From 8fbb1e6d97cda90d588d4263a18906a52d147fba Mon Sep 17 00:00:00 2001
From: Kenneth Reitz <me@kennethreitz.com>
Date: Sat, 20 Aug 2011 19:17:41 -0400
Subject: [PATCH] move encoding methods into utils for external consumption

---
 requests/utils.py | 61 +++++++++++++++++++++++++++++++++++++++++++++++
 1 file changed, 61 insertions(+)

diff --git a/requests/utils.py b/requests/utils.py
index 8ac78b4e..72705734 100644
--- a/requests/utils.py
+++ b/requests/utils.py
@@ -9,7 +9,9 @@ that are also useful for external consumption.
 
 """
 
+import cgi
 import cookielib
+import re
 
 
 def dict_from_cookiejar(cookiejar):
@@ -70,3 +72,62 @@ def add_dict_to_cookiejar(cj, cookie_dict):
         cj.set_cookie(cookie)
 
     return cj
+
+
+def get_encodings_from_content(content):
+    """Returns encodings from given content string."""
+
+    charset_re = re.compile(r'<meta.*?charset=["\']*(.+?)["\'>]', flags=re.I)
+
+    return charset_re.findall(content)
+
+
+
+def get_encoding_from_headers(headers):
+    """Returns encodings from given HTTP Header Dict."""
+
+    content_type = headers.get('content-type')
+    content_type, params = cgi.parse_header(content_type)
+
+    if 'charset' in params:
+        return params['charset'].strip("'\"")
+
+
+def get_unicode_from_response(r):
+    """Returns the requested content back in unicode.
+
+    Tried:
+    1. charset from content-type
+    2. every encodings from <meta ... charset=XXX>
+    3. fall back and replace all unicode characters
+    """
+
+    tried_encodings = []
+
+    # Try charset from content-type
+    encoding = get_encoding_from_headers(r.headers)
+
+    if encoding:
+        try:
+            print '!'
+            return unicode(r.content, encoding)
+        except UnicodeError:
+            tried_encodings.append(encoding)
+
+    # Try every encodings from <meta ... charset=XXX>
+    encodings = get_encodings_from_content(r.content)
+
+    for encoding in encodings:
+        if encoding in tried_encodings:
+            continue
+        try:
+
+            return unicode(r.content, encoding)
+        except (UnicodeError, TypeError):
+            tried_encodings.append(encoding)
+
+    # Fall back:
+    try:
+        return unicode(r.content, encoding, errors='replace')
+    except TypeError:
+        return r.content