Add support for aggregating transforms

[thbar]

See #53.

As seen in this StackOverflow question & other similar situations, it can be helpful to ensure one can aggregate rows together.

This PR introduces an optional #close method on transforms, which can either:

• close resources (e.g. a tracing transform could this way close its output file)
• yield one or more rows (e.g. an aggregating transform can provide the last aggregate row to the rest of the pipeline)

Because of the yielding support, you must use the new StreamingRunner for this to work:

extend Kiba::DSLExtensions::Config
config :kiba, runner: Kiba::StreamingRunner

The bundled AggregateTransform provides an example of use:

class AggregateTransform
  def initialize(aggregate_size:)
    @aggregate_size = aggregate_size
  end

  def process(row)
    @buffer ||= []
    @buffer << row
    if @buffer.size == @aggregate_size
      yield @buffer
      @buffer = []
    end
    nil
  end

  def close
    yield @buffer unless @buffer.empty?
  end
end

[ttilberg]

Hey @thbar, I've been seeing some recent activity with you mentioning Kiba 3 plans. Do you still have intention to officially release Kiba 2.5 with this feature? Or will it be waiting for the full 3.0 release? I've been looking forward to using this, but don't want to target unofficial releases. Cheers!

[thbar]

Hi @ttilberg! Thanks for asking!

I'll release aggregating transforms as part of Kiba 2.5.

I'll make an official release shortly!

[thbar]

@ttilberg it's ready - Kiba v2.5.0 is out. Happy coding!