Merge 0f6fb121ba into f2a4983df7

Closes #11526
Authored by: jshumphrey

yt_dlp/extractor/_extractors.py

@@ -256,7 +256,9 @@ from .bilibili import (
     BiliBiliIE,
     BiliBiliPlayerIE,
     BilibiliPlaylistIE,
+    BiliBiliSearchAllIE,
     BiliBiliSearchIE,
+    BiliBiliSearchPageIE,
     BilibiliSeriesListIE,
     BilibiliSpaceAudioIE,
     BilibiliSpaceVideoIE,

yt_dlp/extractor/archiveorg.py

@@ -205,6 +205,26 @@ class ArchiveOrgIE(InfoExtractor):
                 },
             },
         ],
+    }, {
+        # The reviewbody is None for one of the reviews; just need to extract data without crashing
+        'url': 'https://archive.org/details/gd95-04-02.sbd.11622.sbeok.shnf/gd95-04-02d1t04.shn',
+        'info_dict': {
+            'id': 'gd95-04-02.sbd.11622.sbeok.shnf/gd95-04-02d1t04.shn',
+            'ext': 'mp3',
+            'title': 'Stuck Inside of Mobile with the Memphis Blues Again',
+            'creators': ['Grateful Dead'],
+            'duration': 338.31,
+            'track': 'Stuck Inside of Mobile with the Memphis Blues Again',
+            'description': 'md5:764348a470b986f1217ffd38d6ac7b72',
+            'display_id': 'gd95-04-02d1t04.shn',
+            'location': 'Pyramid Arena',
+            'uploader': 'jon@archive.org',
+            'album': '1995-04-02 - Pyramid Arena',
+            'upload_date': '20040519',
+            'track_number': 4,
+            'release_date': '19950402',
+            'timestamp': 1084927901,
+        },
     }]
 
     @staticmethod
@@ -335,7 +355,7 @@ class ArchiveOrgIE(InfoExtractor):
                 info['comments'].append({
                     'id': review.get('review_id'),
                     'author': review.get('reviewer'),
-                    'text': str_or_none(review.get('reviewtitle'), '') + '\n\n' + review.get('reviewbody'),
+                    'text': join_nonempty('reviewtitle', 'reviewbody', from_dict=review, delim='\n\n'),
                     'timestamp': unified_timestamp(review.get('createdate')),
                     'parent': 'root'})
 

yt_dlp/extractor/bilibili.py

@@ -1660,7 +1660,96 @@ class BilibiliCategoryIE(InfoExtractor):
         return self.playlist_result(self._entries(category, subcategory, query), query, query)
 
