Beacon v2 has 3 main ideas at its core
In Beacon v1 any data analysis involving anything other than `get an event collection, filtered and grouped' involved a lot of complex code. keen-query aims to abstract away the common patterns into a relatively simple syntax. It's worth understanding how it works (see keen-query.md in his gist), but for the purposes of this document it's sufficient to know that
- Given a string which passes a set of parsing rules (see https://github.com/Financial-Times/keen-query/blob/master/lib/parser.js) an object (
kq) is generated which encapsulates either a single call to the keen API, or multiple calls with one or more rules for aggregating them. It may also define post-processing actions to perform on the data once it's fetched from keen. - There is an equivalent JavaScript API to the above which can either be used to construct a
kqinstance from scratch or to create a modified copy of a