Use dict literal instead of constructor where possible and add some p…

…erf related docs

README.md

@@ -447,6 +447,13 @@ wryter.unbind('cid')
 
 The idea behind this is that a cid can be passed into any log message within the same context. "within the same context" is a very abstract defintion, and is up to the developer to implement as it might be thread-related, framework-related, or else. I intend to expand the framework, but for now, that's what it is.
 
+
+## Performance optimizations
+
+* Using dict literal instead of constructor (https://stackoverflow.com/a/6612024/3332312)
+* Using tuples instead of lists (https://stackoverflow.com/a/68817/3332312)
+* Using dict subscription instead of get (https://stackoverflow.com/questions/7631929/python-dictionary-datastructure-which-method-d-or-d-get#comment58682247_7631951)
+
 ## Testing
 
 ```shell

wryte.py

@@ -260,11 +260,12 @@ def _get_base(name, hostname):
         """Generate base fields for each log message.
         """
         # TODO: Document that these are only generated once.
-        return dict(
-            name=name,
-            hostname=hostname or socket.gethostname(),
-            pid=os.getpid(),
-            type='log')
+        return {
+            'name': name,
+            'hostname': hostname or socket.gethostname(),
+            'pid': os.getpid(),
+            'type': 'log'
+        }
 
     @staticmethod
     def _logger(name):
@@ -340,10 +341,11 @@ def _enrich(self, message, level, objects, kwargs=None):
         if kwargs:
             log.update(kwargs)
 
-        log.update(dict(
-            message=message,
-            level=level.upper(),
-            timestamp=self._get_timestamp()))
+        log.update({
+            'message': message,
+            'level': level.upper(),
+            'timestamp': self._get_timestamp()
+        })
         return log
 
     def bind(self, *objects):