EO application developers use Click to create command line tools to process EO data.
Click is a Python package for creating beautiful command line interfaces in a composable way with as little code as necessary. It’s the “Command Line Interface Creation Kit”. It’s highly configurable but comes with sensible defaults out of the box. It aims to make the process of writing command line tools quick and fun while also preventing any frustration caused by the inability to implement an intended CLI API.
To deploy these applications on an Exploitation Plaform on the Cloud, the EO application developers generate a CWL document that becomes an Application Package.
The repo contains a helper Python module that exploits the information contained in a Click context object to generate the CWL document and eventually the parameters.
This helper Python module can thus be used to add "CWL dump" behaviour to a Python CLI that uses Click by adding an import and call to the dump(ctx) function provided in the helper Python module.
Install this Python module in your conda environment using conda or mamba (see this article to learn about mamba):
pip install -c terradue click2cwlNote: Add click2cwl dependency and terradue channel in your environment.yml file, e.g.:
name: my_env
channels:
- terradue
- conda-forge
dependencies:
- python
pip:
- click2cwlFirst import the click2cwl.dump function with:
from click2cwl import dumpThen, in the Python application entry point function, update and enrich the Click function decorator with the information for generating the CWL document.
The information defined at the Click function decorator is explained below:
The @click.command decorator must:
- set the
allow_extra_argsflag toTrue. This allows using the command line interface to pass the runtime information to generate the CWL. - set the
short_helpto define the CWL Workflow class label - set the
helpto define the CWL Workflow class doc
@click.command(
short_help="This is Workflow class label",
help="This is Workflow class doc",
context_settings=dict(
ignore_unknown_options=True,
allow_extra_args=True,
),
)The @click.option decorator must:
- define the option type that can be
click.Pathfor aDirectory, aclick.Filefor aFileor the default, aStringto map asstring - set the
helpto define the CWL Workflow class doc and label - set the
requiredflag toFalseif the parameter is optional
@click.option(
"--input_path",
"-i",
"input_path",
type=click.Path(),
help="this input path",
multiple=False,
required=True,
)
def cli(): ...The @click.argument decorator must:
- define the argument type (same types as for
@click.option) - must OMIT the
help(this is a limitation ofclick, see pallets/click#587) - it will be automatically marked as
required=True
@click.argument(
"input_path",
type=click.Path(),
)
def cli(): ...The advantage of using @click.argument over @click.option is that it will correspond to a positional argument,
for cases where to option flag applies (i.e.: --<some-option>). However, it comes with limitations in terms of
additional parameters offered by click.
Finally, invoke the dump function in your Click decorated function with:
@click.command ...
@click.option ...
@click.pass_context
def main(ctx, **kwargs):
click2cwl.dump(ctx)
print("business as usual")
print(kwargs)
if __name__ == '__main__':
main()Note: See the examples folder to discover typical use cases.
There are multiple exporters to deal with the various contents:
click2cwl.clt.CLTExport: CWL CommandLineTool directlyclick2cwl.cwlexport.CWLExport: CWL Workflows and its underlying CommandLineToolclick2cwl.paramexport.ParamExport: CWL Job Parameters
These can be invoked directly instead of the @click2cwl.dump
operation to extract the contents separately from the CLI invocation.
Build your application with:
hatch buildWhen the app help is invoked, it will only show the arguments defined by the click.decorators:
myapp --helpbut it now supports additional args to drive the generation of the CWL document and associated parameters:
The additional args are:
--dumpcwl|params|clt. Example--dump cwl --dump paramswill dump the CWL document and the CWL parameters template in YAML.cltwill dump the CWlCommandLineToolclass only (no Workflow)--requirementwithrequirement=valuewhere requirement here is one of"coresMin","coresMax","ramMin","ramMax". Example:--requirement ramMax=1 --requirement ramMin=2--docker <docker image>if set, theDockerRequirementhint is set to pull the<docker image>--envsets environment variables in the CWL withenv_var=env_var_value. Example--env A=1 --env B=2where A and B are the environment variables to set in the CWLEnvVarRequirementrequirement--wall-time <wall-time in seconds>if set, theToolTimeLimitcWL requirement is set and the CWL document version is set to v1.1 (instead of default 1.0)--cwl-version, if set, thecwlVersionuses that value