Writing a Pipeline

There are two ways to go about writing a pipeline depending on how much flexibility is required.

1. A linear pipeline defines a simple linear progression of steps where each step has a single input and a single output flowing directly into the next step.

2. A flexible pipeline allows the pipeline to be defined in Python code and all of the tools that implies, such as loops, conditionals and multiple inputs and/or outputs.

Linear pipeline

To create a linear pipeline, one inherits from the LinearPipeline class and adds a special member pipeline_steps to define the order of the steps:

from jwst.stpipe import LinearPipeline

# Some locally-defined steps
from . import FlatField, RampFitting

class ExampleLinearPipeline(LinearPipeline):
    This example linear pipeline has only two steps.
    pipeline_steps = [
        ('flat_field', FlatField),
        ('ramp_fitting', RampFitting)

The pipeline_steps member is a list of tuples. Each tuple is a pair (name, class) where name is the name of the specific step, and class is the step’s class. Both are required so the same step class can be used multiple times in the pipeline. The name is also used for the section headings in the pipeline’s configuration file.

Flexible pipeline

The basics of writing a flexible Pipeline are just like Writing a step, but instead of inheriting from the Step class, one inherits from the Pipeline class.

In addition, a Pipeline subclass defines what its Steps so that the framework can configure parameters for the individual Steps. This is done with the step_defs member, which is a dictionary mapping step names to step classes. This dictionary defines what the Steps are, but says nothing about their order or how data flows from one Step to the next. That is defined in Python code in the Pipeline’s process method. By the time the Pipeline’s process method is called, the Steps in step_defs will be instantiated as member variables.

For example, here is a pipeline with two steps: one that processes each chip of a multi-chip FITS file, and another to combine the chips into a single image:

from jwst.stpipe import Pipeline

from jwst.datamodels import ImageModel

# Some locally-defined steps
from . import FlatField, Combine

class ExamplePipeline(Pipeline):
    This example pipeline demonstrates how to combine steps
    using Python code, in some way that it not necessarily
    a linear progression.

    step_defs = {
        'flat_field': FlatField,
        'combine': Combine,

    def process(self, input):
        with ImageModel(input) as science:

            flattened = self.flat_field(science, self.multiplier)

            combined = self.combine(flattened)

        return combined

    spec = """
    multiplier = float()     # A multiplier constant

When writing the spec member for a Pipeline, only the parameters that apply to the Pipeline as a whole need to be included. The parameters for each Step are automatically loaded in by the framework.

In the case of the above example, we define two new pipeline configuration parameters for the flat field file and the output filename.

The parameters for the individual substeps that make up the Pipeline will be implicitly added by the framework.