Developer Guide - Adding new functionality

Several utility functions (pipeline.utils) have been configured to make adding new features as easy as possible. Some importable functions of key importance:

  • get_proj_dir: Gets the project directory string from command line parameters

  • get/set_proj_file: Open any json formatted project file given a project directory and get/set data.

  • format_str: Pretty formatting, sets the string size to a given length by concatenating or adding whitespace.

  • get/set_codes: Get a list of project codes from any file or sub-directory file in a groupdir.

  • open_kerchunk: Open a kerchunk file with __protection__ from errors and taking into account different types or conditions.

Also a quick note on some pipeline.logs features.

  • FalseLogger(): Test class for when you don’t care about proper logging (not to be used in production).

1. New Modes in Assessor

New modes can be added to the assessor with the following method:

  • Create a new function that takes args and logger as inputs.

  • Add this function to the operations dict, with an appropriate activation string.

  • Add any new command line inputs to ArgParse (try to reuse or expand existing flags where possible)

2. New Processors for Compute

See pipeline.compute.compute_config to configure a new type of processor. You will need:

  • A configuration function to ensure all required parameters are set.

  • A processor class (that most likely inherits from pipeline.compute.ProjectProcessor)

  • A method of activating that processor (override_type option is always available, but could this processor be triggered automatically based on a scan result?)