[jamendo] improve extraction
authorRemita Amine <remitamine@gmail.com>
Mon, 4 Nov 2019 14:43:52 +0000 (15:43 +0100)
committerRemita Amine <remitamine@gmail.com>
Mon, 4 Nov 2019 14:43:52 +0000 (15:43 +0100)
- fix album extraction(closes #18564)
- improve metadata extraction(closes #18565)(closes #21379)

youtube_dl/extractor/jamendo.py

index c218276184ab4acf9086404cdb679a74e24f353b..12e21eb6f3ee73b4e80bc508d31e8d9fd7c21f2c 100644 (file)
@@ -1,38 +1,26 @@
 # coding: utf-8
 from __future__ import unicode_literals
 
-import re
+import hashlib
+import random
 
-from ..compat import compat_urlparse
+from ..compat import compat_str
 from .common import InfoExtractor
-from ..utils import parse_duration
-
-
-class JamendoBaseIE(InfoExtractor):
-    def _extract_meta(self, webpage, fatal=True):
-        title = self._og_search_title(
-            webpage, default=None) or self._search_regex(
-            r'<title>([^<]+)', webpage,
-            'title', default=None)
-        if title:
-            title = self._search_regex(
-                r'(.+?)\s*\|\s*Jamendo Music', title, 'title', default=None)
-        if not title:
-            title = self._html_search_meta(
-                'name', webpage, 'title', fatal=fatal)
-        mobj = re.search(r'(.+) - (.+)', title or '')
-        artist, second = mobj.groups() if mobj else [None] * 2
-        return title, artist, second
-
-
-class JamendoIE(JamendoBaseIE):
+from ..utils import (
+    clean_html,
+    int_or_none,
+    try_get,
+)
+
+
+class JamendoIE(InfoExtractor):
     _VALID_URL = r'''(?x)
                     https?://
                         (?:
                             licensing\.jamendo\.com/[^/]+|
                             (?:www\.)?jamendo\.com
                         )
-                        /track/(?P<id>[0-9]+)/(?P<display_id>[^/?#&]+)
+                        /track/(?P<id>[0-9]+)(?:/(?P<display_id>[^/?#&]+))?
                     '''
     _TESTS = [{
         'url': 'https://www.jamendo.com/track/196219/stories-from-emona-i',
@@ -45,7 +33,9 @@ class JamendoIE(JamendoBaseIE):
             'artist': 'Maya Filipič',
             'track': 'Stories from Emona I',
             'duration': 210,
-            'thumbnail': r're:^https?://.*\.jpg'
+            'thumbnail': r're:^https?://.*\.jpg',
+            'timestamp': 1217438117,
+            'upload_date': '20080730',
         }
     }, {
         'url': 'https://licensing.jamendo.com/en/track/1496667/energetic-rock',
@@ -53,15 +43,19 @@ class JamendoIE(JamendoBaseIE):
     }]
 
     def _real_extract(self, url):
-        mobj = self._VALID_URL_RE.match(url)
-        track_id = mobj.group('id')
-        display_id = mobj.group('display_id')
-
-        webpage = self._download_webpage(
-            'https://www.jamendo.com/track/%s/%s' % (track_id, display_id),
-            display_id)
-
-        title, artist, track = self._extract_meta(webpage)
+        track_id, display_id = self._VALID_URL_RE.match(url).groups()
+        webpage = self._download_webpage(url, track_id)
+        models = self._parse_json(self._html_search_regex(
+            r"data-bundled-models='([^']+)",
+            webpage, 'bundled models'), track_id)
+        track = models['track']['models'][0]
+        title = track_name = track['name']
+        get_model = lambda x: try_get(models, lambda y: y[x]['models'][0], dict) or {}
+        artist = get_model('artist')
+        artist_name = artist.get('name')
+        if artist_name:
+            title = '%s - %s' % (artist_name, title)
+        album = get_model('album')
 
         formats = [{
             'url': 'https://%s.jamendo.com/?trackid=%s&format=%s&from=app-97dab294'
@@ -77,31 +71,58 @@ class JamendoIE(JamendoBaseIE):
         ))]
         self._sort_formats(formats)
 
