Use workflow DOI as crate base URI instead of arcp URI where possible

[elichad]

Looking at the knowledge graph again, most crates use an arcp URI as the base URI, for example:

<arcp://uuid,00e690fb-885b-5c8d-8ac5-bab6bc0bf3e4/> a schema1:Dataset ;
    dct:conformsTo <https://w3id.org/ro/crate/1.1>,
        <https://w3id.org/workflowhub/workflow-ro-crate/1.0> ;
    schema1:description "Workflow for Creating a large disease network from various datasets and databases for IBM, and applying the active subnetwork identification method MOGAMUN." ;
    schema1:hasPart <arcp://uuid,00e690fb-885b-5c8d-8ac5-bab6bc0bf3e4/workflow-681-7e4122fc690aef90cafdd96758c561f2dde60555-diagram.svg>,
        <arcp://uuid,00e690fb-885b-5c8d-8ac5-bab6bc0bf3e4/workflows/workflow.cwl> ;
    schema1:identifier "https://doi.org/10.48546/workflowhub.workflow.681.7" ;
    schema1:isBasedOn "https://github.com/jdwijnbergen/IBM_ASI_workflow.git" ;
    schema1:license "MIT" ;
    schema1:mainEntity <arcp://uuid,00e690fb-885b-5c8d-8ac5-bab6bc0bf3e4/workflows/workflow.cwl> ;
    schema1:name "Research Object Crate for Inclusion Body Myositis Active Subnetwork Identification Workflow" ;
    schema1:sdDatePublished "2024-08-21 15:11:16 +0100" ;
    schema1:url "https://workflowhub.eu/workflows/681/ro_crate?version=7" .

However, many of them have an identifier on the root data entity with a DOI that can be used to persistently identify the crate - using these as base URIs would make the graph more intercompatible with other data sources.

[stain]

I think it should be arcp URIs because the RO-Crates will have content in their ./ but a glue triples calculated from doi/identifier to arcp should be added.

There is an idea that @ptsefton has suggested to add a doi subscheme to arcp, so one can do arcp://doi,10.48546/workflowhub.workflow.681.7/-- this would only work together with Signposting (as we have in WorkflowHub) as you would need programmatic way to link a DOI to one-and-only-one download.