cwlformat

A prettifier for CWL code


Keywords
code-formatter, common-workflow-language, cwl
License
Apache-2.0
Install
pip install cwlformat==2022.2.18

Documentation

CWL Format

Tests PyPI version

CWL Format is a specification and a reference implementation for a very opinionated CWL code formatter.

It outputs CWL in a standardized YAML format. It has no settings or options because you have better things to do with your time. And because CWL Format is always correct.

This repository lists the formatting rules and also contains a Python implementation of the formatter.

pip install cwlformat
cwl-format unformatted.cwl > formatted.cwl

If you don't have a > py3.6 environment, you can use pipx:

pip install pipx  # from your < py3.6 environment
pipx ensurepath # ensures CLI application directory is on your $PATH
pipx install cwlformat --python python3.7 #tells pipx to set up a Py3.7 env for this app

Use programmatically in Python by doing

from cwlformat.formatter import cwl_format

formatted_text = cwl_format(unformatted_text)

or

from cwlformat.formatter import stringify_dict

as_dict = load_yaml(unformatted_text)
formatted_str = stringify_dict(as_dict)

Rules

  • Only comment lines at the top of the file, including blank lines, before the actual CWL code are preserved. All other comments are lost. Do not use this if all comments in the YAML are important to you.

  • If the first line does not start with #!/usr/bin/env the line #!/usr/bin/env cwl-runner is added to the top of the file.

  • All CWL fields are ordered systematically. The field order for specific fields have a defined precedence ("pinned fields"). Any fields not present in this file ("free fields") are printed after the pinned fields and ordered alphabetically.

  • A single blank line is added before the following fields if the parent structure is a process.

    • inputs
    • outputs
    • steps
    • requirements
    • hints
    • baseCommand
  • The pinned fields are defined in this YAML file.

  • Specific pinned field orderings are available for CommandLineTool, ExpressionTool and Workflow processes. All other types follow a generic pinned field list.

  • All strings that fit within 80 columns are expressed in flow style. Longer strings or strings with new lines are expressed in block style.

  • All lists and maps are expressed in block style

  • The ordering of all lists are preserved

  • Indentation is 2 spaces, including for lists

Conformance tests

A series of documents are found in the tests directory that can be used to check correctness of a formatter. The files named original-* are the input files and the files named formatted-* are the corresponding formatted documents. There are a mixture of YAML and JSON input files. Formatted files are always YAML.

CWL Exploder

This takes as input a packed workflow (workflow with all steps in lined) and splits it recursively into parts.

cwl-explode formatted-atac-seq-pipeline.cwl expected-exploded-atac-seq.cwl

Results in the exploded parent workflow and 52 sub-components