Use a persistent requests session instead of recreating it (#39)

As per the documentation for requests, this allows us to have a
persistent TCP connection and uses connection pooling, causing some
speed ups instead of constantly recreating the session object via
`requests.request`.

disco/api/http.py

@@ -202,6 +202,8 @@ def __init__(self, token):
         if token:
             self.headers['Authorization'] = 'Bot ' + token
 
+        self.session = requests.Session()
+
     def __call__(self, route, args=None, **kwargs):
         return self.call(route, args, **kwargs)
 
@@ -257,7 +259,7 @@ def call(self, route, args=None, **kwargs):
         # Make the actual request
         url = self.BASE_URL + route[1].format(**args)
         self.log.info('%s %s (%s)', route[0].value, url, kwargs.get('params'))
-        r = requests.request(route[0].value, url, **kwargs)
+        r = self.session.request(route[0].value, url, **kwargs)
 
         # Update rate limiter
         self.limiter.update(bucket, r)