[patreon] fix extraction(closes #14502)(closes #10471)

youtube_dl/extractor/patreon.py

@@ -2,24 +2,30 @@
 from __future__ import unicode_literals
 
 from .common import InfoExtractor
-from ..utils import js_to_json
+from ..utils import (
+    clean_html,
+    determine_ext,
+    int_or_none,
+    parse_iso8601,
+)
 
 
 class PatreonIE(InfoExtractor):
-    _VALID_URL = r'https?://(?:www\.)?patreon\.com/creation\?hid=(?P<id>[^&#]+)'
+    _VALID_URL = r'https?://(?:www\.)?patreon\.com/(?:creation\?hid=|posts/(?:[\w-]+-)?)(?P<id>\d+)'
-    _TESTS = [
+    _TESTS = [{
-        {
         'url': 'http://www.patreon.com/creation?hid=743933',
         'md5': 'e25505eec1053a6e6813b8ed369875cc',
         'info_dict': {
             'id': '743933',
             'ext': 'mp3',
             'title': 'Episode 166: David Smalley of Dogma Debate',
+            'description': 'md5:713b08b772cd6271b9f3906683cfacdf',
             'uploader': 'Cognitive Dissonance Podcast',
             'thumbnail': 're:^https?://.*$',
+            'timestamp': 1406473987,
+            'upload_date': '20140727',
         },
-        },
+    }, {
-        {
         'url': 'http://www.patreon.com/creation?hid=754133',
         'md5': '3eb09345bf44bf60451b8b0b81759d0a',
         'info_dict': {
@@ -29,8 +35,8 @@ class PatreonIE(InfoExtractor):
             'uploader': 'Cognitive Dissonance Podcast',
             'thumbnail': 're:^https?://.*$',
         },
-        },
+        'skip': 'Patron-only content',
-        {
+    }, {
         'url': 'https://www.patreon.com/creation?hid=1682498',
         'info_dict': {
             'id': 'SU4fj_aEMVw',
@@ -46,8 +52,13 @@ class PatreonIE(InfoExtractor):
             'noplaylist': True,
             'skip_download': True,
         }
-        }
+    }, {
-    ]
+        'url': 'https://www.patreon.com/posts/episode-166-of-743933',
+        'only_matching': True,
+    }, {
+        'url': 'https://www.patreon.com/posts/743933',
+        'only_matching': True,
+    }]
 
     # Currently Patreon exposes download URL via hidden CSS, so login is not
     # needed. Keeping this commented for when this inevitably changes.
@@ -78,38 +89,43 @@ class PatreonIE(InfoExtractor):
 
     def _real_extract(self, url):
         video_id = self._match_id(url)
-        webpage = self._download_webpage(url, video_id)
+        post = self._download_json(
-        title = self._og_search_title(webpage).strip()
+            'https://www.patreon.com/api/posts/' + video_id, video_id)
-
+        attributes = post['data']['attributes']
-        attach_fn = self._html_search_regex(
+        title = attributes['title'].strip()
-            r'<div class="attach"><a target="_blank" href="([^"]+)">',
+        image = attributes.get('image') or {}
-            webpage, 'attachment URL', default=None)
+        info = {
-        embed = self._html_search_regex(
-            r'<div[^>]+id="watchCreation"[^>]*>\s*<iframe[^>]+src="([^"]+)"',
-            webpage, 'embedded URL', default=None)
-
-        if attach_fn is not None:
-            video_url = 'http://www.patreon.com' + attach_fn
-            thumbnail = self._og_search_thumbnail(webpage)
-            uploader = self._html_search_regex(
-                r'<strong>(.*?)</strong> is creating', webpage, 'uploader')
-        elif embed is not None:
-            return self.url_result(embed)
-        else:
-            playlist = self._parse_json(self._search_regex(
-                r'(?s)new\s+jPlayerPlaylist\(\s*\{\s*[^}]*},\s*(\[.*?,?\s*\])',
-                webpage, 'playlist JSON'),
-                video_id, transform_source=js_to_json)
-            data = playlist[0]
-            video_url = self._proto_relative_url(data['mp3'])
-            thumbnail = self._proto_relative_url(data.get('cover'))
-            uploader = data.get('artist')
-
-        return {
             'id': video_id,
-            'url': video_url,
-            'ext': 'mp3',
             'title': title,
-            'uploader': uploader,
+            'description': clean_html(attributes.get('content')),
-            'thumbnail': thumbnail,
+            'thumbnail': image.get('large_url') or image.get('url'),
+            'timestamp': parse_iso8601(attributes.get('published_at')),
+            'like_count': int_or_none(attributes.get('like_count')),
+            'comment_count': int_or_none(attributes.get('comment_count')),
         }
+
+        for i in post.get('included', []):
+            i_type = i.get('type')
+            if i_type == 'attachment':
+                attachment_attributes = i.get('attributes') or {}
+                attachment_url = attachment_attributes.get('url')
+                if attachment_url:
+                    info.update({
+                        'url': attachment_url,
+                        'ext': determine_ext(attachment_attributes.get('name'), 'mp3'),
+                    })
+            elif i_type == 'user':
+                user_attributes = i.get('attributes')
+                if user_attributes:
+                    info.update({
+                        'uploader': user_attributes.get('full_name'),
+                        'uploader_url': user_attributes.get('url'),
+                    })
+
+        if not info.get('url'):
+            info.update({
+                '_type': 'url',
+                'url': attributes['embed']['url'],
+            })
+
+        return info