[ina] improve extraction

youtube_dl/extractor/ina.py

@@ -1,36 +1,72 @@
 # coding: utf-8
 from __future__ import unicode_literals
 
-import re
-
 from .common import InfoExtractor
+from ..utils import (
+    int_or_none,
+    strip_or_none,
+    xpath_attr,
+    xpath_text,
+)
 
 
 class InaIE(InfoExtractor):
-    _VALID_URL = r'https?://(?:www\.)?ina\.fr/video/(?P<id>I?[A-Z0-9]+)'
+    _VALID_URL = r'https?://(?:www\.)?ina\.fr/video/(?P<id>[A-Z0-9_]+)'
-    _TEST = {
+    _TESTS = [{
         'url': 'http://www.ina.fr/video/I12055569/francois-hollande-je-crois-que-c-est-clair-video.html',
         'md5': 'a667021bf2b41f8dc6049479d9bb38a3',
         'info_dict': {
             'id': 'I12055569',
             'ext': 'mp4',
             'title': 'François Hollande "Je crois que c\'est clair"',
+            'description': 'md5:3f09eb072a06cb286b8f7e4f77109663',
         }
-    }
+    }, {
+        'url': 'https://www.ina.fr/video/S806544_001/don-d-organes-des-avancees-mais-d-importants-besoins-video.html',
+        'only_matching': True,
+    }]
 
     def _real_extract(self, url):
-        mobj = re.match(self._VALID_URL, url)
+        video_id = self._match_id(url)
+        info_doc = self._download_xml(
+            'http://player.ina.fr/notices/%s.mrss' % video_id, video_id)
+        item = info_doc.find('channel/item')
+        title = xpath_text(item, 'title', fatal=True)
+        media_ns_xpath = lambda x: self._xpath_ns(x, 'http://search.yahoo.com/mrss/')
+        content = item.find(media_ns_xpath('content'))
 
-        video_id = mobj.group('id')
+        get_furl = lambda x: xpath_attr(content, media_ns_xpath(x), 'url')
-        mrss_url = 'http://player.ina.fr/notices/%s.mrss' % video_id
+        formats = []
-        info_doc = self._download_xml(mrss_url, video_id)
+        for q, w, h in (('bq', 400, 300), ('mq', 512, 384), ('hq', 768, 576)):
+            q_url = get_furl(q)
+            if not q_url:
+                continue
+            formats.append({
+                'format_id': q,
+                'url': q_url,
+                'width': w,
+                'height': h,
+            })
+        if not formats:
+            formats = [{
+                'url': get_furl('player') or content.attrib['url'],
+            }]
 
-        self.report_extraction(video_id)
+        thumbnails = []
-
+        for thumbnail in content.findall(media_ns_xpath('thumbnail')):
-        video_url = info_doc.find('.//{http://search.yahoo.com/mrss/}player').attrib['url']
+            thumbnail_url = thumbnail.get('url')
+            if not thumbnail_url:
+                continue
+            thumbnails.append({
+                'url': thumbnail_url,
+                'height': int_or_none(thumbnail.get('height')),
+                'width': int_or_none(thumbnail.get('width')),
+            })
 
         return {
             'id': video_id,
-            'url': video_url,
+            'formats': formats,
-            'title': info_doc.find('.//title').text,
+            'title': title,
+            'description': strip_or_none(xpath_text(item, 'description')),
+            'thumbnails': thumbnails,
         }