Fix merge problem in pull request #43. Previously, there was just one for loop that iterated thru all of the sites. But, with the addition of the parallel functionality, there are now two for loops. The dictionary changes were not done in the second loop, which caused bogus results.

sherlock.py

@@ -143,7 +143,7 @@ def sherlock(username, verbose=False, tor=False, unique_tor=False):
         results_total[social_network] = results_site
 
     # Core logic: If tor requests, make them here. If multi-threaded requests, wait for responses
-    for social_network in data:
+    for social_network, net_info in data.items():
 
         # Retrieve results again
         results_site = results_total.get(social_network)