[extractor/Ximalaya] Fix extractors (#4339)

Authored by: lockmatrix

yt_dlp/extractor/ximalaya.py

@@ -1,7 +1,7 @@
-import itertools
+import math
-import re
 
 from .common import InfoExtractor
+from ..utils import traverse_obj, try_call, InAdvancePagedList
 
 
 class XimalayaBaseIE(InfoExtractor):
@@ -11,11 +11,10 @@ class XimalayaBaseIE(InfoExtractor):
 class XimalayaIE(XimalayaBaseIE):
     IE_NAME = 'ximalaya'
     IE_DESC = '喜马拉雅FM'
-    _VALID_URL = r'https?://(?:www\.|m\.)?ximalaya\.com/(?P<uid>[0-9]+)/sound/(?P<id>[0-9]+)'
+    _VALID_URL = r'https?://(?:www\.|m\.)?ximalaya\.com/(:?(?P<uid>\d+)/)?sound/(?P<id>[0-9]+)'
-    _USER_URL_FORMAT = '%s://www.ximalaya.com/zhubo/%i/'
     _TESTS = [
         {
-            'url': 'http://www.ximalaya.com/61425525/sound/47740352/',
+            'url': 'http://www.ximalaya.com/sound/47740352/',
             'info_dict': {
                 'id': '47740352',
                 'ext': 'm4a',
@@ -24,19 +23,20 @@ class XimalayaIE(XimalayaBaseIE):
                 'uploader_url': 'http://www.ximalaya.com/zhubo/61425525/',
                 'title': '261.唐诗三百首.卷八.送孟浩然之广陵.李白',
                 'description': "contains:《送孟浩然之广陵》\n作者：李白\n故人西辞黄鹤楼，烟花三月下扬州。\n孤帆远影碧空尽，惟见长江天际流。",
+                'thumbnail': r're:^https?://.*\.jpg',
                 'thumbnails': [
                     {
                         'name': 'cover_url',
-                        'url': r're:^https?://.*\.jpg$',
+                        'url': r're:^https?://.*\.jpg',
                     },
                     {
                         'name': 'cover_url_142',
-                        'url': r're:^https?://.*\.jpg$',
+                        'url': r're:^https?://.*\.jpg',
                         'width': 180,
                         'height': 180
                     }
                 ],
-                'categories': ['renwen', '人文'],
+                'categories': ['人文'],
                 'duration': 93,
                 'view_count': int,
                 'like_count': int,
@@ -52,77 +52,42 @@ class XimalayaIE(XimalayaBaseIE):
                 'uploader_url': 'http://www.ximalaya.com/zhubo/61425525/',
                 'title': '261.唐诗三百首.卷八.送孟浩然之广陵.李白',
                 'description': "contains:《送孟浩然之广陵》\n作者：李白\n故人西辞黄鹤楼，烟花三月下扬州。\n孤帆远影碧空尽，惟见长江天际流。",
+                'thumbnail': r're:^https?://.*\.jpg',
                 'thumbnails': [
                     {
                         'name': 'cover_url',
-                        'url': r're:^https?://.*\.jpg$',
+                        'url': r're:^https?://.*\.jpg',
                     },
                     {
                         'name': 'cover_url_142',
-                        'url': r're:^https?://.*\.jpg$',
+                        'url': r're:^https?://.*\.jpg',
                         'width': 180,
                         'height': 180
                     }
                 ],
-                'categories': ['renwen', '人文'],
+                'categories': ['人文'],
                 'duration': 93,
                 'view_count': int,
                 'like_count': int,
             }
-        },
-        {
-            'url': 'https://www.ximalaya.com/11045267/sound/15705996/',
-            'info_dict': {
-                'id': '15705996',
-                'ext': 'm4a',
-                'uploader': '李延隆老师',
-                'uploader_id': 11045267,
-                'uploader_url': 'https://www.ximalaya.com/zhubo/11045267/',
-                'title': 'Lesson 1 Excuse me!',
-                'description': "contains:Listen to the tape then answer\xa0this question. Whose handbag is it?\n"
-                               "听录音，然后回答问题，这是谁的手袋？",
-                'thumbnails': [
-                    {
-                        'name': 'cover_url',
-                        'url': r're:^https?://.*\.jpg$',
-                    },
-                    {
-                        'name': 'cover_url_142',
-                        'url': r're:^https?://.*\.jpg$',
-                        'width': 180,
-                        'height': 180
         }
-                ],
-                'categories': ['train', '外语'],
-                'duration': 40,
-                'view_count': int,
-                'like_count': int,
-            }
-        },
     ]
 
