For this recipe, we will be using the results of the previous two recipes and the twitter_oauth_login() function. Also, we will be working in IPython or the default Python REPL, if you prefer that instead. Feel free to use an editor in order to start capturing and modifying the code as it grows in complexity.

The following steps will allow you to determine all of your Twitter friends and followers:

This recipe demonstrates just how useful the Twython package is and how easy it makes certain tasks. After we logged in using the twitter_oauth_login function, we make two basic calls using the twitter object, one to get friends' IDs and one to get followers' IDs. Note that we set the count parameter to 5000, which is the maximum value allowed by the Twitter API. The twitter object returns a dictionary from which we extract the actual IDs.

One of the nice things about the Twython interface is how closely it mirrors the Twitter API. If you ever have a question about a particular function, just check the Twitter documents.

Once we have collected our list of friends' and followers' IDs, we turn to the Python set type for some quick navel gazing. The Python set type, which has been built into Python since Version 2.4, is an unordered collection of unique objects. The key word for us is the word unique. If we create a set from a list with duplicates, we will get a set with only unique elements; set([1, 2, 2, 3, 3, 3]) will return {1, 2, 3}.

We unite the set of friends' IDs with the followers' IDs to determine the total set of unique IDs of Twitter users that either follow or are followed by us. In the preceding code, we use the union method of the set type, but there are several other ways in which we could have done this:

(friends_set | followers_set)
(set(friends_ids + followers_ids))

While Twython's beautiful abstraction hides some of the complexity of using the API, this simplicity or magic can be problematic if we don't have an understanding of what is actually happening behind the scenes. When we call the twitter.get_friends_ids(count=5000) method, we are sending an HTTP GET request to a particular URL. In the case of twitter.get_friends_ids(), the URL is https://api.twitter.com/1.1/friends/ids.json.

The count=5000 input parameter to the function call shows up as field-value pairs in the URL and as such, the URL becomes https://api.twitter.com/1.1/friends/ids.json?count=5000.

Now, the actual API endpoint requires some default parameter values that Twython fills in for us, as shown in the following URL for clarity:

https://api.twitter.com/1.1/friends/ids.json?cursor=-1&screen_name=sayhitosean&count=5000

The Twitter v1.1 API requires all requests to be authenticated using OAuth. The required information is actually embedded in the header of the GET request, and the process of constructing the appropriate header is extensive (for more information, go to https://dev.twitter.com/docs/auth/authorizing-request). Thus, Twython not only forms the proper URL for making the request, but also handles the relatively painful Open Authorization so that we don't have to. If you are interested, you can go down a level lower and construct your own GET requests using the excellent request library or an alternative of your choosing. We leave this for the reader to explore.