/python-cwlgen

Generation of CWL programmatically. Available types: CommandLineTool and DockerRequirement

Primary LanguagePythonMIT LicenseMIT

python-cwlgen (deprecated)

Deprecated

python-cwlgen is now deprecated, please use cwl-utils >= 0.4.

from cwl_utils import parser_v1_0

# You could alias this as cwlgen to simplify the migration
from cwl_utils import parser_v1_0 as cwlgen

Migration notes:

  • Method changes

    • get_dict() → save()
    • parse_cwl(cwlfile)load_document(cwlfile)
    • parse_dict → No super clear analogue, but loaded through _RecordLoader(CommandLineTool) || _UnionLoader((CommandLineToolLoader, ...workflow + other loaders)
  • Field names:

    • Uses camelCase instead of snake_case
    • No more special field names, eg:
      • tool_id | workflow_id | input_id | etc → id
      • StepInput: inputsin_
  • Other notes:

    • Classes aren't nested anymore, ie: cwlgen.InitialWorkDirRequirement.Direntcwl_utils.parser_v1_0.Dirent.

    • Take care if you're migrating to a newer spec, as some classes might have changed names (notably: InputParameter -> WorkflowInputParameter)

    • Don't forget to catch all references of cwlgen, as missing one (or using mismatch versions of the parser) will cause:

      raise RepresenterError('cannot represent an object: %s' % (data,))
      ruamel.yaml.representer.RepresenterError: cannot represent an object: 
      <cwlgen.common.CommandInputArraySchema object at 0x1100a5780>

If you have issues with the migration, please see this thread or raise an issue on CWLUtils.


Original README

Build Status codecov Documentation Status PyPI version

Python-cwlgen is a python library for the generation of CWL programmatically. It supports the generation of CommandLineTool, Workflow and DockerRequirement. The library works for both Python 2.7.12+ and 3.6.0.


Common Workflow Language

Common Workflow Language (CWL) is a language to describe workflows. The user guide gives a gentle explanation of what its goals are, but broadly:

  1. Stop writing bash scripts for long complex jobs.
  2. Take pipelines anywhere (portability).
  3. Enforce reproducibility guidelines.

This python repository is a python wrapper for most of the classes (work in progress), allowing you to build the structure of the workflow in Python and have this module generate and export CWL for you.

Nb: This doesn't check the logic of Workflows or CommandLineTools for you.
CWLTool has a --validate mode that you can use.

Quick-start guide

You can install python-cwlgen through pip with the following command:

pip install cwlgen

How it works?

This repository contains a number of python classes that mirror the CWL specifications (Workflow| CommandLineTool). In essence, each class's initializer has all of the properties it expects, which may be another object. The classes include the relevant docstrings to give you context of classes and their properties.

The examples/ folder contains some simple examples, however in essence you simply initialize the class you're trying to build. An initializer for a class has all of the properties it expects which may be another object.

Creating a CommandLineTool

import cwlgen

tool_object = cwlgen.CommandLineTool(tool_id="echo-tool", base_command="echo", label=None, doc=None,
                 cwl_version="v1.0", stdin=None, stderr=None, stdout=None, path=None)
tool_object.inputs.append(
    cwlgen.CommandInputParameter("myParamId", param_type="string", label=None, secondary_files=None, param_format=None,
                 streamable=None, doc=None, input_binding=None, default=None)
)

# to get the dictionary representation:
dict_to_export = tool_object.get_dict()

# to get the string representation (YAML)
yaml_export = tool_object.export_string()

# print to console
tool_object.export()

# print to file
tool_object.export("echotool.cwl")

References

CWL is developed by an informal, multi-vendor working group consisting of organizations and individuals aiming to enable scientists to share data analysis workflows. The CWL project is on Github.

Known issues

  • SchemaDefRequirement doesn't parse the types subfield into the specific types (InputRecordSchema | InputEnumSchema | InputArraySchema), but leaves them as a simple dictionary.