     def _real_extract(self, url):
-
-        is_m = 'm.ximalaya' in url
         scheme = 'https' if url.startswith('https') else 'http'
 
         audio_id = self._match_id(url)
-        webpage = self._download_webpage(url, audio_id,
-                                         note='Download sound page for %s' % audio_id,
-                                         errnote='Unable to get sound page')
-
         audio_info_file = '%s://m.ximalaya.com/tracks/%s.json' % (scheme, audio_id)
         audio_info = self._download_json(audio_info_file, audio_id,
                                          'Downloading info json %s' % audio_info_file,
                                          'Unable to download info file')
 
-        formats = []
+        formats = [{
-        for bps, k in (('24k', 'play_path_32'), ('64k', 'play_path_64')):
+            'format_id': f'{bps}k',
-            if audio_info.get(k):
-                formats.append({
-                    'format_id': bps,
             'url': audio_info[k],
-                })
+            'abr': bps,
+            'vcodec': 'none'
+        } for bps, k in ((24, 'play_path_32'), (64, 'play_path_64')) if audio_info.get(k)]
 
         thumbnails = []
         for k in audio_info.keys():
@@ -136,30 +101,18 @@ def _real_extract(self, url):
 
         audio_uploader_id = audio_info.get('uid')
 
-        if is_m:
+        audio_description = try_call(
-            audio_description = self._html_search_regex(r'(?s)<section\s+class=["\']content[^>]+>(.+?)</section>',
+            lambda: audio_info['intro'].replace('\r\n\r\n\r\n ', '\n').replace('\r\n', '\n'))
-                                                        webpage, 'audio_description', fatal=False)
-        else:
-            audio_description = self._html_search_regex(r'(?s)<div\s+class=["\']rich_intro[^>]*>(.+?</article>)',
-                                                        webpage, 'audio_description', fatal=False)
-
-        if not audio_description:
-            audio_description_file = '%s://www.ximalaya.com/sounds/%s/rich_intro' % (scheme, audio_id)
-            audio_description = self._download_webpage(audio_description_file, audio_id,
-                                                       note='Downloading description file %s' % audio_description_file,
-                                                       errnote='Unable to download descrip file',
-                                                       fatal=False)
-            audio_description = audio_description.strip() if audio_description else None
 
         return {
             'id': audio_id,
             'uploader': audio_info.get('nickname'),
             'uploader_id': audio_uploader_id,
-            'uploader_url': self._USER_URL_FORMAT % (scheme, audio_uploader_id) if audio_uploader_id else None,
+            'uploader_url': f'{scheme}://www.ximalaya.com/zhubo/{audio_uploader_id}/' if audio_uploader_id else None,
             'title': audio_info['title'],
             'thumbnails': thumbnails,
             'description': audio_description,
-            'categories': list(filter(None, (audio_info.get('category_name'), audio_info.get('category_title')))),
+            'categories': list(filter(None, [audio_info.get('category_name')])),
             'duration': audio_info.get('duration'),
             'view_count': audio_info.get('play_count'),
             'like_count': audio_info.get('favorites_count'),
@@ -170,60 +123,38 @@ def _real_extract(self, url):
 class XimalayaAlbumIE(XimalayaBaseIE):
     IE_NAME = 'ximalaya:album'
     IE_DESC = '喜马拉雅FM 专辑'
-    _VALID_URL = r'https?://(?:www\.|m\.)?ximalaya\.com/(?P<uid>[0-9]+)/album/(?P<id>[0-9]+)'
+    _VALID_URL = r'https?://(?:www\.|m\.)?ximalaya\.com/\d+/album/(?P<id>[0-9]+)'
-    _TEMPLATE_URL = '%s://www.ximalaya.com/%s/album/%s/'
-    _BASE_URL_TEMPL = '%s://www.ximalaya.com%s'
-    _LIST_VIDEO_RE = r'<a[^>]+?href="(?P<url>/%s/sound/(?P<id>\d+)/?)"[^>]+?title="(?P<title>[^>]+)">'
     _TESTS = [{
         'url': 'http://www.ximalaya.com/61425525/album/5534601/',
         'info_dict': {
             'title': '唐诗三百首（含赏析）',
             'id': '5534601',
         },
-        'playlist_count': 312,
+        'playlist_mincount': 323,
-    }, {
+    }]
-        'url': 'http://m.ximalaya.com/61425525/album/5534601',
-        'info_dict': {
-            'title': '唐诗三百首（含赏析）',
-            'id': '5534601',
-        },
-        'playlist_count': 312,
-    },
-    ]
 
