Pipelines

Writing a Pipeline

The basics of writing a 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 are 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 stdatamodels.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 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.