CQ3 - Configuration files

[simleo]

What are the configuration files used in a workflow execution step?

ChooseAction? Though maybe the crate generator should just merge the params with the other ones if it can parse the config file. To link to the config file as a black box instead we probably need a new property.

[simleo]

There are two main cases here:

• Configuration of the application being run (the workflow, or a tool). We have no concrete use case for this yet, but we could simply represent this as one of the inputs of the workflow / tool, likely an optional one ("valueRequired": "False") since in most application there's a default configuration and the configuration file is only used to tweak some features.
• Configuration of the workflow engine. We have the StreamFlow use case from @GlassOfWhiskey, and there are likely other workflow engines where this is relevant. Looks like there is no suitable property in SoftwareApplication, so we need to add a custom one to ro-terms. Looking at other SoftwareApplication properties, softwareConfiguration seems the most logical name.

Example:

{
    "@id": "#c0080ce4-e220-44bf-92b3-c59e6d590059",
    "@type": "SoftwareApplication",
    "name": "StreamFlow 0.2.0.dev2",
    "softwareConfiguration": "streamflow.yml"
},
{
    "@id": "streamflow.yml",
    "@type": "File",
    "name": "StreamFlow configuration file",
    "encodingFormat": "application/yaml"
}

On a side note, it seems that the application/yaml media type is not official yet.

[GlassOfWhiskey]

Maybe it is worth adding a pointer to the SoftwareApplication it refers to

[GlassOfWhiskey]

Maybe it is worth adding a pointer to the SoftwareApplication it refers to

Oh sorry I didn't see it. I was looking for it the other way around :)

[simleo]

I've been rethinking this, and the above does not work. The reason is that the configuration is not an intrinsic property of the workflow engine, but rather one of the parameters of a specific execution of the engine, which can change from one execution to another, so the description of multiple engine runs (not forbidden by the spec, and there's been talk about a use case for it yesterday in a BY-COVID meeting) in general is not possible. The configuration file should instead be added to the object of the OrganizeAction that represents the engine run:

{
    "@id": "#e55c4723-7814-4cef-b3b6-96c1dbf1ae9b",
    "@type": "SoftwareApplication",
    "name": "StreamFlow 0.2.0.dev2"
},
{
    "@id": "#7ff2f0b6-0294-4da5-9ecc-5846b8aa4e66",
    "@type": "OrganizeAction",
    "instrument": {"@id": "#e55c4723-7814-4cef-b3b6-96c1dbf1ae9b"},
    "name": "Run of StreamFlow 0.2.0.dev2",
    "object": [
        {"@id": "7ff2f0b6-0294-4da5-9ecc-5846b8aa4e66/streamflow.yml"},
        {"@id": "#a203c665-668d-4488-bf57-5b2eedf77905"},
        ...
    ],
    "result": {"@id": "#9984d778-7cd8-49ea-984d-7c58a0404f85"}
},
{
    "@id": "7ff2f0b6-0294-4da5-9ecc-5846b8aa4e66/streamflow.yml",
    "@type": "File",
    "encodingFormat": "application/yaml"
},
{
    "@id": "#a203c665-668d-4488-bf57-5b2eedf77905",
    "@type": "ControlAction",
    "instrument": {"@id": "predictions.cwl#extract-tissue-low"},
    "object": {"@id": "#465dafb2-66cf-4af9-a6cf-b8fea0e8acc9"}
},
...

[simleo]

See #39