API for the Aggregator Endpoint

[faktas2]

Hi,

This package currently doesn't support the aggregator endpoint, which is super useful at decreasing the amount of requests that could be made to PivotalTracker.

https://www.pivotaltracker.com/help/api/#Using_the_GET_Request_Aggregator

This PR adds a new service to the package for using the aggregator. It currently only supports getting the stories, comments of stories and reviews of stories. However, if the pattern is approved, I believe the API could be extended to support other parts of PivotalTracker.

Note: It uses some parts from the fork https://github.com/oschwald/go-pivotaltracker.

A sample use would be similar to below,

ptClient := pivotal.NewClient("TOKEN")
storiesToGet := []int{
	1,
	2,
	3,
	4,
}
projectID := 123
aggregation := ptClient.Aggregator.GetBuilder().Stories(projectID, storiesToGet).Story(projectID, 5)
aggregation.CommentsOfStories(projectID, storiesToGet).ReviewsOfStories(projectID, storiesToGet)
aggregation, err := aggregation.Send()
_ = err

story, _ := aggregation.GetStory(projectID, 1)
comments, _ := aggregation.GetComments(projectID, 2)
reviews, _ := aggregation.GetReviews(projectID, 2)
_ = story

[oschwald]

I had a few small comments.

[oschwald]

It is idiomatic to start the documentation with the name of the thing you are documenting , e.g., this might be "Story adds the url for the story to the aggregation data."

Would you be willing to update this for all the public member in this PR?

[oschwald]

Do these URL helpers need to be public? I think they would only be used internally.

[oschwald]

It seems like all of the other equivalent methods in the module return the *http.Response as well, e.g., List(). It'd probably make sense to do that for consistency. I am not sure the exact use case.