mirrors/yt-dlp
1
0
Fork 0
mirror of https://github.com/yt-dlp/yt-dlp.git synced 2024-11-27 10:31:29 +00:00
Code Issues Packages Projects Releases Wiki Activity

[extractor/youtube:tab] Support reporthistory page

Browse source 
Closes #4929
This commit is contained in:
pukkandan 2022-09-27 03:55:58 +05:30
parent 0f60ba6e65
commit 0a5095fe8d
No known key found for this signature in database
GPG key ID: 7EEE9E1E817D0A39
1 changed files with 18 additions and 6 deletions
Show stats Download patch file Download diff file Expand all files Collapse all files

24
yt_dlp/extractor/youtube.py
View file

@ -292,7 +292,7 @@ class YoutubeBaseInfoExtractor(InfoExtractor):
r'channel|c|user|playlist|watch|w|v|embed|e|watch_popup|clip|' r'channel|c|user|playlist|watch|w|v|embed|e|watch_popup|clip|'
r'shorts|movies|results|search|shared|hashtag|trending|explore|feed|feeds|' r'shorts|movies|results|search|shared|hashtag|trending|explore|feed|feeds|'
r'browse|oembed|get_video_info|iframe_api|s/player|' r'browse|oembed|get_video_info|iframe_api|s/player|'
r'storefront|oops|index|account|reporthistory|t/terms|about|upload|signin|logout') r'storefront|oops|index|account|t/terms|about|upload|signin|logout')
_PLAYLIST_ID_RE = r'(?:(?:PL|LL|EC|UU|FL|RD|UL|TL|PU|OLAK5uy_)[0-9A-Za-z-_]{10,}|RDMM|WL|LL|LM)' _PLAYLIST_ID_RE = r'(?:(?:PL|LL|EC|UU|FL|RD|UL|TL|PU|OLAK5uy_)[0-9A-Za-z-_]{10,}|RDMM|WL|LL|LM)'
@ -673,7 +673,7 @@ def _extract_continuation(cls, renderer):
return next_continuation return next_continuation
contents = [] contents = []
for key in ('contents', 'items'): for key in ('contents', 'items', 'rows'):
contents.extend(try_get(renderer, lambda x: x[key], list) or []) contents.extend(try_get(renderer, lambda x: x[key], list) or [])
for content in contents: for content in contents:
@ -4405,6 +4405,13 @@ def _rich_grid_entries(self, contents):
yield entry yield entry
''' '''
def _report_history_entries(self, renderer):
for url in traverse_obj(renderer, (
'rows', ..., 'reportHistoryTableRowRenderer', 'cells', ...,
'reportHistoryTableCellRenderer', 'cell', 'reportHistoryTableTextCellRenderer', 'text', 'runs', ...,
'navigationEndpoint', 'commandMetadata', 'webCommandMetadata', 'url')):
yield self.url_result(urljoin('https://www.youtube.com', url), YoutubeIE)
def _extract_entries(self, parent_renderer, continuation_list): def _extract_entries(self, parent_renderer, continuation_list):
# continuation_list is modified in-place with continuation_list = [continuation_token] # continuation_list is modified in-place with continuation_list = [continuation_token]
continuation_list[:] = [None] continuation_list[:] = [None]
@ -4416,12 +4423,16 @@ def _extract_entries(self, parent_renderer, continuation_list):
content, 'itemSectionRenderer', 'musicShelfRenderer', 'musicShelfContinuation', content, 'itemSectionRenderer', 'musicShelfRenderer', 'musicShelfContinuation',
expected_type=dict) expected_type=dict)
if not is_renderer: if not is_renderer:
renderer = content.get('richItemRenderer') if content.get('richItemRenderer'):
if renderer: for entry in self._rich_entries(content['richItemRenderer']):
for entry in self._rich_entries(renderer):
yield entry yield entry
continuation_list[0] = self._extract_continuation(parent_renderer) continuation_list[0] = self._extract_continuation(parent_renderer)
elif content.get('reportHistorySectionRenderer'): # https://www.youtube.com/reporthistory
table = traverse_obj(content, ('reportHistorySectionRenderer', 'table', 'tableRenderer'))
yield from self._report_history_entries(table)
continuation_list[0] = self._extract_continuation(table)
continue continue
isr_contents = try_get(is_renderer, lambda x: x['contents'], list) or [] isr_contents = try_get(is_renderer, lambda x: x['contents'], list) or []
for isr_content in isr_contents: for isr_content in isr_contents:
if not isinstance(isr_content, dict): if not isinstance(isr_content, dict):
@ -4510,7 +4521,8 @@ def _entries(self, tab, item_id, ytcfg, account_syncid, visitor_data):
'playlistVideoRenderer': (self._playlist_entries, 'contents'), 'playlistVideoRenderer': (self._playlist_entries, 'contents'),
'itemSectionRenderer': (extract_entries, 'contents'), # for feeds 'itemSectionRenderer': (extract_entries, 'contents'), # for feeds
'richItemRenderer': (extract_entries, 'contents'), # for hashtag 'richItemRenderer': (extract_entries, 'contents'), # for hashtag
'backstagePostThreadRenderer': (self._post_thread_continuation_entries, 'contents') 'backstagePostThreadRenderer': (self._post_thread_continuation_entries, 'contents'),
'reportHistoryTableRowRenderer': (self._report_history_entries, 'rows'),
} }
on_response_received = dict_get(response, ('onResponseReceivedActions', 'onResponseReceivedEndpoints')) on_response_received = dict_get(response, ('onResponseReceivedActions', 'onResponseReceivedEndpoints'))
continuation_items = try_get( continuation_items = try_get(