Pipeline

class Pipeline(steps, breakpointers, pruners=None, once_in_iterations=1)

Bases: BaseEstimator

An estimator that fits a mixture model via a configurable iterative process.

The pipeline executes a sequence of defined steps in a loop. After each full sequence of steps, pruning strategies are applied, and stopping conditions are checked. The loop continues until a breakpointer signals to stop. This allows for building complex, multi-stage estimation algorithms, such as variants of the EM algorithm.

The core components used for configuration are:

• PipelineStep

• Breakpointer

• Pruner

Parameters:

• steps (Sequence[PipelineStep]) – An ordered sequence of steps to be executed in each iteration of the pipeline.

• breakpointers (Sequence[Breakpointer]) – A sequence of strategies that define the stopping conditions for the iterative process. This list cannot be empty.

• pruners (Sequence[Pruner] | None, optional) – A sequence of strategies for removing components from the mixture model during fitting. Defaults to None, meaning no pruning is performed.

• once_in_iterations (int, optional) – The logging frequency. A value of n means logging occurs every n iterations. Defaults to 1 (log every iteration).

Variables:

• steps (list[PipelineStep]) – The ordered list of operations to be performed in each iteration.

• breakpointers (list[Breakpointer]) – The list of objects that determine when the fitting process should terminate.

• pruners (list[Pruner]) – The list of objects that may remove components from the mixture during the fitting process.

• logger (IterationsHistory) – object that collects comprehensive information about each iteration of a Pipeline estimator.

Raises:

ValueError – If the sequence of steps is empty or invalid (i.e., a step is followed by a step not listed in its available_next_steps), or if the sequence of breakpointers is empty.

Methods

fit(X, mixture)

Fits the mixture model to the data using the configured pipeline.