[manyvids] Improve extraction (#31172)

* extract all formats from page
* extract description, uploader, views, likes
* downrate previews
* fix tests
* use txt_or_none()

youtube_dl/extractor/manyvids.py

@@ -1,11 +1,16 @@
 # coding: utf-8
 from __future__ import unicode_literals
 
+import re
+
 from .common import InfoExtractor
+from ..compat import compat_str
 from ..utils import (
     determine_ext,
+    extract_attributes,
     int_or_none,
     str_to_int,
+    url_or_none,
     urlencode_postdata,
 )
 
@@ -20,17 +25,20 @@ class ManyVidsIE(InfoExtractor):
             'id': '133957',
             'ext': 'mp4',
             'title': 'everthing about me (Preview)',
+            'uploader': 'ellyxxix',
             'view_count': int,
             'like_count': int,
         },
     }, {
         # full video
         'url': 'https://www.manyvids.com/Video/935718/MY-FACE-REVEAL/',
-        'md5': 'f3e8f7086409e9b470e2643edb96bdcc',
+        'md5': 'bb47bab0e0802c2a60c24ef079dfe60f',
         'info_dict': {
             'id': '935718',
             'ext': 'mp4',
             'title': 'MY FACE REVEAL',
+            'description': 'md5:ec5901d41808b3746fed90face161612',
+            'uploader': 'Sarah Calanthe',
             'view_count': int,
             'like_count': int,
         },
@@ -41,15 +49,43 @@ class ManyVidsIE(InfoExtractor):
 
         webpage = self._download_webpage(url, video_id)
 
-        video_url = self._search_regex(
+        info = self._search_regex(
-            r'data-(?:video-filepath|meta-video)\s*=s*(["\'])(?P<url>(?:(?!\1).)+)\1',
+            r'''(<div\b[^>]*\bid\s*=\s*(['"])pageMetaDetails\2[^>]*>)''',
-            webpage, 'video URL', group='url')
+            webpage, 'meta details', default='')
+        info = extract_attributes(info)
 
-        title = self._html_search_regex(
+        player = self._search_regex(
-            (r'<span[^>]+class=["\']item-title[^>]+>([^<]+)',
+            r'''(<div\b[^>]*\bid\s*=\s*(['"])rmpPlayerStream\2[^>]*>)''',
-             r'<h2[^>]+class=["\']h2 m-0["\'][^>]*>([^<]+)'),
+            webpage, 'player details', default='')
-            webpage, 'title', default=None) or self._html_search_meta(
+        player = extract_attributes(player)
-            'twitter:title', webpage, 'title', fatal=True)
+
+        video_urls_and_ids = (
+            (info.get('data-meta-video'), 'video'),
+            (player.get('data-video-transcoded'), 'transcoded'),
+            (player.get('data-video-filepath'), 'filepath'),
+            (self._og_search_video_url(webpage, secure=False, default=None), 'og_video'),
+        )
+
+        def txt_or_none(s, default=None):
+            return (s.strip() or default) if isinstance(s, compat_str) else default
+
+        uploader = txt_or_none(info.get('data-meta-author'))
+
+        def mung_title(s):
+            if uploader:
+                s = re.sub(r'^\s*%s\s+[|-]' % (re.escape(uploader), ), '', s)
+            return txt_or_none(s)
+
+        title = (
+            mung_title(info.get('data-meta-title'))
+            or self._html_search_regex(
+                (r'<span[^>]+class=["\']item-title[^>]+>([^<]+)',
+                 r'<h2[^>]+class=["\']h2 m-0["\'][^>]*>([^<]+)'),
+                webpage, 'title', default=None)
+            or self._html_search_meta(
+                'twitter:title', webpage, 'title', fatal=True))
+
+        title = re.sub(r'\s*[|-]\s+ManyVids\s*$', '', title) or title
 
         if any(p in webpage for p in ('preview_videos', '_preview.mp4')):
             title += ' (Preview)'
@@ -70,23 +106,56 @@ class ManyVidsIE(InfoExtractor):
                     'X-Requested-With': 'XMLHttpRequest'
                 })
 
-        if determine_ext(video_url) == 'm3u8':
+        formats = []
-            formats = self._extract_m3u8_formats(
+        for v_url, fmt in video_urls_and_ids:
-                video_url, video_id, 'mp4', entry_protocol='m3u8_native',
+            v_url = url_or_none(v_url)
-                m3u8_id='hls')
+            if not v_url:
-        else:
+                continue
-            formats = [{'url': video_url}]
+            if determine_ext(v_url) == 'm3u8':
+                formats.extend(self._extract_m3u8_formats(
+                    v_url, video_id, 'mp4', entry_protocol='m3u8_native',
+                    m3u8_id='hls'))
+            else:
+                formats.append({
+                    'url': v_url,
+                    'format_id': fmt,
+                })
 
-        like_count = int_or_none(self._search_regex(
+        self._remove_duplicate_formats(formats)
-            r'data-likes=["\'](\d+)', webpage, 'like count', default=None))
+
-        view_count = str_to_int(self._html_search_regex(
+        for f in formats:
-            r'(?s)<span[^>]+class="views-wrapper"[^>]*>(.+?)</span', webpage,
+            if f.get('height') is None:
-            'view count', default=None))
+                f['height'] = int_or_none(
+                    self._search_regex(r'_(\d{2,3}[02468])_', f['url'], 'video height', default=None))
+            if '/preview/' in f['url']:
+                f['format_id'] = '_'.join(filter(None, (f.get('format_id'), 'preview')))
+                f['preference'] = -10
+            if 'transcoded' in f['format_id']:
+                f['preference'] = f.get('preference', -1) - 1
+
+        self._sort_formats(formats)
+
+        def get_likes():
+            likes = self._search_regex(
+                r'''(<a\b[^>]*\bdata-id\s*=\s*(['"])%s\2[^>]*>)''' % (video_id, ),
+                webpage, 'likes', default='')
+            likes = extract_attributes(likes)
+            return int_or_none(likes.get('data-likes'))
+
+        def get_views():
+            return str_to_int(self._html_search_regex(
+                r'''(?s)<span\b[^>]*\bclass\s*=["']views-wrapper\b[^>]+>.+?<span\b[^>]+>\s*(\d[\d,.]*)\s*</span>''',
+                webpage, 'view count', default=None))
 
         return {
             'id': video_id,
             'title': title,
-            'view_count': view_count,
-            'like_count': like_count,
             'formats': formats,
+            'description': txt_or_none(info.get('data-meta-description')),
+            'uploader': txt_or_none(info.get('data-meta-author')),
+            'thumbnail': (
+                url_or_none(info.get('data-meta-image'))
+                or url_or_none(player.get('data-video-screenshot'))),
+            'view_count': get_views(),
+            'like_count': get_likes(),
         }