Add support for direct links to a video (#1973)

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

diff --git a/youtube_dl/extractor/generic.py b/youtube_dl/extractor/generic.py
index 216e032186297b7b91a488fb1edd1421e3270b39..209f68204ef7fd342edf8200e148f249a5d99266 100644 (file)
--- a/youtube_dl/extractor/generic.py
+++ b/youtube_dl/extractor/generic.py
@@ -13,6 +13,8 @@ from ..utils import (
     ExtractorError,
     smuggle_url,
     unescapeHTML,
+    unified_strdate,
+    url_basename,
 )
 from .brightcove import BrightcoveIE
 
@@ -71,6 +73,17 @@ class GenericIE(InfoExtractor):
                 u'skip_download': True,
             },
         },
+        # Direct link to a video
+        {
+            u'url': u'http://media.w3.org/2010/05/sintel/trailer.mp4',
+            u'file': u'trailer.mp4',
+            u'md5': u'67d406c2bcb6af27fa886f31aa934bbe',
+            u'info_dict': {
+                u'id': u'trailer',
+                u'title': u'trailer',
+                u'upload_date': u'20100513',
+            }
+        }
     ]
 
     def report_download_webpage(self, video_id):
@@ -83,7 +96,7 @@ class GenericIE(InfoExtractor):
         """Report information extraction."""
         self._downloader.to_screen(u'[redirect] Following redirect to %s' % new_url)
 
-    def _test_redirect(self, url):
+    def _send_head(self, url):
         """Check if it is a redirect, like url shorteners, in case return the new url."""
         class HeadRequest(compat_urllib_request.Request):
             def get_method(self):
@@ -131,29 +144,46 @@ class GenericIE(InfoExtractor):
         response = opener.open(HeadRequest(url))
         if response is None:
             raise ExtractorError(u'Invalid URL protocol')
-        new_url = response.geturl()
-
-        if url == new_url:
-            return False
-
-        self.report_following_redirect(new_url)
-        return new_url
+        return response
 
     def _real_extract(self, url):
         parsed_url = compat_urlparse.urlparse(url)
         if not parsed_url.scheme:
             self._downloader.report_warning('The url doesn\'t specify the protocol, trying with http')
             return self.url_result('http://' + url)
+        video_id = os.path.splitext(url.split('/')[-1])[0]
 
         try:
-            new_url = self._test_redirect(url)
-            if new_url:
-                return [self.url_result(new_url)]
+            response = self._send_head(url)
+
+            # Check for redirect
+            new_url = response.geturl()
+            if url != new_url:
+                self.report_following_redirect(new_url)
+                return self.url_result(new_url)
+
+            # Check for direct link to a video
+            content_type = response.headers.get('Content-Type', '')
+            m = re.match(r'^(?:audio|video)/(?P<format_id>.+)$', content_type)
+            if m:
+                upload_date = response.headers.get('Last-Modified')
+                if upload_date:
+                    upload_date = unified_strdate(upload_date)
+                assert (url_basename(url) == 'trailer.mp4')
+                return {
+                    'id': video_id,
+                    'title': os.path.splitext(url_basename(url))[0],
+                    'formats': [{
+                        'format_id': m.group('format_id'),
+                        'url': url,
+                    }],
+                    'upload_date': upload_date,
+                }
+
         except compat_urllib_error.HTTPError:
             # This may be a stupid server that doesn't like HEAD, our UA, or so
             pass
 
-        video_id = url.split('/')[-1]
         try:
             webpage = self._download_webpage(url, video_id)
         except ValueError:
@@ -222,6 +252,18 @@ class GenericIE(InfoExtractor):
                 'id': video_id,
             }
 
+        # Look for embedded blip.tv player
+        mobj = re.search(r'<meta\s[^>]*https?://api.blip.tv/\w+/redirect/\w+/(\d+)', webpage)
+        if mobj:
+            return self.url_result('http://blip.tv/seo/-'+mobj.group(1), 'BlipTV')
+        mobj = re.search(r'<(?:iframe|embed|object)\s[^>]*https?://(?:\w+\.)?blip.tv/(?:play/|api\.swf#)([a-zA-Z0-9]+)', webpage)
+        if mobj:
+            player_url = 'http://blip.tv/play/%s.x?p=1' % mobj.group(1)
+            player_page = self._download_webpage(player_url, mobj.group(1))
+            blip_video_id = self._search_regex(r'data-episode-id="(\d+)', player_page, u'blip_video_id', fatal=False)
+            if blip_video_id:
+                return self.url_result('http://blip.tv/seo/-'+blip_video_id, 'BlipTV')
+
         # Look for Bandcamp pages with custom domain
         mobj = re.search(r'<meta property="og:url"[^>]*?content="(.*?bandcamp\.com.*?)"', webpage)
         if mobj is not None:
@@ -229,6 +271,12 @@ class GenericIE(InfoExtractor):
             # Don't set the extractor because it can be a track url or an album
             return self.url_result(burl)
 
+        # Look for embedded Vevo player
+        mobj = re.search(
+            r'<iframe[^>]+?src=(["\'])(?P<url>(?:https?:)?//(?:cache\.)?vevo\.com/.+?)\1', webpage)
+        if mobj is not None:
+            return self.url_result(mobj.group('url'))
+
         # Start with something easy: JW Player in SWFObject
         mobj = re.search(r'flashvars: [\'"](?:.*&)?file=(http[^\'"&]*)', webpage)
         if mobj is None: