Monty.Experiment.RecognitionStrategy.MontyIsDone (ExTbpMonty.Core v0.1.0)

Copy Markdown View Source

The original policy: the model says it is done, or it does not.

Mirrors tbp/monty/experiment/recognition_strategy.py::MontyIsDone, which is what Monty uses when a config names no policy at all — so this is the floor every experiment stands on, and the one an episode with no step limit falls back to.

Summary

Functions

call(model, counter, state)

Done when the model says so; asks no other question.

new(options \\ [])

The policy, taking no configuration.

Functions

call(model, counter, state)

@spec call(map(), term(), map()) ::
  Monty.Experiment.RecognitionStrategy.RecognitionResult.t()

Done when the model says so; asks no other question.

new(options \\ [])

@spec new(keyword()) :: {:ok, {module(), %{}}}

The policy, taking no configuration.

Monty's constructor takes no arguments either; the empty option list is here so that every policy is built the same way, and an option is still refused — by the factory, before this is asked (Monty.Experiment.RecognitionStrategy).