Pipeline Base Class

Pipeline Base Class#

The Pipeline class is the base class for all pipelines. It provides the following methods:

Base class for building pipelines in anomalib.

This module provides the abstract base class for creating pipelines that can execute jobs in a configurable way. Pipelines handle setting up runners, parsing configs, and orchestrating job execution.

Example

>>> from anomalib.pipelines.components.base import Pipeline
>>> class MyPipeline(Pipeline):
...     def _setup_runners(self, args: dict) -> list[Runner]:
...         # Configure and return list of runners
...         pass
...     def run(self, args: Namespace | None = None):
...         # Execute pipeline logic
...         pass

The base pipeline interface defines key methods that subclasses must implement:

  • _setup_runners(): Configure the runners that will execute pipeline jobs

  • run(): Execute the core pipeline logic

Pipelines can be used to implement workflows like training, inference, or benchmarking by composing jobs and runners in a modular way.

class anomalib.pipelines.components.base.pipeline.Pipeline#

Bases: ABC

Base class for pipeline.

static get_parser(parser=None)#

Create a new parser if none is provided.

Return type:

ArgumentParser

run(args=None)#

Run the pipeline.

Parameters:

args (Namespace) – Arguments to run the pipeline. These are the args returned by ArgumentParser.

Return type:

None

Contents