Pipeline (skypy.pipeline)¶
Running SkyPy¶
The pipeline package contains the functionality to run a SkyPy
simulation from end to end. This is implemented in the SkyPyDriver
class and can be called using the command line script
skypy.
Reference/API¶
skypy.pipeline Package¶
This module provides methods to pipeline together multiple models with dependencies and handle their outputs.
skypy.pipeline.driver Module¶
Driver module.
This module provides methods to run pipelines of functions with dependencies and handle their results.
Classes¶
Class for running pipelines. |
Class Inheritance Diagram¶
skypy.pipeline.scripts.skypy Module¶
skypy Command Line Script¶
skypy is a command line script that runs a pipeline of functions defined in
a config file to generate tables of objects and write them to file.
Using skypy to run one of the example pipelines and write the outputs to
fits files:
$ skypy –config examples/herbel_galaxies.yaml –format fits
Config Files¶
Config files are written in yaml format. The top level should contain the
fields cosmology and/or tables. cosmology should contain a
dictionary configuring a function that returns an
astropy.cosmology.Cosmology object. tables should contain a set of
nested dictionaries, first giving the names of each table, then the names of
each column within each table. Each column should contain a dictionary
configuring a function that returns an array-like object.
Each step in the pipeline is configured by a dictionary specifying:
‘function’ : the name of the function
‘module’ : the name of the the module to import ‘function’ from
‘args’ : a list of positional arguments (by value)
‘kwargs’ : a dictionary of keyword arguments
‘requires’ : a dictionary of keyword arguments
Note that ‘kwargs’ specifices keyword arguments by value, wheras ‘requires’ specifices the names of previous steps in the pipeline and uses their return values as keyword arguments.