Caching

I didn't want to create a data layer, hence I am wondering what you think about caching the GET requests? Otherwise, the rate limit happens so quick. ;D I think all that's potentially needed is exposing the http client to register a transport.

Start by locating the HTTP client and the GET request entry points in the Go Fitbit API client, then assess how exposing the client for transport registration would support caching. The work is complete when GET requests can be cached to reduce Fitbit rate-limit hits, with the intended behavior and scope established.

Written by the indexing model from the issue text.