     def _real_extract(self, url):
-        self.scheme = scheme = 'https' if url.startswith('https') else 'http'
+        playlist_id = self._match_id(url)
 
-        mobj = self._match_valid_url(url)
+        first_page = self._fetch_page(playlist_id, 1)
-        uid, playlist_id = mobj.group('uid'), mobj.group('id')
+        page_count = math.ceil(first_page['trackTotalCount'] / first_page['pageSize'])
 
-        webpage = self._download_webpage(self._TEMPLATE_URL % (scheme, uid, playlist_id), playlist_id,
+        entries = InAdvancePagedList(
-                                         note='Download album page for %s' % playlist_id,
+            lambda idx: self._get_entries(self._fetch_page(playlist_id, idx + 1) if idx else first_page),
-                                         errnote='Unable to get album info')
+            page_count, first_page['pageSize'])
 
-        title = self._html_search_regex(r'detailContent_title[^>]*><h1(?:[^>]+)?>([^<]+)</h1>',
+        title = traverse_obj(first_page, ('tracks', 0, 'albumTitle'), expected_type=str)
-                                        webpage, 'title', fatal=False)
 
-        return self.playlist_result(self._entries(webpage, playlist_id, uid), playlist_id, title)
+        return self.playlist_result(entries, playlist_id, title)
 
-    def _entries(self, page, playlist_id, uid):
+    def _fetch_page(self, playlist_id, page_idx):
-        html = page
+        return self._download_json(
-        for page_num in itertools.count(1):
+            'https://www.ximalaya.com/revision/album/v1/getTracksList',
-            for entry in self._process_page(html, uid):
+            playlist_id, note=f'Downloading tracks list page {page_idx}',
-                yield entry
+            query={'albumId': playlist_id, 'pageNum': page_idx, 'sort': 1})['data']
 
-            next_url = self._search_regex(r'<a\s+href=(["\'])(?P<more>[\S]+)\1[^>]+rel=(["\'])next\3',
+    def _get_entries(self, page_data):
-                                          html, 'list_next_url', default=None, group='more')
+        for e in page_data['tracks']:
-            if not next_url:
+            yield self.url_result(
-                break
+                self._proto_relative_url(f'//www.ximalaya.com{e["url"]}'),
-
+                XimalayaIE, e.get('trackId'), e.get('title'))
-            next_full_url = self._BASE_URL_TEMPL % (self.scheme, next_url)
-            html = self._download_webpage(next_full_url, playlist_id)
-
-    def _process_page(self, html, uid):
-        find_from = html.index('album_soundlist')
-        for mobj in re.finditer(self._LIST_VIDEO_RE % uid, html[find_from:]):
-            yield self.url_result(self._BASE_URL_TEMPL % (self.scheme, mobj.group('url')),
-                                  XimalayaIE.ie_key(),
-                                  mobj.group('id'),
-                                  mobj.group('title'))