 
-class BiliBiliSearchIE(SearchInfoExtractor):
+class BiliBiliSearchBaseIE(BilibiliBaseIE):
+    def _extract_search_result(self, result_data):
+        live_room_prefix = 'https://live.bilibili.com/'
+        bili_user_prefix = 'https://space.bilibili.com/'
+        result_type = result_data.get('type')
+        if result_type == 'video':
+            return self.url_result(result_data['arcurl'])
+        elif result_type == 'live_room':
+            return self.url_result(live_room_prefix + str(result_data['roomid']))
+        elif result_type in ['media_ft', 'media_bangumi']:
+            return self.url_result(result_data['url'])
+        elif result_type == 'bili_user':
+            return self.url_result(bili_user_prefix + str(result_data['mid']))
+
+
+class BiliBiliSearchAllIE(SearchInfoExtractor, BiliBiliSearchBaseIE):
+    IE_DESC = 'Bilibili all search'
+    _MAX_RESULTS = 100000
+    _SEARCH_KEY = 'biliallsearch'
+    _TESTS = [{
+        'url': 'biliallsearch3:靡烟 出道一年，我怎么还在等你单推的女人睡觉后开播啊',
+        'playlist_count': 3,
+        'info_dict': {
+            'id': '靡烟 出道一年，我怎么还在等你单推的女人睡觉后开播啊',
+            'title': '靡烟 出道一年，我怎么还在等你单推的女人睡觉后开播啊',
+        },
+        'playlist': [{
+            'info_dict': {
+                'id': 'BV1n44y1Q7sc',
+                'ext': 'mp4',
+                'title': '“出道一年，我怎么还在等你单推的女人睡觉后开播啊？”【一分钟了解靡烟miya】',
+                'timestamp': 1669889987,
+                'upload_date': '20221201',
+                'description': 'md5:43343c0973defff527b5a4b403b4abf9',
+                'tags': list,
+                'uploader': '靡烟miya',
+                'duration': 123.156,
+                'uploader_id': '1958703906',
+                'comment_count': int,
+                'view_count': int,
+                'like_count': int,
+                'thumbnail': r're:^https?://.*\.(jpg|jpeg|png)$',
+                '_old_archive_ids': ['bilibili 988222410_part1'],
+            },
+        }],
+    }, {
+        'url': 'biliallsearch:LOL',
+        'playlist_count': 1,
+        'info_dict': {
+            'id': 'LOL',
+            'title': 'LOL',
+        },
+    }]
+
+    def _search_results(self, query):
+        headers = self.geo_verification_headers()
+        headers['Referer'] = 'https://www.bilibili.com/'
+        page_size = 50
+        if not self._get_cookies('https://api.bilibili.com').get('buvid3'):
+            self._set_cookie('.bilibili.com', 'buvid3', f'{uuid.uuid4()}infoc')
+        for page_num in itertools.count(1):
+            query_params = {
+                'keyword': query,
+                'page': page_num,
+                'dynamic_offset': (page_num - 1) * page_size,
+                'platform': 'pc',
+            }
+            api_url = r'https://api.bilibili.com/x/web-interface/wbi/search/all/v2'
+            try:
+                search_all_result = self._download_json(
+                    api_url, video_id=query, query=self._sign_wbi(query_params, query),
+                    headers=headers,
+                )
+            except ExtractorError as e:
+                if isinstance(e.cause, HTTPError) and e.cause.status == 412:
+                    raise ExtractorError('Request is blocked by server (-412).', expected=True)
+                raise
+            status_code = search_all_result['code']
+            if status_code == -400:
+                raise ExtractorError('Invalid request (-400).', expected=True)
+            result_list = search_all_result['data'].get('result')
+            if not result_list:
+                self.write_debug(f'Response: {search_all_result}')
+                raise ExtractorError(f'Result not found in the response ({status_code}).',
+                                     expected=True)
+            for result_data in traverse_obj(result_list, (..., 'data', ...)):
+                yield self._extract_search_result(result_data)
+
+
+class BiliBiliSearchIE(SearchInfoExtractor, BilibiliBaseIE):
     IE_DESC = 'Bilibili video search'
     _MAX_RESULTS = 100000
     _SEARCH_KEY = 'bilisearch'
@@ -1695,21 +1784,16 @@ class BiliBiliSearchIE(SearchInfoExtractor):
     def _search_results(self, query):
         if not self._get_cookies('https://api.bilibili.com').get('buvid3'):
             self._set_cookie('.bilibili.com', 'buvid3', f'{uuid.uuid4()}infoc')
+        headers = self.geo_verification_headers()
+        headers['Referer'] = 'https://www.bilibili.com/'
         for page_num in itertools.count(1):
             videos = self._download_json(
-                'https://api.bilibili.com/x/web-interface/search/type', query,
-                note=f'Extracting results from page {page_num}', query={
-                    'Search_key': query,
+                'https://api.bilibili.com/x/web-interface/wbi/search/type', query,
+                note=f'Extracting results from page {page_num}', query=self._sign_wbi({
                     'keyword': query,
                     'page': page_num,
-                    'context': '',
-                    'duration': 0,
-                    'tids_2': '',
-                    '__refresh__': 'true',
                     'search_type': 'video',
-                    'tids': 0,
-                    'highlight': 1,
-                })['data'].get('result')
+                }, query), headers=headers)['data'].get('result')
             if not videos:
                 break
             for video in videos:
@@ -2406,3 +2490,126 @@ class BiliLiveIE(InfoExtractor):
                 'Referer': url,
             },
         }
+
+
+class BiliBiliSearchPageIE(BiliBiliSearchBaseIE):
+    IE_DESC = 'Bilibili Search Page URL Extractor'
