseIE(InfoExtractor):
class CiscoLiveSessionIE(CiscoLiveBaseIE): class CiscoLiveSessionIE(CiscoLiveBaseIE):
_VALID_URL = r'https?://(?:www\.)?ciscolive(?:\.cisco)?\.com/[^#]*#/session/(?P<id>[^/?&]+)' _VALID_URL = r'https?://(?:www\.)?ciscolive(?:\.cisco)?\.com/[^#]*#/session/(?P<id>[^/?&]+)'
_TESTS = [{ _TESTS = [{
'url': 'https://ciscolive.cisco.com/on-demand-library/?#/session/1423353499155001FoSs', 'url': 'https://www.ciscolive.com/on-demand/on-demand-library.html?search=#/session/16360600004400017rMx',
'md5': 'c98acf395ed9c9f766941c70f5352e22', 'md5': 'e0f5b0b2927b586ebff619294fec6926',
'info_dict': { 'info_dict': {
'id': '5803694304001', 'id': '6128601216001',
'ext': 'mp4', 'ext': 'mp4',
'title': '13 Smart Automations to Monitor Your Cisco IOS Network', 'title': 'A Deeper Dive into the Telco Cloud Architecture Evolution to support 5G and MEC - BRKSPG-1565',
'description': 'md5:ec4a436019e09a918dec17714803f7cc', 'description': 'md5:04bc54ec8ede2bbd9d21264bfaf16cb3',
'timestamp': 1530305395, 'timestamp': 1580474594,
'upload_date': '20180629', 'upload_date': '20200131',
'uploader_id': '5647924234001', 'uploader_id': '5647924234001',
'location': '16B Mezz.',
}, },
}, { }, {
'url': 'https://www.ciscolive.com/global/on-demand-library.html?search.event=ciscoliveemea2019#/session/15361595531500013WOU', 'url': 'https://www.ciscolive.com/global/on-demand-library.html?search.event=ciscoliveemea2019#/session/15361595531500013WOU',
@ -85,6 +100,9 @@ class CiscoLiveSessionIE(CiscoLiveBaseIE):
}, { }, {
'url': 'https://www.ciscolive.com/global/on-demand-library.html?#/session/1490051371645001kNaS', 'url': 'https://www.ciscolive.com/global/on-demand-library.html?#/session/1490051371645001kNaS',
'only_matching': True, 'only_matching': True,
}, {
'url': 'https://www.ciscolive.com/on-demand/on-demand-library.html?#/session/16360600774720017Iby',
'only_matching': True,
}] }]
def _real_extract(self, url): def _real_extract(self, url):
@ -94,16 +112,15 @@ class CiscoLiveSessionIE(CiscoLiveBaseIE):
class CiscoLiveSearchIE(CiscoLiveBaseIE): class CiscoLiveSearchIE(CiscoLiveBaseIE):
_VALID_URL = r'https?://(?:www\.)?ciscolive(?:\.cisco)?\.com/(?:global/)?on-demand-library(?:\.html|/)' _VALID_URL = r'https?://(?:www\.)?ciscolive(?:\.cisco)?\.com/(?:global/|on-demand/)?on-demand-library(?:\.html|/)'
_TESTS = [{ _TESTS = [{
'url': 'https://ciscolive.cisco.com/on-demand-library/?search.event=ciscoliveus2018&search.technicallevel=scpsSkillLevel_aintroductory&search.focus=scpsSessionFocus_designAndDeployment#/', 'url': 'https://www.ciscolive.com/on-demand/on-demand-library.html?search.technology=1614262524988009b6j9&search.technicallevel=scpsSkillLevel_bintermediate#/',
'info_dict': { 'info_dict': {
'title': 'Search query', 'title': 'Search query',
}, },
'playlist_count': 5, 'playlist_count': 6,
}, { }, {
'url': 'https://ciscolive.cisco.com/on-demand-library/?search.technology=scpsTechnology_applicationDevelopment&search.technology=scpsTechnology_ipv6&search.focus=scpsSessionFocus_troubleshootingTroubleshooting#/', 'url': 'https://www.ciscolive.com/on-demand/on-demand-library.html?search.technology=scpsTechnology_automation&search.technicallevel=scpsSkillLevel_cadvanced#/',
'only_matching': True,
}, { }, {
'url': 'https://www.ciscolive.com/global/on-demand-library.html?search.technicallevel=scpsSkillLevel_aintroductory&search.event=ciscoliveemea2019&search.technology=scpsTechnology_dataCenter&search.focus=scpsSessionFocus_bestPractices#/', 'url': 'https://www.ciscolive.com/global/on-demand-library.html?search.technicallevel=scpsSkillLevel_aintroductory&search.event=ciscoliveemea2019&search.technology=scpsTechnology_dataCenter&search.focus=scpsSessionFocus_bestPractices#/',
'only_matching': True, 'only_matching': True,
@ -118,31 +135,25 @@ class CiscoLiveSearchIE(CiscoLiveBaseIE):
return int_or_none(try_get(rf_item, lambda x: x['videos'][0]['url'])) is not None return int_or_none(try_get(rf_item, lambda x: x['videos'][0]['url'])) is not None
def _entries(self, query, url): def _entries(self, query, url):
query['size'] = 50 results = self._call_api(
query['from'] = 0 'search', None, query, url,
for page_num in itertools.count(1): 'Downloading search JSON')
results = self._call_api( if int(results['totalSearchItems']) == 0:
'search', None, query, url, raise ExtractorError(
'Downloading search JSON page %d' % page_num) 'Search api returned no items (if matches are expected rfApiProfileId may be invalid)',
sl = try_get(results, lambda x: x['sectionList'][0], dict) expected=True)
if sl: sl = try_get(results, lambda x: x['sectionList'], list)
results = sl if sl is not None:
items = results.get('items') for s in sl:
if not items or not isinstance(items, list): items = s.get('items')
break if not items or not isinstance(items, list):
for item in items: break
if not isinstance(item, dict): for item in items:
continue if not isinstance(item, dict):
if not self._check_bc_id_exists(item): continue
continue if not self._check_bc_id_exists(item):
yield self._parse_rf_item(item) continue
size = int_or_none(results.get('size')) yield self._parse_rf_item(item)
if size is not None:
query['size'] = size
total = int_or_none(results.get('total'))
if total is not None and query['from'] + query['size'] > total:
break
query['from'] += query['size']
def _real_extract(self, url): def _real_extract(self, url):
query = compat_parse_qs(compat_urllib_parse_urlparse(url).query) query = compat_parse_qs(compat_urllib_parse_urlparse(url).query)