Merge 311dc3d438 into f2a4983df7

Closes #11526
Authored by: jshumphrey

yt_dlp/extractor/_extractors.py

@@ -2083,6 +2083,10 @@ from .theplatform import (
 from .thestar import TheStarIE
 from .thesun import TheSunIE
 from .theweatherchannel import TheWeatherChannelIE
+from .thirtydaysinger import (
+    ThirtyDaySingerIE,
+    ThirtyDaySingerPlaylistIE
+)
 from .thisamericanlife import ThisAmericanLifeIE
 from .thisoldhouse import ThisOldHouseIE
 from .thisvid import (

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/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

yt_dlp/extractor/thirtydaysinger.py

@@ -0,0 +1,107 @@
+import re
+
+from .wistia import WistiaBaseIE
+from ..utils import (
+    clean_html,
+    get_elements_html_by_class
+)
+
+
+class ThirtyDaySingerBase(WistiaBaseIE):
+    _INDEX_EXTRACTION_RE = r'/tutorial/[\w-]+/(?P<index>[\w-]+)'
+
+    def _extract_for_url(self, url):
+        lesson_index = re.search(self._INDEX_EXTRACTION_RE, url).group('index')
+        webpage = self._download_webpage(url, lesson_index)
+        match = next(self._extract_wistia_async_embed(webpage))
+        embed_config = self._download_embed_config('medias', match.group('id'), url)
+
+        embed_infojson = self._extract_media(embed_config)
+        webpage_infojson = self._extract_webpage_data(webpage)
+
+        return {**embed_infojson, **webpage_infojson}
+
+    def _extract_webpage_data(self, webpage):
+        title = self._html_search_regex(r'<h1>([^<]+)</h1>', webpage, 'title')
+        fallback_title = self._html_extract_title(webpage)
+        description = self._html_search_meta('description', webpage, fatal=False)
+
+        return {
+            'title': title or fallback_title,
+            'description': clean_html(self._format_html_list(description))
+        }
+
+    # The site makes extensive use of HTML lists for formatting and `clean_html`
+    # doesn't handle them well. This is needed to keep lists readable.
+    def _format_html_list(self, html):
+        replacements = {
+            '<ul>': '<br>',
+            '</ul': '<br>',
+            '<li>': '<br>- ',
+            '</li>': ''
+        }
+
+        for k, v in replacements.items():
+            html = html.replace(k, v)
+
+        return html
+
+
+class ThirtyDaySingerIE(ThirtyDaySingerBase):
+    _VALID_URL = r'https?://www\.30daysinger\.com/tutorial/[\w-]+/[\w-]+'
+
+    _TESTS = [{
+        'url': 'https://www.30daysinger.com/tutorial/30-day-beginner-course-with-jonathan-estabrooks/1',
+        'md5': '56bb11529b9777899b27b599d4b16cf6',
+        'info_dict': {
+            'id': 'tegd38l3d5',
+            'ext': 'mp4',
+            'thumbnail': 'http://embed.wistia.com/deliveries/c26a85cb98e32efa8a5e12a0576e63355af66230.jpg',
+            'upload_date': '20190608',
+            'description': 'md5:d3291de8988be57b1d3e411126ba4d33',
+            'duration': 344.22,
+            'timestamp': 1559952526,
+            'title': 'Welcome to 30 Day Singer'
+        }
+    }]
+
+    def _real_extract(self, url):
+        return self._extract_for_url(url)
+
+
+class ThirtyDaySingerPlaylistIE(ThirtyDaySingerBase):
+    _URI_BASE = 'https://www.30daysinger.com'
+    _VALID_URL = r'https?://www\.30daysinger\.com/tutorial/(?P<playlist_id>[\w-]+)/?(?:$|[#?])'
+
+    _TESTS = [{
+        'url': 'https://www.30daysinger.com/tutorial/30-day-beginner-course-with-jonathan-estabrooks',
+        'info_dict': {
+            'id': '30-day-beginner-course-with-jonathan-estabrooks',
+            'description': 'md5:8cf6d6c7c377895653c9cde9dfc4104f',
+            'title': '30 Day Beginner Course with Jonathan Estabrooks',
+        },
+        'playlist_count': 1,
+        'expected_warnings': ['This video is for premium members only']
+    }]
+
+    def _real_extract(self, url):
+        playlist_id = self._match_valid_url(url).group('playlist_id')
+        webpage = self._download_webpage(url, playlist_id)
+        playlist_attrs = self._extract_webpage_data(webpage)
+
+        entries = []
+        for html_element in get_elements_html_by_class('playlist-item-link', webpage):
+            href = self._search_regex(r'href="([^"]+)"', html_element, 'href')
+
+            if not href:
+                continue
+            # Often _some_ content is free so we should still download that but warn the user
+            # when we encounter premium content.
+            # NOTE: this only applies to the playlist extractor, not the single video extractor
+            if 'upgrade' in href:
+                self.report_warning('This video is for premium members only')
+                continue
+
+            entries.append(self._extract_for_url(self._URI_BASE + href))
+
+        return self.playlist_result(entries, playlist_id, playlist_attrs['title'], playlist_attrs['description'])