Merge 0d26f152b0 into c699bafc50

Closes #11537

Authored by: bashonly

test/test_traversal.py

@@ -481,7 +481,7 @@ class TestTraversalHelpers:
             'id': 'name',
             'data': 'content',
             'url': 'url',
-        }, all, {subs_list_to_dict}]) == {
+        }, all, {subs_list_to_dict(lang=None)}]) == {
             'de': [{'url': 'https://example.com/subs/de.ass'}],
             'en': [{'data': 'content'}],
         }, 'subs with mandatory items missing should be filtered'
@@ -507,6 +507,54 @@ class TestTraversalHelpers:
             {'url': 'https://example.com/subs/en1', 'ext': 'ext'},
             {'url': 'https://example.com/subs/en2', 'ext': 'ext'},
         ]}, '`quality` key should sort subtitle list accordingly'
+        assert traverse_obj([
+            {'name': 'de', 'url': 'https://example.com/subs/de.ass'},
+            {'name': 'de'},
+            {'name': 'en', 'content': 'content'},
+            {'url': 'https://example.com/subs/en'},
+        ], [..., {
+            'id': 'name',
+            'url': 'url',
+            'data': 'content',
+        }, all, {subs_list_to_dict(lang='en')}]) == {
+            'de': [{'url': 'https://example.com/subs/de.ass'}],
+            'en': [
+                {'data': 'content'},
+                {'url': 'https://example.com/subs/en'},
+            ],
+        }, 'optionally provided lang should be used if no id available'
+        assert traverse_obj([
+            {'name': 1, 'url': 'https://example.com/subs/de1'},
+            {'name': {}, 'url': 'https://example.com/subs/de2'},
+            {'name': 'de', 'ext': 1, 'url': 'https://example.com/subs/de3'},
+            {'name': 'de', 'ext': {}, 'url': 'https://example.com/subs/de4'},
+        ], [..., {
+            'id': 'name',
+            'url': 'url',
+            'ext': 'ext',
+        }, all, {subs_list_to_dict(lang=None)}]) == {
+            'de': [
+                {'url': 'https://example.com/subs/de3'},
+                {'url': 'https://example.com/subs/de4'},
+            ],
+        }, 'non str types should be ignored for id and ext'
+        assert traverse_obj([
+            {'name': 1, 'url': 'https://example.com/subs/de1'},
+            {'name': {}, 'url': 'https://example.com/subs/de2'},
+            {'name': 'de', 'ext': 1, 'url': 'https://example.com/subs/de3'},
+            {'name': 'de', 'ext': {}, 'url': 'https://example.com/subs/de4'},
+        ], [..., {
+            'id': 'name',
+            'url': 'url',
+            'ext': 'ext',
+        }, all, {subs_list_to_dict(lang='de')}]) == {
+            'de': [
+                {'url': 'https://example.com/subs/de1'},
+                {'url': 'https://example.com/subs/de2'},
+                {'url': 'https://example.com/subs/de3'},
+                {'url': 'https://example.com/subs/de4'},
+            ],
+        }, 'non str types should be replaced by default id'
 
     def test_trim_str(self):
         with pytest.raises(TypeError):
@@ -525,7 +573,7 @@ class TestTraversalHelpers:
     def test_unpack(self):
         assert unpack(lambda *x: ''.join(map(str, x)))([1, 2, 3]) == '123'
         assert unpack(join_nonempty)([1, 2, 3]) == '1-2-3'
-        assert unpack(join_nonempty(delim=' '))([1, 2, 3]) == '1 2 3'
+        assert unpack(join_nonempty, delim=' ')([1, 2, 3]) == '1 2 3'
         with pytest.raises(TypeError):
             unpack(join_nonempty)()
         with pytest.raises(TypeError):

test/test_utils.py

@@ -72,7 +72,6 @@ from yt_dlp.utils import (
     intlist_to_bytes,
     iri_to_uri,
     is_html,
-    join_nonempty,
     js_to_json,
     limit_length,
     locked_file,
@@ -2158,10 +2157,6 @@ Line 1
         assert int_or_none(v=10) == 10, 'keyword passed positional should call function'
         assert int_or_none(scale=0.1)(10) == 100, 'call after partial application should call the function'
 
