Improve WrrocRenderer

[fbartusch]

Add workflow file to crate. Correct path to output files in metadata file. Add CreateActions for tool executions. Add ControlActions. Add SoftwareApplications for processes. Add HowToSteps. Add agent information (orcid, name) to nextflow.config that is incorporated into metadata file.

[simleo]

I've installed the plugin with make install and used the following nextflow.config:

plugins {
    id 'nf-prov@1.1.0'
}

params {
    outdir = 'results'
}

prov {
    enabled = true
    formats {
        wrroc {
            file = "${params.outdir}/ro-crate-metadata.json"
            overwrite = true
            agent {
                name = "John Doe"
                orcid = "https://orcid.org/0000-0000-0000-0000"
            }
        }
    }
}

I had to pin the plugin's version because otherwise Nextflow downloads nf-prov-1.2.2 and uses that (which fails because it does not support wrroc as a format). Then I ran nextflow run tests/test.nf and got a results RO-Crate. Here are the problems I've found:

• The CreateAction corresponding to test.nf (#606ce08c-cac1-4777-8a1b-83f497a0c63d) has an "id" property which should be "@id" instead, so the RO-Crate is not valid. I've manually changed that to "@id" in order to continue with the review.

• The other CreateAction entities have a "result" property that has a nested list, which shouldn't be there. For instance, this:

"result": [
    [
        {
            "@id": "r1.foo.1.txt"
        },
        {
            "@id": "r1.foo.2.txt"
        }
    ]
],

should be:

"result": [
    {
        "@id": "r1.foo.1.txt"
    },
    {
        "@id": "r1.foo.2.txt"
    }
],

I made these changes manually to move on with the review. At this point, Runcrate reads the crate (runcrate report results) without crashing. Here's the output:

action: #606ce08c-cac1-4777-8a1b-83f497a0c63d
  instrument: test.nf (['File', 'SoftwareSourceCode', 'ComputationalWorkflow', 'HowTo'])
  started: 2024-06-20T11:59:14.123266278+02:00
  ended: 2024-06-20T11:59:14.897242324+02:00
  outputs:
    r1.foo.1.txt
    r1.foo.2.txt
    r2.foo.2.txt
    r2.foo.1.txt
    r3.foo.1.txt
    r3.foo.2.txt

action: #71e33a99d3115bb57a7cdfc1b220fbbf
  step: test.nf#main/RNG
  instrument: #Script_923b8caac4a54f39@66e17eff (SoftwareApplication)
  outputs:
    r1.foo.1.txt
    r1.foo.2.txt

action: #a9b6314416b001fb728ff45fa29d6f5e
  step: test.nf#main/RNG
  instrument: #Script_923b8caac4a54f39@66e17eff (SoftwareApplication)
  outputs:
    r2.foo.1.txt
    r2.foo.2.txt

action: #45b19d72e22ba819cf02b88e4977bc86
  step: test.nf#main/RNG
  instrument: #Script_923b8caac4a54f39@66e17eff (SoftwareApplication)
  outputs:
    r3.foo.1.txt
    r3.foo.2.txt

• The "@id" of ControlAction 3968b3e5-b5c8-4b24-9e03-d118801e589e should be #3968b3e5-b5c8-4b24-9e03-d118801e589e.

• The ComputationalWorkflow has outdir as an input parameter, but the only parameter you can set when launching the workflow is constant (which is not listed as an input parameter in the RO-Crate metadata). If I try to change the value of outdir from results to crate (either in nextflow.config or on the command line with --outdir crate), files are still output into results; moreover, in this case the ro-crate-metadata.json and test.nf files are missing (they can be found in the crate dir but only if you create it manually before running the workflow).

I've also noticed a weird thing: if you run rm results -rf && nextflow run tests/test.nf (without changing the value of outdir in nextflow.config) repeatedly, sometimes (like 1 in 20 times) the ro-crate-metadata.json file is missing from the results dir.

[fbartusch]

Thanks for the comprehensive feedback @simleo . The next time I will check if runcrate can parse the document before asking for feedback ...

I fixed most of the issues, but this causes some headache:

sometimes (like 1 in 20 times) the ro-crate-metadata.json file is missing from the results dir.

