Extending a built-in collection – a list that does statistics

In the Designing classes with lots of processing recipe, we looked at a way to distinguish between a complex algorithm and a collection. We showed how to encapsulate the algorithm and the data into separate classes. The alternative design strategy is to extend the collection to incorporate a useful algorithm.

How can we extend Python's built-in collections? How can we add features to the built-in list?

Getting ready

We'll create a sophisticated list class where each instance can compute the sums and averages of the items in the list. This will require an application to only put numbers in the list; otherwise, there will be ValueError exceptions.

We're going to show methods that explicitly use generator expressions as places where additional processing can be included. Rather than use sum(self), we're going to emphasize sum(v for v in self) because there are two common future extensions...