From 636e61d68b9176c614f8243295c81552dd0ee199 Mon Sep 17 00:00:00 2001
From: Kenneth Reitz <me@kennethreitz.com>
Date: Tue, 23 Aug 2011 23:09:09 -0400
Subject: [PATCH] utils.unicode_from_html

---
 requests/utils.py | 19 +++++++++++++++++++
 1 file changed, 19 insertions(+)

diff --git a/requests/utils.py b/requests/utils.py
index 8a419acb..35eb005b 100644
--- a/requests/utils.py
+++ b/requests/utils.py
@@ -110,6 +110,25 @@ def get_encoding_from_headers(headers):
         return params['charset'].strip("'\"")
 
 
+def unicode_from_html(content):
+    """Attempts to decode an HTML string into unicode.
+    If unsuccessful, the original content is returned.
+    """
+
+    encodings = get_encodings_from_content(content)
+
+    for encoding in encodings:
+
+        try:
+            return unicode(content, encoding)
+        except (UnicodeError, TypeError):
+            pass
+
+    return content
+
+
+
+
 def get_unicode_from_response(r):
     """Returns the requested content back in unicode.