I can reproduce the problem. I wrote a script that removes the results directory, reruns the pipeline and checks if ro-crate-metadata.json ist present. It seems that the workflowOutputs mapping is missing entries for some output files.

Next I will use the newest Nextflow version and maybe the newest nf-prov version, maybe this problem is addressed there. For me it doesn't seem that the problem originates from the WrrocRenderer itself.

[simleo]

I've run the workflow again with the plugin at ac05da2, using the same nextflow.config. The issues I've reported in my previous comment seem to be solved (the sometimes-missing metadata file problem should be solved after the integration of c92b21f, if I understand correctly). Great job! This is another round of review.

• if the workflow is run without passing parameters (nextflow run tests/test.nf), the #constant FormalParameter is missing from the RO-Crate metadata (only #outdir is present). If it's run with nextflow run tests/test.nf --constant bar, #constant is present. It should be present in both cases.

• The CreateAction corresponding to the workfow execution has an empty "object". It should contain references to PropertyValue entities corresponding to the values taken by the parameters during the specific run. For instance, when run with nextflow run tests/test.nf --outdir crate --constant bar, the result should look like this:

{
    "@id": "#outdir",
    "@type": "FormalParameter",
    "name": "outdir",
    ...
},
{
    "@id": "#constant",
    "@type": "FormalParameter",
    "name": "constant",
    ...
},
{
    "@id": "#b5236af9-95e9-42b8-9121-a6a57755ee1b",
    "@type": "CreateAction",
    "instrument": {
        "@id": "test.nf"
    },
    "object": [
        {"@id": "#outdir-pv"},
        {"@id": "#constant-pv"}
    ],
    "result": [
        {"@id": "r3.bar.2.txt"},
        {"@id": "r3.bar.1.txt"},
        {"@id": "r2.bar.1.txt"},
        {"@id": "r2.bar.2.txt"},
        {"@id": "r1.bar.1.txt"},
        {"@id": "r1.bar.2.txt"}
    ]
    ...
},
{
    "@id": "#outdir-pv",
    "@type": "PropertyValue",
    "exampleOfWork": {"@id": "#outdir"},
    "name": "outdir",
    "value": "crate"
},
{
    "@id": "#constant-pv",
    "@type": "PropertyValue",
    "exampleOfWork": {"@id": "#constant"},
    "name": "constant",
    "value": "bar"
}

When run with nextflow run tests/test.nf the outcome should be the same except for the "value" fields of the PropertyValue entities, which should be "results" and "foo" (the defaults). Note that the PropertyValue instances link to the corresponding FormalParameters via exampleOfWork.

• The three CreateActions corresponding to the executions of RNG should also have an "object" pointing to PropertyValue instances, which in turn should be connected via exampleOfWork to FormalParameters of the SoftwareApplication corresponding to RNG.

• Since the output RO-Crate conforms to the Workflow RO-Crate profile, its Root Data Entity must have a "license". This can be specified in nextflow.config, like agent.

• nextflow.config should be included in the output RO-Crate. In the metadata file, it should be linked via object from the OrganizeAction, as specified in the Provenance Run Crate profile.

[fbartusch]

First, thank you for the review @simleo.

• Parameters are now added to RO-Crate metadata regardless if they are specified in the configuration or inside Nextflow's workflow files. This was simple, just use session.getBinding().getParams() this should return all parameters that are active during workflow execution.

• CreateAction of the workflow references now PropertyValues via object. PropertyValue link to the FormalParameter via exampleOfWork.

• If a license is specified in the configuration (e.g. license = "Apache-2.0") it's added to the Root Data Entity. If the specified string is a valid URL it's added with "license": "{@ id: < URL >}". But the CreativeWork entry for the URL of the license is missing in the RO-Crate metadata. Is this needed or does the URL is enough? If yes I try to add it.

• I started to think about the nextflow.config thing. There several places where the configuration can reside. The documentation explains it . There can be four different configuration files that take precedence in a specific order. Maybe it doesn't suffice to just copy all the files into the crate, because how can we decide which one takes precedence over the others? I'm thinking about compiling one 'effective' configuration file containing the values that are valid during workflow exeuction and put that one into the crate.

• I still have to do the PropertyValue stuff to the three CreateActions corresponding to the RNG executions.

[bentsherman]

session.config should contain the resolved configuration. it also contains the resolved params as session.config.params