From 46720279c28afb646b6ac19bcb11e85bb4bea726 Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?Jaime=20Marqui=CC=81nez=20Ferra=CC=81ndiz?=
 <jaime.marquinez.ferrandiz@gmail.com>
Date: Fri, 12 Jul 2013 19:00:19 +0200
Subject: InfoExtractor: add some helper methods to extract OpenGraph info

---
 youtube_dl/extractor/common.py | 24 ++++++++++++++++++++++++
 1 file changed, 24 insertions(+)

(limited to 'youtube_dl/extractor/common.py')

diff --git a/youtube_dl/extractor/common.py b/youtube_dl/extractor/common.py
index 1bd5538ca..0a0c4047d 100644
--- a/youtube_dl/extractor/common.py
+++ b/youtube_dl/extractor/common.py
@@ -257,6 +257,30 @@ class InfoExtractor(object):
         
         return (username, password)
 
+    # Helper functions for extracting OpenGraph info
+    @staticmethod
+    def _og_regex(property):
+        return r'<meta.+?property=[\'"]og:%s[\'"].+?content=(?:"(.+?)"|\'(.+?)\')' % property
+
+    def _og_search_property(self, property, html, name=None, **kargs):
+        if name is None:
+            name = 'OpenGraph %s' % property
+        return self._html_search_regex(self._og_regex(property), html, name, **kargs)
+
+    def _og_search_thumbnail(self, html, **kargs):
+        return self._og_search_property('image', html, 'thumbnail url', fatal=False, **kargs)
+
+    def _og_search_description(self, html, **kargs):
+        return self._og_search_property('description', html, fatal=False, **kargs)
+
+    def _og_search_title(self, html, **kargs):
+        return self._og_search_property('title', html, **kargs)
+
+    def _og_search_video_url(self, html, name='video url', **kargs):
+        return self._html_search_regex([self._og_regex('video:secure_url'),
+                                        self._og_regex('video')],
+                                       html, name, **kargs)
+
 class SearchInfoExtractor(InfoExtractor):
     """
     Base class for paged search queries extractors.
-- 
cgit v1.2.3


From 44dbe8903580e1e62ea6f3881b2fa469853c4a83 Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?Jaime=20Marqui=CC=81nez=20Ferra=CC=81ndiz?=
 <jaime.marquinez.ferrandiz@gmail.com>
Date: Sat, 13 Jul 2013 11:29:08 +0200
Subject: Use re.DOTALL by default when searching OpenGraph properties

---
 youtube_dl/extractor/common.py | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

(limited to 'youtube_dl/extractor/common.py')

diff --git a/youtube_dl/extractor/common.py b/youtube_dl/extractor/common.py
index 0a0c4047d..05b243871 100644
--- a/youtube_dl/extractor/common.py
+++ b/youtube_dl/extractor/common.py
@@ -265,7 +265,7 @@ class InfoExtractor(object):
     def _og_search_property(self, property, html, name=None, **kargs):
         if name is None:
             name = 'OpenGraph %s' % property
-        return self._html_search_regex(self._og_regex(property), html, name, **kargs)
+        return self._html_search_regex(self._og_regex(property), html, name, flags=re.DOTALL, **kargs)
 
     def _og_search_thumbnail(self, html, **kargs):
         return self._og_search_property('image', html, 'thumbnail url', fatal=False, **kargs)
-- 
cgit v1.2.3