-        assert callable(join_nonempty(delim=', ')), 'varargs positional should apply partially'
-        assert callable(join_nonempty()), 'varargs positional should apply partially'
-        assert join_nonempty(None, delim=', ') == '', 'passed varargs should call the function'
-
 
 if __name__ == '__main__':
     unittest.main()

yt_dlp/YoutubeDL.py

@@ -4381,7 +4381,9 @@ class YoutubeDL:
             return None
 
         for idx, t in list(enumerate(thumbnails))[::-1]:
-            thumb_ext = (f'{t["id"]}.' if multiple else '') + determine_ext(t['url'], 'jpg')
+            thumb_ext = t.get('ext') or determine_ext(t['url'], 'jpg')
+            if multiple:
+                thumb_ext = f'{t["id"]}.{thumb_ext}'
             thumb_display_id = f'{label} thumbnail {t["id"]}'
             thumb_filename = replace_extension(filename, thumb_ext, info_dict.get('ext'))
             thumb_filename_final = replace_extension(thumb_filename_base, thumb_ext, info_dict.get('ext'))

yt_dlp/extractor/_extractors.py

@@ -2252,6 +2252,10 @@ from .ufctv import (
 from .ukcolumn import UkColumnIE
 from .uktvplay import UKTVPlayIE
 from .umg import UMGDeIE
+from .underline import (
+    UnderlinePosterIE,
+    UnderlinePosterListIE,
+)
 from .unistra import UnistraIE
 from .unity import UnityIE
 from .unsupported import (

yt_dlp/extractor/afreecatv.py

@@ -66,6 +66,14 @@ class AfreecaTVBaseIE(InfoExtractor):
             extensions={'legacy_ssl': True}), display_id,
             'Downloading API JSON', 'Unable to download API JSON')
 
+    @staticmethod
+    def _fixup_thumb(thumb_url):
+        if not url_or_none(thumb_url):
+            return None
+        # Core would determine_ext as 'php' from the url, so we need to provide the real ext
+        # See: https://github.com/yt-dlp/yt-dlp/issues/11537
+        return [{'url': thumb_url, 'ext': 'jpg'}]
+
 
 class AfreecaTVIE(AfreecaTVBaseIE):
     IE_NAME = 'soop'
@@ -155,7 +163,7 @@ class AfreecaTVIE(AfreecaTVBaseIE):
             'uploader': ('writer_nick', {str}),
             'uploader_id': ('bj_id', {str}),
             'duration': ('total_file_duration', {int_or_none(scale=1000)}),
-            'thumbnail': ('thumb', {url_or_none}),
+            'thumbnails': ('thumb', {self._fixup_thumb}),
         })
 
         entries = []
@@ -226,8 +234,7 @@ class AfreecaTVCatchStoryIE(AfreecaTVBaseIE):
 
         return self.playlist_result(self._entries(data), video_id)
 
-    @staticmethod
-    def _entries(data):
+    def _entries(self, data):
         # 'files' is always a list with 1 element
         yield from traverse_obj(data, (
             'data', lambda _, v: v['story_type'] == 'catch',
@@ -238,7 +245,7 @@ class AfreecaTVCatchStoryIE(AfreecaTVBaseIE):
                 'title': ('title', {str}),
                 'uploader': ('writer_nick', {str}),
                 'uploader_id': ('writer_id', {str}),
-                'thumbnail': ('thumb', {url_or_none}),
+                'thumbnails': ('thumb', {self._fixup_thumb}),
                 'timestamp': ('write_timestamp', {int_or_none}),
             }))
 

yt_dlp/extractor/common.py

@@ -279,6 +279,7 @@ class InfoExtractor:
     thumbnails:     A list of dictionaries, with the following entries:
                         * "id" (optional, string) - Thumbnail format ID
                         * "url"
+                        * "ext" (optional, string) - actual image extension if not given in URL
                         * "preference" (optional, int) - quality of the image
                         * "width" (optional, int)
                         * "height" (optional, int)

yt_dlp/extractor/underline.py

@@ -0,0 +1,123 @@
+import functools
+
+from .common import InfoExtractor
+from ..utils import (
+    OnDemandPagedList,
+    filter_dict,
+    update_url,
+    url_or_none,
+)
+from ..utils.traversal import traverse_obj
+
+
+class UnderlinePosterIE(InfoExtractor):
+    _VALID_URL = r'https://(?:www\.)?underline\.io/events/\d+/posters/\d+/poster/(?P<id>\d+)-[\w-]+/?(?:[#?]|$)'
+
+    _TESTS = [{
+        # Video and PPT
+        'url': 'https://underline.io/events/342/posters/12863/poster/66466-towards-a-general-purpose-machine-translation-system-for-sranantongo',
+        'info_dict': {
+            'id': '66466',
+            'ext': 'mp4',
+            'title': 'Towards a general purpose machine translation system for Sranantongo',
+            'thumbnail': 'https://assets.underline.io/lecture/66466/poster_document_thumbnail_extract/2a8d3abd5a5edbf0117d8d9bdf018e6d.jpg',
+            'live_status': 'not_live',
+        },
+    }, {
+        # PPT only
+        'url': 'https://underline.io/events/342/posters/12863/poster/66459-low-resourced-multilingual-neural-machine-translation-for-ometo-english',
+        'info_dict': {
+            'id': '66459',
+            'title': 'Low Resourced Multilingual Neural Machine Translation for Ometo-English',
+            'thumbnail': 'https://assets.underline.io/lecture/66459/poster_document_thumbnail_extract/69a3fd48e3bb137ddcf3099db637d184.jpg',
+            'live_status': 'not_live',
+        },
+        'params': {
+            'skip_download': True,
+            'ignore_no_formats_error': True,
+        },
+        'expected_warnings': ['No video formats found!', 'Requested format is not available'],
+    }, {
+        'url': 'https://www.underline.io/events/342/posters/12863/poster/66463-mbti-personality-prediction-approach-on-persian-twitter/?tab=video',
+        'only_matching': True,
+    }]
+
+    def _real_extract(self, url):
+        video_id = self._match_id(url)
+        webpage = self._download_webpage(url, video_id)
+
+        data = self._search_nextjs_data(webpage, video_id)['props']['pageProps']['fallback']
+
+        m3u8_url = traverse_obj(data, (..., 'data', 'attributes', 'playlist', {url_or_none}), get_all=False)
+        subtitle_urls = filter_dict(dict(traverse_obj(data, (
+            ..., 'data', lambda _, v: v['type'] == 'transcripts', {lambda x: (
+                x['relationships']['language']['data']['id'],
+                url_or_none(x['attributes']['subtitleUrl']),
+            )}))))
+
+        return {
+            'id': video_id,
+            'title': traverse_obj(data, (..., 'data', 'attributes', 'title', {str}), get_all=False),
+            'thumbnail': traverse_obj(data, (
+                ..., 'data', 'attributes', 'originalPosterDocumentThumbnailExtractUrl', {url_or_none}), get_all=False),
+            'formats': self._extract_m3u8_formats(m3u8_url, video_id) if m3u8_url else [],
+            'subtitles': filter_dict(dict(traverse_obj(data, (
+                ..., 'included', lambda _, v: v['type'] == 'transcript_languages', {lambda x: (
+                    x['attributes']['locale'],
+                    [{'url': subtitle_urls[x['id']]}],
+                )})))),
+            'live_status': 'not_live',
+        }
+
+
+class UnderlinePosterListIE(InfoExtractor):
+    _VALID_URL = r'https://(?:www\.)?underline\.io/events/(?P<event_id>\d+)/posters/?\?(?:[^&]+&)*?eventSessionId=(?P<event_session_id>\d+)(?:&|#)?'
+    _HEADERS = {'Accept': 'application/vnd.api+json'}
+    _PAGE_SIZE = 16
+
+    _TESTS = [{
+        'url': 'https://underline.io/events/342/posters?eventSessionId=12863',
+        'info_dict': {
+            'id': '12863',
+            'title': 'W13 (WiNLP) The Sixth Widening NLP Workshop',
+            'description': r're:.+Isidora Tourni and Surangika Ranathunga.+',
+        },
+        'playlist_mincount': 10,
+    }, {
+        'url': 'https://www.underline.io/events/342/posters?searchGroup=lecture&eventSessionId=12860',
+        'only_matching': True,
+    }]
+
+    def _fetch_paged_channel_video_list(self, list_url, event_id, event_session_id, page):
+        page_num = page + 1
+        video_url_tmpl = f'https://underline.io/events/{event_id}/posters/{event_session_id}/poster/{{id}}-{{slug}}'
+
+        page_info = self._download_json(update_url(
+            list_url, query_update={'page[number]': page_num}), event_session_id, headers=self._HEADERS,
+            note=f'Downloading list info (page {page_num})', errnote=f'Failed to download list info (page {page_num})')
+
+        yield from traverse_obj(page_info, (
+            'data', ..., {lambda x: video_url_tmpl.format(id=x['id'], slug=x['attributes']['slug'])}, {self.url_result}))
+
+    def _real_extract(self, url):
+        event_id, event_session_id = self._match_valid_url(url).groups()
+
+        event_info = self._download_json(
+            f'https://app.underline.io/api/v1/thin_event_sessions/{event_session_id}', event_session_id, headers=self._HEADERS,
+            note='Downloading event session info', errnote='Failed to download event session info')
+
+        return self.playlist_result(OnDemandPagedList(functools.partial(
+            self._fetch_paged_channel_video_list,
+            update_url('https://app.underline.io/api/v1/thin_lectures', query_update={
+                'filter[poster_lecture]': 'true',
+                'filter[published]': 'all',
+                'sort': 'held_at,id',
+                'include': 'sorted_profiles,event_session,tag,event',
+                'filter[scope]': 'all',
+                'filter[event_id]': event_id,
+                'filter[event_session_id]': event_session_id,
+                'page[size]': self._PAGE_SIZE,
+            }), event_id, event_session_id), self._PAGE_SIZE),
+            event_session_id,
+            traverse_obj(event_info, ('data', 'attributes', 'name'), get_all=False),
+            traverse_obj(event_info, ('data', 'attributes', 'description'), get_all=False))

yt_dlp/utils/_utils.py

@@ -216,7 +216,7 @@ def partial_application(func):
     sig = inspect.signature(func)
     required_args = [
         param.name for param in sig.parameters.values()
-        if param.kind in (inspect.Parameter.POSITIONAL_ONLY, inspect.Parameter.POSITIONAL_OR_KEYWORD, inspect.Parameter.VAR_POSITIONAL)
+        if param.kind in (inspect.Parameter.POSITIONAL_ONLY, inspect.Parameter.POSITIONAL_OR_KEYWORD)
         if param.default is inspect.Parameter.empty
     ]
 
@@ -4837,7 +4837,6 @@ def number_of_digits(number):
     return len('%d' % number)
 
 
-@partial_application
 def join_nonempty(*values, delim='-', from_dict=None):
     if from_dict is not None:
         values = (traversal.traverse_obj(from_dict, variadic(v)) for v in values)

yt_dlp/utils/traversal.py

@@ -332,14 +332,14 @@ class _RequiredError(ExtractorError):
 
 
 @typing.overload
-def subs_list_to_dict(*, ext: str | None = None) -> collections.abc.Callable[[list[dict]], dict[str, list[dict]]]: ...
+def subs_list_to_dict(*, lang: str | None = 'und', ext: str | None = None) -> collections.abc.Callable[[list[dict]], dict[str, list[dict]]]: ...
 
 
 @typing.overload
-def subs_list_to_dict(subs: list[dict] | None, /, *, ext: str | None = None) -> dict[str, list[dict]]: ...
+def subs_list_to_dict(subs: list[dict] | None, /, *, lang: str | None = 'und', ext: str | None = None) -> dict[str, list[dict]]: ...
 
 
-def subs_list_to_dict(subs: list[dict] | None = None, /, *, ext=None):
+def subs_list_to_dict(subs: list[dict] | None = None, /, *, lang='und', ext=None):
     """
     Convert subtitles from a traversal into a subtitle dict.
     The path should have an `all` immediately before this function.
@@ -352,7 +352,7 @@ def subs_list_to_dict(subs: list[dict] | None = None, /, *, ext=None):
     `quality`  The sort order for each subtitle
     """
     if subs is None:
-        return functools.partial(subs_list_to_dict, ext=ext)
+        return functools.partial(subs_list_to_dict, lang=lang, ext=ext)
 
     result = collections.defaultdict(list)
 
@@ -360,10 +360,16 @@ def subs_list_to_dict(subs: list[dict] | None = None, /, *, ext=None):
         if not url_or_none(sub.get('url')) and not sub.get('data'):
             continue
         sub_id = sub.pop('id', None)
-        if sub_id is None:
-            continue
-        if ext is not None and not sub.get('ext'):
-            sub['ext'] = ext
+        if not isinstance(sub_id, str):
+            if not lang:
+                continue
+            sub_id = lang
+        sub_ext = sub.get('ext')
+        if not isinstance(sub_ext, str):
+            if not ext:
+                sub.pop('ext', None)
+            else:
+                sub['ext'] = ext
         result[sub_id].append(sub)
     result = dict(result)
 
@@ -452,9 +458,9 @@ def trim_str(*, start=None, end=None):
     return trim
 
 
-def unpack(func):
+def unpack(func, **kwargs):
     @functools.wraps(func)
-    def inner(items, **kwargs):
+    def inner(items):
         return func(*items, **kwargs)
 
     return inner