Fix ESPN extractor

Fixes the ESPN extractor to extract clips from the site.

youtube_dl/extractor/espn.py

@@ -51,20 +51,6 @@ class ESPNIE(OnceIE):
         'params': {
             'skip_download': True,
         },
-    }, {
-        'url': 'https://broadband.espn.go.com/video/clip?id=18910086',
-        'info_dict': {
-            'id': '18910086',
-            'ext': 'mp4',
-            'title': 'Kyrie spins around defender for two',
-            'description': 'md5:2b0f5bae9616d26fba8808350f0d2b9b',
-            'timestamp': 1489539155,
-            'upload_date': '20170315',
-        },
-        'params': {
-            'skip_download': True,
-        },
-        'expected_warnings': ['Unable to download f4m manifest'],
     }, {
         'url': 'http://nonredline.sports.espn.go.com/video/clip?id=19744672',
         'only_matching': True,
@@ -103,27 +89,29 @@ class ESPNIE(OnceIE):
     def _real_extract(self, url):
         video_id = self._match_id(url)
 
-        clip = self._download_json(
+        webpage = self._download_webpage(url, video_id)
-            'http://api-app.espn.com/v1/video/clips/%s' % video_id,
-            video_id)['videos'][0]
 
-        title = clip['headline']
+        data_id=self._search_regex(r'data-cerebro-id="(.*?)"',webpage,'data_id',group=1)
+
+        request_url = 'https://watch.auth.api.espn.com/video/auth/getclip/%s?apikey=5p8m6dw513q716wt2os04mec3' % data_id
+
+        clip = self._download_xml(
+            request_url,            
+            video_id).findall('clip')[0]
+
+
+        title = clip.findall('headline')[0].text
 
         format_urls = set()
         formats = []
 
-        def traverse_source(source, base_source_id=None):
-            for source_id, source in source.items():
-                if source_id == 'alert':
-                    continue
-                elif isinstance(source, compat_str):
-                    extract_source(source, base_source_id)
-                elif isinstance(source, dict):
-                    traverse_source(
-                        source,
-                        '%s-%s' % (base_source_id, source_id)
-                        if base_source_id else source_id)
         
+        def traverse_source(source):
+            for element in source.iter():
+                if element.tag == 'transcodes':
+                    continue
+               
+                extract_source(element.text, element.tag)
         def extract_source(source_url, source_id=None):
             if source_url in format_urls:
                 return
@@ -134,9 +122,6 @@ class ESPNIE(OnceIE):
             elif ext == 'smil':
                 formats.extend(self._extract_smil_formats(
                     source_url, video_id, fatal=False))
-            elif ext == 'f4m':
-                formats.extend(self._extract_f4m_formats(
-                    source_url, video_id, f4m_id=source_id, fatal=False))
             elif ext == 'm3u8':
                 formats.extend(self._extract_m3u8_formats(
                     source_url, video_id, 'mp4', entry_protocol='m3u8_native',
@@ -157,21 +142,18 @@ class ESPNIE(OnceIE):
                     f['preference'] = 1
                 formats.append(f)
 
-        links = clip.get('links', {})
+        links = clip.findall('transcodes')[0]
-        traverse_source(links.get('source', {}))
+        traverse_source(links)
-        traverse_source(links.get('mobile', {}))
         self._sort_formats(formats)
 
-        description = clip.get('caption') or clip.get('description')
+        description = clip.findall('caption')[0].text
-        thumbnail = clip.get('thumbnail')
+        duration = int_or_none(clip.findall('durationSeconds')[0].text)
-        duration = int_or_none(clip.get('duration'))
+        timestamp = unified_timestamp(clip.findall('publishDate')[0].text)
-        timestamp = unified_timestamp(clip.get('originalPublishDate'))
 
         return {
             'id': video_id,
             'title': title,
             'description': description,
-            'thumbnail': thumbnail,
             'timestamp': timestamp,
             'duration': duration,
             'formats': formats,