Case insensitive dictionary for request headers

brubeck/request.py

@@ -5,12 +5,19 @@
 import urlparse
 import re
 
+try:
+    from collections import OrderedDict
+except:
+    from ordereddict import OrderedDict
+
+
 def parse_netstring(ns):
     length, rest = ns.split(':', 1)
     length = int(length)
     assert rest[length] == ',', "Netstring did not end in ','"
     return rest[:length], rest[length + 1:]
 
+
 def to_bytes(data, enc='utf8'):
     """Convert anything to bytes
     """
@@ -23,6 +30,18 @@ def to_unicode(s, enc='utf8'):
     return s if isinstance(s, unicode) else unicode(str(s), encoding=enc)
 
 
+def uncgi(headers):
+    """Cleaned up WSGI headers
+    """
+    new_headers = Headers()
+    for key, value in headers.iteritems():
+        if key.upper().startswith('HTTP_'):
+            new_headers[key[5:].replace('_', '-')] = value
+        elif '_' in key:
+            new_headers[key.replace('_', '-')] = value
+    return new_headers
+
+
 class Request(object):
     """Word.
     """
@@ -90,7 +109,7 @@ def _parse_mime_body(self, boundary, data, arguments, files):
                 continue
             #headers = HTTPHeaders.parse(part[:eoh].decode("utf-8"))
             header_string = part[:eoh].decode("utf-8")
-            headers = dict()
+            headers = Headers()
             last_key = ''
             for line in header_string.splitlines():
                 if line[0].isspace():
@@ -193,7 +212,7 @@ def parse_msg(msg):
         sender, conn_id, path, rest = msg.split(' ', 3)
         headers, rest = parse_netstring(rest)
         body, _ = parse_netstring(rest)
-        headers = json.loads(headers)
+        headers = Headers(json.loads(headers))
         # construct url from request
         scheme = headers.get('URL_SCHEME', 'http')
         netloc = headers.get('host')
@@ -216,20 +235,14 @@ def parse_wsgi_request(environ):
             body = environ["wsgi.input"].read(int(environ['CONTENT_LENGTH']))
             del environ["CONTENT_LENGTH"]
             del environ["wsgi.input"]
-        #setting headers to environ dict with no manipulation
-        headers = environ
-        # normalize request dict
+        # normalize environ dict
+        headers = Headers(environ)
+        headers.update(uncgi(environ))
         if 'REQUEST_METHOD' in headers:
             headers['METHOD'] = headers['REQUEST_METHOD']
         if 'QUERY_STRING' in headers:
             headers['QUERY'] = headers['QUERY_STRING']
-        if 'CONTENT_TYPE' in headers:
-            headers['content-type'] = headers['CONTENT_TYPE']
-        headers['version'] = 1.1  #TODO: hardcoded!
-        if 'HTTP_COOKIE' in headers:
-            headers['cookie'] = headers['HTTP_COOKIE']
-        if 'HTTP_CONNECTION' in headers:
-            headers['connection'] = headers['HTTP_CONNECTION']
+        headers['VERSION'] = headers.get('SERVER_PROTOCOL', 'HTTP/1.1')
         # construct url from request
         scheme = headers['wsgi.url_scheme']
         netloc = headers.get('HTTP_HOST')
@@ -294,3 +307,76 @@ def get_argument(self, name, default=None, strip=True):
         if not args:
             return default
         return args[-1]
+
+
+class Headers(OrderedDict):
+    """Ordered dictionary with case insensitive lookup that preserves original
+    case when listing.
+    """
+
+    class CaseInsensitiveString(str):
+        def __hash__(self):
+            """
+            >>> hash(Headers.CaseInsensitiveString('Doc')) == hash('doc')
+            True
+            """
+            return hash(self.lower())
+
+        def __eq__(self, other):
+            """
+            >>> Headers.CaseInsensitiveString('Doc') == 'doc'
+            True
+            """
+            return self.lower() == other.lower()
+
+    def __keytransform__(self, key):
+        return self.CaseInsensitiveString(key)
+
+    def __contains__(self, key):
+        """
+        >>> 'USER-AGENT' in Headers({'User-Agent': 'doctest'})
+        True
+        """
+        return super(Headers, self).__contains__(self.__keytransform__(key))
+
+    def __getitem__(self, key):
+        """
+        >>> h = Headers({'User-Agent': 'doctest'})
+        >>> h['USER-AGENT']
+        'doctest'
+        """
+        return super(Headers, self).__getitem__(self.__keytransform__(key))
+
+    def __setitem__(self, key, value):
+        """
+        >>> h = Headers()
+        >>> h['User-Agent'] = 'doctest'
+        >>> h['USER-AGENT']
+        'doctest'
+        """
+        return super(Headers, self).__setitem__(self.__keytransform__(key), value)
+
+    def __delitem__(self, key):
+        """
+        >>> h = Headers({'User-Agent': 'doctest'})
+        >>> del h['USER-AGENT']
+        >>> 'User-Agent' not in h
+        True
+        """
+        return super(Headers, self).__delitem__(self.__keytransform__(key))
+
+    def has_key(self, key):
+        """
+        >>> Headers({'User-Agent': 'doctest'}).has_key('USER-AGENT')
+        True
+        """
+        return super(Headers, self).has_key(self.__keytransform__(key))
+
+    def get(self, key, *args):
+        """
+        >>> Headers({'User-Agent': 'doctest'}).get('USER-AGENT')
+        'doctest'
+        >>> Headers().get('User-Agent', 'doctest')
+        'doctest'
+        """
+        return super(Headers, self).get(self.__keytransform__(key), *args)