Add support for embed.ly

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

diff --git a/youtube_dl/extractor/generic.py b/youtube_dl/extractor/generic.py
index 082da9c77b138f35651e2de8a2d035f8fee1f517..e84c022a587cf7cae22f3d6ec33fdbf15f07098b 100644 (file)
--- a/youtube_dl/extractor/generic.py
+++ b/youtube_dl/extractor/generic.py
@@ -4,6 +4,7 @@ from __future__ import unicode_literals
 
 import os
 import re
+import xml.etree.ElementTree
 
 from .common import InfoExtractor
 from .youtube import YoutubeIE
@@ -12,6 +13,7 @@ from ..utils import (
     compat_urllib_parse,
     compat_urllib_request,
     compat_urlparse,
+    compat_xml_parse_error,
 
     ExtractorError,
     HEADRequest,
@@ -100,6 +102,18 @@ class GenericIE(InfoExtractor):
                 'title': '2cc213299525360.mov',  # that's what we get
             },
         },
+        # embed.ly video
+        {
+            'url': 'http://www.tested.com/science/weird/460206-tested-grinding-coffee-2000-frames-second/',
+            'info_dict': {
+                'id': '9ODmcdjQcHQ',
+                'ext': 'mp4',
+            },
+            # No need to test YoutubeIE here
+            'params': {
+                'skip_download': True,
+            },
+        },
     ]
 
     def report_download_webpage(self, video_id):
@@ -159,6 +173,25 @@ class GenericIE(InfoExtractor):
             raise ExtractorError('Invalid URL protocol')
         return response
 
+    def _extract_rss(self, url, video_id, doc):
+        playlist_title = doc.find('./channel/title').text
+        playlist_desc_el = doc.find('./channel/description')
+        playlist_desc = None if playlist_desc_el is None else playlist_desc_el.text
+
+        entries = [{
+            '_type': 'url',
+            'url': e.find('link').text,
+            'title': e.find('title').text,
+        } for e in doc.findall('./channel/item')]
+
+        return {
+            '_type': 'playlist',
+            'id': url,
+            'title': playlist_title,
+            'description': playlist_desc,
+            'entries': entries,
+        }
+
     def _real_extract(self, url):
         parsed_url = compat_urlparse.urlparse(url)
         if not parsed_url.scheme:
@@ -219,6 +252,14 @@ class GenericIE(InfoExtractor):
 
         self.report_extraction(video_id)
 
+        # Is it an RSS feed?
+        try:
+            doc = xml.etree.ElementTree.fromstring(webpage.encode('utf-8'))
+            if doc.tag == 'rss':
+                return self._extract_rss(url, video_id, doc)
+        except compat_xml_parse_error:
+            pass
+
         # it's tempting to parse this further, but you would
         # have to take into account all the variations like
         #   Video Title - Site Name
@@ -234,11 +275,21 @@ class GenericIE(InfoExtractor):
             r'^(?:https?://)?([^/]*)/.*', url, 'video uploader')
 
         # Look for BrightCove:
-        bc_url = BrightcoveIE._extract_brightcove_url(webpage)
-        if bc_url is not None:
+        bc_urls = BrightcoveIE._extract_brightcove_urls(webpage)
+        if bc_urls:
             self.to_screen('Brightcove video detected.')
-            surl = smuggle_url(bc_url, {'Referer': url})
-            return self.url_result(surl, 'Brightcove')
+            entries = [{
+                '_type': 'url',
+                'url': smuggle_url(bc_url, {'Referer': url}),
+                'ie_key': 'Brightcove'
+            } for bc_url in bc_urls]
+
+            return {
+                '_type': 'playlist',
+                'title': video_title,
+                'id': video_id,
+                'entries': entries,
+            }
 
         # Look for embedded (iframe) Vimeo player
         mobj = re.search(
@@ -342,6 +393,14 @@ class GenericIE(InfoExtractor):
         if mobj is not None:
             return self.url_result(mobj.group('url'), 'HuffPost')
 
+        # Look for embed.ly
+        mobj = re.search(r'class=["\']embedly-card["\'][^>]href=["\'](?P<url>[^"\']+)', webpage)
+        if mobj is not None:
+            return self.url_result(mobj.group('url'))
+        mobj = re.search(r'class=["\']embedly-embed["\'][^>]src=["\'][^"\']*url=(?P<url>[^&]+)', webpage)
+        if mobj is not None:
+            return self.url_result(compat_urllib_parse.unquote(mobj.group('url')))
+
         # Start with something easy: JW Player in SWFObject
         mobj = re.search(r'flashvars: [\'"](?:.*&)?file=(http[^\'"&]*)', webpage)
         if mobj is None: