Workflow RO-Crate 1.0 suggestions #22

stain · 2021-11-12T11:44:43Z

Split into 0.2 and 1.0-DRAFT (to become 1.0/) with symlinks for current versions. Redirect instead?
Make RO-Crate 1.1 compatibility explicit 86af752 although with a SHOULD for the declared conformsTo (thus forward compatible)
HowTo should reference programmingLanguage to CWL 52a804f
license can be overridden on individual files 766be08
Entity for README.md should have encodingFormat and about: ./ dfd8e20
ComputationalWorkflow Bioschemas conformance 415c59c - just a SHOULD? What about FormalParameter ?
More about zipping 10e4e8b
conformsTo also include PID of this profile ad31675 according to https://www.researchobject.org/ro-crate/1.2-DRAFT/profiles#declaring-conformance-of-an-ro-crate-profile

Content negotiation would be via https://w3id.org/workflowhub/workflow-ro-crate/1.0 (see https://github.com/perma-id/w3id.org/tree/master/workflowhub) which seems to work:

(base) stain@xena:~/src/about/Workflow-RO-Crate/1.0-DRAFT$ curl -H "Accept: application/ld+json;profile=https://w3id.org/ro/crate" https://w3id.org/workflowhub/workflow-ro-crate/1.0
<!DOCTYPE HTML PUBLIC "-//IETF//DTD HTML 2.0//EN">
<html><head>
<title>302 Found</title>
</head><body>
<h1>Found</h1>
<p>The document has moved <a href="https://about.workflowhub.eu/Workflow-RO-Crate/1.0/ro-crate-metadata.json">here</a>.</p>
<hr>
<address>Apache/2.4.29 (Ubuntu) Server at w3id.org Port 443</address>
</body></html>

Now I think we may need to discuss several of these items before releasing this.

You can preview this in https://github.com/workflowhub-eu/about/blob/workflow-ro-crate-1.0/Workflow-RO-Crate/1.0-DRAFT/index.md

Also CWL 1.2 has abstract CWL which we want

I set this as a SHOULD to not force the FormalParameters - not yet supported in WorkflowHub.

I don't think we should be picky about the filename, as anyone consuming a zip can simply look for ro-crate-metadata.json inside and thus determine it is an RO-Crate. I think it is true you want the ro-crate-metadata.json straight in the root, and do not like somefolder/ro-crate-metadata.json? .jsonld -> .json

moved signposting to the profile

…ut into workflow-ro-crate-1.0

stain · 2021-11-12T12:04:06Z

Rendering of the updated Profile Crate
https://raw.githack.com/workflowhub-eu/about/workflow-ro-crate-1.0/Workflow-RO-Crate/1.0-DRAFT/ro-crate-preview.html

assuming authors are Finn, Alan, Stian

fbacall · 2021-11-15T14:09:44Z

Workflow-RO-Crate/1.0-DRAFT/example/ro-crate-metadata.json

+      "@type": [
+        "File",
+        "SoftwareSourceCode",
+        "HowTo"


Shouldn't this be ComputationalWorkflow? Since this file is the actual workflow, not the abstract CWL.

You are right, this example should perhaps be amended to have both a ComputationalWorkflow (which could be Nextflow) and a Abstract CWL HowTo

simleo · 2021-11-18T10:34:01Z

@stain should we have explicit types for the workflow languages (instead of #cwl, #galaxy, etc.), like you suggested for Workflow Testing RO-Crate test service / engine types? I remember starting with things like #jenkins and #planemo, trying to do the same as Workflow RO-Crate, but then we changed to https://w3id.org/ro/terms/test#JenkinsService and https://w3id.org/ro/terms/test#PlanemoEngine after your suggestion. Since Workflow Testing RO-Crate is derived from Workflow RO-Crate I think they should be consistent in this respect.

simleo · 2021-11-18T10:44:51Z

Related to the above: since https://w3id.org/ro/terms/test#PlanemoEngine should be a well-defined entity we ended up putting the version in the referencing entity instead:

        {
            "@id": "test/test1/sort-and-change-case-test.yml",
            "@type": [
                "File",
                "TestDefinition"
            ],
            "conformsTo": {"@id": "https://w3id.org/ro/terms/test#PlanemoEngine"},
            "engineVersion": ">=0.70"
        },
        {
            "@id": "https://w3id.org/ro/terms/test#PlanemoEngine",
            "@type": "SoftwareApplication",
            "name": "Planemo",
            "url": {"@id": "https://github.com/galaxyproject/planemo"}
        }

The use case I had in mind was multiple tests needing different engine versions, since this would not be valid:

        {
            "@id": "https://w3id.org/ro/terms/test#PlanemoEngine",
            "@type": "SoftwareApplication",
            "name": "Planemo",
            "url": {"@id": "https://github.com/galaxyproject/planemo"},
            "version": 0.70
        },
        {
            "@id": "https://w3id.org/ro/terms/test#PlanemoEngine",
            "@type": "SoftwareApplication",
            "name": "Planemo",
            "url": {"@id": "https://github.com/galaxyproject/planemo"},
            "version": 0.71
        }

I think the same would happen in the Workflow RO-Crate case. What if there's a secondary workflow that's written in the same language as the main workflow, but in a different version? You can't repeat the same workflow language entity with two different versions, can you?

…o-crate/1.0

stain · 2021-11-18T11:52:49Z

Agree, @simleo - rather than everyone defining their own new #cwl etc. they can now use the PID from this profile.

However I don't want to use #cwl then inside the profile as that would mean expand to the versioned URL of the profile, languages should survive across.

So I will change the languages to be under the https://w3id.org/workflowhub/workflow-ro-crate# namespace which would resolve to the current profile crate.

Not sure what are implication of this on https://github.com/seek4science/seek/blob/master/config/default_data/workflow_classes.yml and corresponding [match_from_metadata algorithm]https://github.com/seek4science/seek/blob/workflowhub/app/models/workflow_class.rb#L43()

For now think they should still also be explicit in the crate, that means algorithm would still be able to match on identifier etc. Longer term perhaps the Ruby code would have the Workflow Crate loaded for dereferencing licenses and languages.

stain · 2021-11-18T11:58:15Z

On the problem with versions, I think you would need a second contextual entity. And then it is harder to make @id as most workflow systems are not good at defining PIDs for their releases or even to have a page about their language syntax.

Now this profile itself does not say you need a version on the programmingLanguage, but https://www.researchobject.org/ro-crate/1.1/workflows.html#workflow-runtime-and-programming-language does.

Ideally some kind of prov:specializationOf the versionless PID? We've not done that kind of versioning of entities or crates in RO-Crate yet. See https://practicalprovenance.wordpress.com/2016/05/07/tracking-versions-with-pav/ for possibilities.

I tried for instance to improve on the URL for Knime, since https://www.knime.com/ is a company and not a programming language - but they don't have a page about the .knwf format.

stain · 2021-11-18T12:02:28Z

I also fixed that the old examples had odd nesting with @id on identifier and url - as these are defined to have the data type URL they should be flat strings (the identifier is the string itself, not whatever typed object you point at)

But I think that algorithm currently only tries to go into @id of a nested object.

See #22 (comment)

stain · 2021-11-18T12:09:45Z

https://w3id.org/workflowhub/workflow-ro-crate/1.0 now live.

RO-Crates previews:

In the Profile Crate I changed to use http://schema.org/Guide as discussed in hackathon. Not sure about PID entity which is needed to not violate MUST end in / requirement of https://www.researchobject.org/ro-crate/1.1/root-data-entity.html#direct-properties-of-the-root-data-entity which also does not permit additional @type. To fix in RO-Crate 1.2

stain added 15 commits November 12, 2021 11:30

Moved to 1.0-DRAFT to track progress

b188446

0.2 as-is with symlinks back

e1a582e

1.0 metadata

6bdcc94

RO-Crate 1.1 and .json not .jsonld

86af752

HowTo should reference #cwl ?

52a804f

Also CWL 1.2 has abstract CWL which we want

About license override

766be08

README.md should have encodingFormat and be about crate

dfd8e20

About ComputationalWorkflow conformance

415c59c

I set this as a SHOULD to not force the FormalParameters - not yet supported in WorkflowHub.

conformsTo should include this permalink

ad31675

Include markdown and about on README.md entity

f8b814a

somewhat updated to match the index.md

86d9ab4

Updated previews

b95a474

moved signposting to the profile

Updated previews

d9f3f87

moved signposting to the profile

Merge branch 'workflow-ro-crate-1.0' of github.com:workflowhub-eu/abo…

2bf2d60

…ut into workflow-ro-crate-1.0

Added metadata to HTML

6c518f6

assuming authors are Finn, Alan, Stian

fbacall reviewed Nov 15, 2021

View reviewed changes

stain added 2 commits November 18, 2021 02:51

Galaxy example

7269c76

link to CWL Viewer

259664c

fbacall approved these changes Nov 18, 2021

View reviewed changes

Updated example and now conformsTo BioSchemas

2cf1725

stain added 4 commits November 18, 2021 11:33

Use Guide and Recommendation

25a0db8

recommendations under Guide

0992587

clarify identifier on test/examples

94fb9d6

Release Workflow-RO-Crate 1.0 https://w3id.org/workflowhub/workflow-r…

9e5a55f

…o-crate/1.0

stain added 2 commits November 18, 2021 12:02

PIDs for programmingLanguages

b28f326

See #22 (comment)

fixed timestamp

4223a07

stain merged commit 3cf2d15 into master Nov 18, 2021

stain deleted the workflow-ro-crate-1.0 branch November 18, 2021 12:05

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Workflow RO-Crate 1.0 suggestions #22

Workflow RO-Crate 1.0 suggestions #22

stain commented Nov 12, 2021 •

edited

Loading

stain commented Nov 12, 2021

fbacall Nov 15, 2021

stain Nov 16, 2021

simleo commented Nov 18, 2021 •

edited

Loading

simleo commented Nov 18, 2021 •

edited

Loading

stain commented Nov 18, 2021

stain commented Nov 18, 2021

stain commented Nov 18, 2021

stain commented Nov 18, 2021

Workflow RO-Crate 1.0 suggestions #22

Workflow RO-Crate 1.0 suggestions #22

Conversation

stain commented Nov 12, 2021 • edited Loading

stain commented Nov 12, 2021

fbacall Nov 15, 2021

Choose a reason for hiding this comment

stain Nov 16, 2021

Choose a reason for hiding this comment

simleo commented Nov 18, 2021 • edited Loading

simleo commented Nov 18, 2021 • edited Loading

stain commented Nov 18, 2021

stain commented Nov 18, 2021

stain commented Nov 18, 2021

stain commented Nov 18, 2021

stain commented Nov 12, 2021 •

edited

Loading

simleo commented Nov 18, 2021 •

edited

Loading

simleo commented Nov 18, 2021 •

edited

Loading