-        thumbnail = self._html_search_meta(
-            'image', webpage, 'thumbnail', fatal=False)
-        duration = parse_duration(self._search_regex(
-            r'<span[^>]+itemprop=["\']duration["\'][^>]+content=["\'](.+?)["\']',
-            webpage, 'duration', fatal=False))
+        urls = []
+        thumbnails = []
+        for _, covers in track.get('cover', {}).items():
+            for cover_id, cover_url in covers.items():
+                if not cover_url or cover_url in urls:
+                    continue
+                urls.append(cover_url)
+                size = int_or_none(cover_id.lstrip('size'))
+                thumbnails.append({
+                    'id': cover_id,
+                    'url': cover_url,
+                    'width': size,
+                    'height': size,
+                })
+
+        tags = []
+        for tag in track.get('tags', []):
+            tag_name = tag.get('name')
+            if not tag_name:
+                continue
+            tags.append(tag_name)
+
+        stats = track.get('stats') or {}
 
         return {
             'id': track_id,
             'display_id': display_id,
-            'thumbnail': thumbnail,
+            'thumbnails': thumbnails,
             'title': title,
-            'duration': duration,
-            'artist': artist,
-            'track': track,
-            'formats': formats
+            'description': track.get('description'),
+            'duration': int_or_none(track.get('duration')),
+            'artist': artist_name,
+            'track': track_name,
+            'album': album.get('name'),
+            'formats': formats,
+            'license': '-'.join(track.get('licenseCC', [])) or None,
+            'timestamp': int_or_none(track.get('dateCreated')),
+            'view_count': int_or_none(stats.get('listenedAll')),
+            'like_count': int_or_none(stats.get('favorited')),
+            'average_rating': int_or_none(stats.get('averageNote')),
+            'tags': tags,
         }
 
 
-class JamendoAlbumIE(JamendoBaseIE):
-    _VALID_URL = r'https?://(?:www\.)?jamendo\.com/album/(?P<id>[0-9]+)/(?P<display_id>[\w-]+)'
+class JamendoAlbumIE(InfoExtractor):
+    _VALID_URL = r'https?://(?:www\.)?jamendo\.com/album/(?P<id>[0-9]+)'
     _TEST = {
         'url': 'https://www.jamendo.com/album/121486/duck-on-cover',
         'info_dict': {
             'id': '121486',
-            'title': 'Shearer - Duck On Cover'
+            'title': 'Duck On Cover',
+            'description': 'md5:c2920eaeef07d7af5b96d7c64daf1239',
         },
         'playlist': [{
             'md5': 'e1a2fcb42bda30dfac990212924149a8',
@@ -111,6 +132,8 @@ class JamendoAlbumIE(JamendoBaseIE):
                 'title': 'Shearer - Warmachine',
                 'artist': 'Shearer',
                 'track': 'Warmachine',
+                'timestamp': 1368089771,
+                'upload_date': '20130509',
             }
         }, {
             'md5': '1f358d7b2f98edfe90fd55dac0799d50',
@@ -120,6 +143,8 @@ class JamendoAlbumIE(JamendoBaseIE):
                 'title': 'Shearer - Without Your Ghost',
                 'artist': 'Shearer',
                 'track': 'Without Your Ghost',
+                'timestamp': 1368089771,
+                'upload_date': '20130509',
             }
         }],
         'params': {
@@ -127,24 +152,35 @@ class JamendoAlbumIE(JamendoBaseIE):
         }
     }
 
+    def _call_api(self, resource, resource_id):
+        path = '/api/%ss' % resource
+        rand = compat_str(random.random())
+        return self._download_json(
+            'https://www.jamendo.com' + path, resource_id, query={
+                'id[]': resource_id,
+            }, headers={
+                'X-Jam-Call': '$%s*%s~' % (hashlib.sha1((path + rand).encode()).hexdigest(), rand)
+            })[0]
+
     def _real_extract(self, url):
-        mobj = self._VALID_URL_RE.match(url)
-        album_id = mobj.group('id')
-
-        webpage = self._download_webpage(url, mobj.group('display_id'))
-
-        title, artist, album = self._extract_meta(webpage, fatal=False)
-
-        entries = [{
-            '_type': 'url_transparent',
-            'url': compat_urlparse.urljoin(url, m.group('path')),
-            'ie_key': JamendoIE.ie_key(),
-            'id': self._search_regex(
-                r'/track/(\d+)', m.group('path'), 'track id', default=None),
-            'artist': artist,
-            'album': album,
-        } for m in re.finditer(
-            r'<a[^>]+href=(["\'])(?P<path>(?:(?!\1).)+)\1[^>]+class=["\'][^>]*js-trackrow-albumpage-link',
-            webpage)]
-
-        return self.playlist_result(entries, album_id, title)
+        album_id = self._match_id(url)
+        album = self._call_api('album', album_id)
+        album_name = album.get('name')
+
+        entries = []
+        for track in album.get('tracks', []):
+            track_id = track.get('id')
+            if not track_id:
+                continue
+            track_id = compat_str(track_id)
+            entries.append({
+                '_type': 'url_transparent',
+                'url': 'https://www.jamendo.com/track/' + track_id,
+                'ie_key': JamendoIE.ie_key(),
+                'id': track_id,
+                'album': album_name,
+            })
+
+        return self.playlist_result(
+            entries, album_id, album_name,
+            clean_html(try_get(album, lambda x: x['description']['en'], compat_str)))