[youtube] Fix extraction.

[youtube-dl] / youtube_dl / extractor / aparat.py

diff --git a/youtube_dl/extractor/aparat.py b/youtube_dl/extractor/aparat.py
index 63429780e8abf528165daf7e50a6317bce9a6c7d..883dcee7aa4cae953fff16dbca8cbc5fbf07e64e 100644 (file)
--- a/youtube_dl/extractor/aparat.py
+++ b/youtube_dl/extractor/aparat.py
@@ -1,60 +1,95 @@
 # coding: utf-8
 from __future__ import unicode_literals
 
-import re
-
 from .common import InfoExtractor
 from ..utils import (
-    ExtractorError,
-    HEADRequest,
+    int_or_none,
+    merge_dicts,
+    mimetype2ext,
+    url_or_none,
 )
 
 
 class AparatIE(InfoExtractor):
-    _VALID_URL = r'^https?://(?:www\.)?aparat\.com/(?:v/|video/video/embed/videohash/)(?P<id>[a-zA-Z0-9]+)'
+    _VALID_URL = r'https?://(?:www\.)?aparat\.com/(?:v/|video/video/embed/videohash/)(?P<id>[a-zA-Z0-9]+)'
 
-    _TEST = {
+    _TESTS = [{
         'url': 'http://www.aparat.com/v/wP8On',
-        'md5': '6714e0af7e0d875c5a39c4dc4ab46ad1',
+        'md5': '131aca2e14fe7c4dcb3c4877ba300c89',
         'info_dict': {
             'id': 'wP8On',
             'ext': 'mp4',
             'title': 'تیم گلکسی 11 - زومیت',
-            'age_limit': 0,
+            'description': 'md5:096bdabcdcc4569f2b8a5e903a3b3028',
+            'duration': 231,
+            'timestamp': 1387394859,
+            'upload_date': '20131218',
+            'view_count': int,
         },
-        # 'skip': 'Extremely unreliable',
-    }
+    }, {
+        # multiple formats
+        'url': 'https://www.aparat.com/v/8dflw/',
+        'only_matching': True,
+    }]
 
     def _real_extract(self, url):
         video_id = self._match_id(url)
 
-        # Note: There is an easier-to-parse configuration at
-        # http://www.aparat.com/video/video/config/videohash/%video_id
-        # but the URL in there does not work
-        embed_url = ('http://www.aparat.com/video/video/embed/videohash/' +
-                     video_id + '/vt/frame')
-        webpage = self._download_webpage(embed_url, video_id)
-
-        video_urls = [video_url.replace('\\/', '/') for video_url in re.findall(
-            r'(?:fileList\[[0-9]+\]\s*=|"file"\s*:)\s*"([^"]+)"', webpage)]
-        for i, video_url in enumerate(video_urls):
-            req = HEADRequest(video_url)
-            res = self._request_webpage(
-                req, video_id, note='Testing video URL %d' % i, errnote=False)
-            if res:
-                break
-        else:
-            raise ExtractorError('No working video URLs found')
-
-        title = self._search_regex(r'\s+title:\s*"([^"]+)"', webpage, 'title')
-        thumbnail = self._search_regex(
-            r'image:\s*"([^"]+)"', webpage, 'thumbnail', fatal=False)
-
-        return {
+        # Provides more metadata
+        webpage = self._download_webpage(url, video_id, fatal=False)
+
+        if not webpage:
+            # Note: There is an easier-to-parse configuration at
+            # http://www.aparat.com/video/video/config/videohash/%video_id
+            # but the URL in there does not work
+            webpage = self._download_webpage(
+                'http://www.aparat.com/video/video/embed/vt/frame/showvideo/yes/videohash/' + video_id,
+                video_id)
+
+        options = self._parse_json(
+            self._search_regex(
+                r'options\s*=\s*JSON\.parse\(\s*(["\'])(?P<value>(?:(?!\1).)+)\1\s*\)',
+                webpage, 'options', group='value'),
+            video_id)
+
+        player = options['plugins']['sabaPlayerPlugin']
+
+        formats = []
+        for sources in player['multiSRC']:
+            for item in sources:
+                if not isinstance(item, dict):
+                    continue
+                file_url = url_or_none(item.get('src'))
+                if not file_url:
+                    continue
+                item_type = item.get('type')
+                if item_type == 'application/vnd.apple.mpegurl':
+                    formats.extend(self._extract_m3u8_formats(
+                        file_url, video_id, 'mp4',
+                        entry_protocol='m3u8_native', m3u8_id='hls',
+                        fatal=False))
+                else:
+                    ext = mimetype2ext(item.get('type'))
+                    label = item.get('label')
+                    formats.append({
+                        'url': file_url,
+                        'ext': ext,
+                        'format_id': 'http-%s' % (label or ext),
+                        'height': int_or_none(self._search_regex(
+                            r'(\d+)[pP]', label or '', 'height',
+                            default=None)),
+                    })
+        self._sort_formats(
+            formats, field_preference=('height', 'width', 'tbr', 'format_id'))
+
+        info = self._search_json_ld(webpage, video_id, default={})
+
+        if not info.get('title'):
+            info['title'] = player['title']
+
+        return merge_dicts(info, {
             'id': video_id,
-            'title': title,
-            'url': video_url,
-            'ext': 'mp4',
-            'thumbnail': thumbnail,
-            'age_limit': self._family_friendly_search(webpage),
-        }
+            'thumbnail': url_or_none(options.get('poster')),
+            'duration': int_or_none(player.get('duration')),
+            'formats': formats,
+        })