+    _VALID_URL = r'https?://search\.bilibili\.com/(?P<type>all|video|bangumi|pgc|live|upuser).*'
+    _TESTS = [{
+        'url': r'https://search.bilibili.com/all?keyword=yt+-+dlp+%E4%B8%8B%E8%BD%BD%E5%99%A8',
+        'playlist_count': 36,
+        'info_dict': {
+            'id': 'yt - dlp 下载器',
+            'title': 'yt - dlp 下载器',
+        },
+    }, {
+        'url': r'https://search.bilibili.com/bangumi/?keyword=%E5%AD%A4%E7%8B%AC%E6%91%87%E6%BB%9A&from_source=webtop_search&spm_id_from=333.1007&search_source=5',
+        'playlist_mincount': 1,
+        'info_dict': {
+            'id': '孤独摇滚',
+            'title': '孤独摇滚',
+        },
+        'skip': 'geo-restricted',
+    }, {
+        'url': r'https://search.bilibili.com/video?keyword=%E8%AE%A9%E5%AD%90%E5%BC%B9%E9%A3%9E&from_source=webtop_search&spm_id_from=333.1007&search_source=5&order=dm&duration=4&tids=181&page=3&o=72',
+        'playlist_mincount': 4,
+        'info_dict': {
+            'id': '让子弹飞',
+            'title': '让子弹飞',
+        },
+    }]
+
+    def _real_extract(self, url):
+        headers = self.geo_verification_headers()
+        headers['Referer'] = url
+        entries = []
+        params = parse_qs(url)
+        query = {
+            'platform': 'pc',
+            'page_size': 36,
+        }
+        if not self._get_cookies('https://api.bilibili.com').get('buvid3'):
+            self._set_cookie('.bilibili.com', 'buvid3', f'{uuid.uuid4()}infoc')
+        search_type = self._match_valid_url(url).group('type')
+        raw_playlist_id = traverse_obj(params, ('keyword', 0))
+        if not raw_playlist_id:
+            raise ExtractorError('Please specify the keyword to search for!', expected=True)
+        playlist_id = urllib.parse.unquote_plus(raw_playlist_id)
+        search_type_mapping = {
+            'video': 'video',
+            'bangumi': 'media_bangumi',
+            'pgc': 'media_ft',
+            'live': 'live_room',
+            'upuser': 'bili_user',
+            'all': 'video',  # 'all' search calls video search after page 1
+        }
+        valid_params = [
+            'keyword',
+            'page',
+            'order',
+            'duration',
+            'tids',
+            'search_type',  # Only when searching for live_room or live_user
+            'order_sort',
+            'user_type',
+        ]
+        for valid_param in valid_params:
+            param_value = traverse_obj(params, (valid_param, 0))
+            if param_value is not None:
+                query[valid_param] = param_value
+        page_num = int(query.get('page', 1))
+        param_offset = int_or_none(traverse_obj(params, ('o', 0)))
+        if page_num == 1:
+            query['dynamic_offset'] = 0
+        elif param_offset is not None:
+            query['dynamic_offset'] = param_offset
+        else:
+            query['dynamic_offset'] = query['page_size'] * (page_num - 1)
+        if search_type == 'live' and traverse_obj(params, ('search_type', 0)) == 'live_user':
+            raise ExtractorError('Live users are not downloadable!', expected=True)
+        if search_type == 'all' and page_num == 1:
+            try:
+                search_all_result = self._download_json(
+                    r'https://api.bilibili.com/x/web-interface/wbi/search/all/v2',
+                    video_id=playlist_id, query=self._sign_wbi(query, playlist_id), headers=headers)
+            except ExtractorError as e:
+                if isinstance(e.cause, HTTPError) and e.cause.status == 412:
+                    raise ExtractorError('Request is blocked by server (-412).', expected=True)
+                raise
+            status_code = search_all_result['code']
+            if status_code == -400:
+                raise ExtractorError('Invalid request (-400).', expected=True)
+            result_list = search_all_result['data'].get('result')
+            if not result_list:
+                self.write_debug(f'Response: {search_all_result}')
+                raise ExtractorError(f'Result not found in the response ({status_code}).',
+                                     expected=True)
+
+            entries = [self._extract_search_result(result_data) for result_data in traverse_obj(result_list, (..., 'data', ...))]
+        else:
+            query = {
+                'search_type': search_type_mapping[search_type],
+                **query,  # search_type in type is overridden when specified in url params
+            }
+            try:
+                search_type_result = self._download_json(
+                    r'https://api.bilibili.com/x/web-interface/wbi/search/type',
+                    video_id=playlist_id, query=self._sign_wbi(query, playlist_id), headers=headers,
+                )
+            except ExtractorError as e:
+                if isinstance(e.cause, HTTPError) and e.cause.status == 412:
+                    raise ExtractorError('Request is blocked by server (-412).')
+                raise
+            status_code = search_type_result['code']
+            if status_code == -400:
+                raise ExtractorError('Invalid request (-400).')
+            result_list = search_type_result['data'].get('result')
+            if not result_list:
+                self.write_debug(f'Response: {search_type_result}')
+                raise ExtractorError(
+                    f'Result not found in the response ({status_code}). '
+                    'You might want to try a VPN or a proxy server (with --proxy)', expected=True)
+
+            entries = [self._extract_search_result(result_data) for result_data in result_list]
+
+        return self.playlist_result(entries, playlist_id=playlist_id, playlist_title=playlist_id)

