Code review changes

youtube_dl/extractor/myvideoge.py

@@ -1,9 +1,19 @@
 # coding: utf-8
 from __future__ import unicode_literals
 
-from .common import InfoExtractor
+import re
-from ..utils import js_to_json
 
+from .common import InfoExtractor
+from ..utils import (
+    clean_html,
+    get_element_by_id,
+    get_element_by_class,
+    int_or_none,
+    js_to_json,
+    MONTH_NAMES,
+    qualities,
+    unified_strdate,
+)
 
 class MyVideoGeIE(InfoExtractor):
     _VALID_URL = r'https?://(?:www\.)?myvideo\.ge/v/(?P<id>[0-9]+)'
@@ -14,37 +24,51 @@ class MyVideoGeIE(InfoExtractor):
             'id': '3941048',
             'ext': 'mp4',
             'title': 'The best prikol',
+            'upload_date': '20200611',
             'thumbnail': r're:^https?://.*\.jpg$',
-            'uploader': 'md5:d72addd357b0dd914e704781f7f777d8',
+            'uploader': 'chixa33',
-            'description': 'md5:5c0371f540f5888d603ebfedd46b6df3'
+            'description': 'md5:5b067801318e33c2e6eea4ab90b1fdd3',
         }
     }
+    _MONTH_NAMES_KA = ['იანვარი', 'თებერვალი', 'მარტი', 'აპრილი', 'მაისი', 'ივნისი','ივლისი', 'აგვისტო', 'სექტემბერი', 'ოქტომბერი', 'ნოემბერი', 'დეკემბერი']
+
+    _quality = staticmethod(qualities(('SD', 'HD')))
 
     def _real_extract(self, url):
         video_id = self._match_id(url)
         webpage = self._download_webpage(url, video_id)
 
-        title = self._html_search_regex(r'<h1[^>]*>([^<]+)</h1>', webpage, 'title')
+        title = (
-        description = self._og_search_description(webpage)
+            self._og_search_title(webpage, default=None)
-        thumbnail = self._html_search_meta(['og:image'], webpage)
+            or clean_html(get_element_by_class('my_video_title', webpage))
-        uploader = self._search_regex(r'<a[^>]+class="mv_user_name"[^>]*>([^<]+)<', webpage, 'uploader', fatal=False)
+            or self._html_search_regex(r'<title\b[^>]*>([^<]+)</title\b', webpage, 'title'))
 
         jwplayer_sources = self._parse_json(
             self._search_regex(
-                r"(?s)jwplayer\(\"mvplayer\"\).setup\(.*?sources: (.*?])", webpage, 'jwplayer sources'),
+                r'''(?s)jwplayer\s*\(\s*['"]mvplayer['"]\s*\)\s*\.\s*setup\s*\(.*?\bsources\s*:\s*(\[.*?])\s*[,});]''', webpage, 'jwplayer sources', fatal=False)
-            video_id, transform_source=js_to_json)
+            or '',
+            video_id, transform_source=js_to_json, fatal=False)
 
-        def _formats_key(f):
+        formats = self._parse_jwplayer_formats(jwplayer_sources or [], video_id)
-            if f['label'] == 'SD':
+        for f in formats or []:
-                return -1
+            f['preference'] = self._quality(f['format_id'])
-            elif f['label'] == 'HD':
+        self._sort_formats(formats)
-                return 1
-            else:
-                return 0
 
-        jwplayer_sources = sorted(jwplayer_sources, key=_formats_key)
+        description = (
+            self._og_search_description(webpage)
+            or get_element_by_id('long_desc_holder', webpage)
+            or self._html_search_meta('description', webpage))
 
-        formats = self._parse_jwplayer_formats(jwplayer_sources, video_id)
+        uploader = self._search_regex(r'<a[^>]+class="mv_user_name"[^>]*>([^<]+)<', webpage, 'uploader', fatal=False)
+
+        upload_date = get_element_by_class('mv_vid_upl_date', webpage)
+        # as ka locale may not be present roll a local date conversion
+        upload_date = (unified_strdate(
+            # translate any ka month to an en one
+            re.sub('|'.join(self._MONTH_NAMES_KA),
+                   lambda m: MONTH_NAMES['en'][self._MONTH_NAMES_KA.index(m.group(0))],
+                   upload_date, re.I))
+            if upload_date else None)
 
         return {
             'id': video_id,
@@ -52,5 +76,9 @@ class MyVideoGeIE(InfoExtractor):
             'description': description,
             'uploader': uploader,
             'formats': formats,
-            'thumbnail': thumbnail
+            'thumbnail': self._og_search_thumbnail(webpage),
+            'upload_date': upload_date,
+            'view_count': int_or_none(get_element_by_class('mv_vid_views', webpage)),
+            'like_count': int_or_none(get_element_by_id('likes_count', webpage)),
+            'dislike_count': int_or_none(get_element_by_id('dislikes_count', webpage)),
         }