Learn how to aggregate data from the series.

Aggregators

Aggregators allows you to perform calculations on top of the whole series. They plug seamelessly into Series and can be used through dot notation.

For example you want to know the number of events that happened during the specific timeframe. Of course you can calculate that manually with some form of reduce, but you can also use build in Trifle::Stats methods to calculate this.

Aggregator allows you to pass also slices: Integer that will split the series into equal chunks and calculate specific action on top of each slice. It defaults to 1 (obviously).

series = Trifle::Stats.series(...)
=> #<Trifle::Stats::Series:0x0000ffffa14256e8 @series={:at=>[2024-03-22 19:38:00 +0000, 2024-03-22 19:39:00 +0000], :values=>[{events: {count: 42, sum: 2184}}, {events: {count: 33, sum: 1553}}]}>
series.aggregate.sum(path: 'events.count')
=> [75]
series.aggregate.sum(path: 'events.count', slices: 2)
=> [42, 33]

Note: path is a list of keys joined by dot. Ie orders.shipped.count would represent value at {orders: { shipped: { count: ... } } }.

Custom Aggregators

You can use one of predefined aggregators or write your own. Aggregator needs to implement at least one method that accepts key parameters series: and path: and returns something. The series:is passed in automatically.

class MyAggregator
  def aggregate(series:, path:)
    # perform some calculation on top of series
    rand(1000)
  end
end

If you write your own aggregator and would like to access it through dot notation, you need to register it through Trifle::Stats::Series.register_aggregator(NAME, self) inside of your aggregator class.

class MyAggregator
  Trifle::Stats::Series.register_aggregator(:rand, self)

  def aggregate(series:, path:)
    # perform some calculation on top of series
    rand(1000)
  end
end

And from there you can access it through series.aggregate.rand(path: 'a.count')