yt_dlp/extractor/facebook.py

@@ -563,13 +563,13 @@ class FacebookIE(InfoExtractor):
                 return extract_video_data(try_get(
                     js_data, lambda x: x['jsmods']['instances'], list) or [])
 
-        def extract_dash_manifest(video, formats):
+        def extract_dash_manifest(vid_data, formats, mpd_url=None):
             dash_manifest = traverse_obj(
-                video, 'dash_manifest', 'playlist', 'dash_manifest_xml_string', expected_type=str)
+                vid_data, 'dash_manifest', 'playlist', 'dash_manifest_xml_string', 'manifest_xml', expected_type=str)
             if dash_manifest:
                 formats.extend(self._parse_mpd_formats(
                     compat_etree_fromstring(urllib.parse.unquote_plus(dash_manifest)),
-                    mpd_url=url_or_none(video.get('dash_manifest_url'))))
+                    mpd_url=url_or_none(video.get('dash_manifest_url')) or mpd_url))
 
         def process_formats(info):
             # Downloads with browser's User-Agent are rate limited. Working around
@@ -619,9 +619,12 @@ class FacebookIE(InfoExtractor):
                         video = video['creation_story']
                         video['owner'] = traverse_obj(video, ('short_form_video_context', 'video_owner'))
                         video.update(reel_info)
-                    fmt_data = traverse_obj(video, ('videoDeliveryLegacyFields', {dict})) or video
+
                     formats = []
                     q = qualities(['sd', 'hd'])
+
+                    # Legacy formats extraction
+                    fmt_data = traverse_obj(video, ('videoDeliveryLegacyFields', {dict})) or video
                     for key, format_id in (('playable_url', 'sd'), ('playable_url_quality_hd', 'hd'),
                                            ('playable_url_dash', ''), ('browser_native_hd_url', 'hd'),
                                            ('browser_native_sd_url', 'sd')):
@@ -629,7 +632,7 @@ class FacebookIE(InfoExtractor):
                         if not playable_url:
                             continue
                         if determine_ext(playable_url) == 'mpd':
-                            formats.extend(self._extract_mpd_formats(playable_url, video_id))
+                            formats.extend(self._extract_mpd_formats(playable_url, video_id, fatal=False))
                         else:
                             formats.append({
                                 'format_id': format_id,
@@ -638,6 +641,28 @@ class FacebookIE(InfoExtractor):
                                 'url': playable_url,
                             })
                     extract_dash_manifest(fmt_data, formats)
+
+                    # New videoDeliveryResponse formats extraction
+                    fmt_data = traverse_obj(video, ('videoDeliveryResponseFragment', 'videoDeliveryResponseResult'))
+                    mpd_urls = traverse_obj(fmt_data, ('dash_manifest_urls', ..., 'manifest_url', {url_or_none}))
+                    dash_manifests = traverse_obj(fmt_data, ('dash_manifests', lambda _, v: v['manifest_xml']))
+                    for idx, dash_manifest in enumerate(dash_manifests):
+                        extract_dash_manifest(dash_manifest, formats, mpd_url=traverse_obj(mpd_urls, idx))
+                    if not dash_manifests:
+                        # Only extract from MPD URLs if the manifests are not already provided
+                        for mpd_url in mpd_urls:
+                            formats.extend(self._extract_mpd_formats(mpd_url, video_id, fatal=False))
+                    for prog_fmt in traverse_obj(fmt_data, ('progressive_urls', lambda _, v: v['progressive_url'])):
+                        format_id = traverse_obj(prog_fmt, ('metadata', 'quality', {str.lower}))
+                        formats.append({
+                            'format_id': format_id,
+                            # sd, hd formats w/o resolution info should be deprioritized below DASH
+                            'quality': q(format_id) - 3,
+                            'url': prog_fmt['progressive_url'],
+                        })
+                    for m3u8_url in traverse_obj(fmt_data, ('hls_playlist_urls', ..., 'hls_playlist_url', {url_or_none})):
+                        formats.extend(self._extract_m3u8_formats(m3u8_url, video_id, 'mp4', fatal=False, m3u8_id='hls'))
+
                     if not formats:
                         # Do not append false positive entry w/o any formats
                         return