niiri prefixes > urn / id are now uuids / project info removed

[thomasbtnfr]

This PR solves the following issues: Inria-Empenn#6 Inria-Empenn#7 Inria-Empenn#8 and Inria-Empenn#9

Fork linked

• [Bug]: Replace all niiri by urn Inria-Empenn/BEP028_BIDSprov#6 : all niiri prefixes have been replaced by urn

• [Bug]: Use UUID as identifiers Inria-Empenn/BEP028_BIDSprov#7 : all @id identifiers have been replaced by uuid. Normally the links have not been broken especially with the following example :

  • Error (two entities) :
    
  • Correct solution
    

• [Bug]: Remove project info Inria-Empenn/BEP028_BIDSprov#8 : the project information has been removed

"wasGeneratedBy": {  
    "@id": "INRIA",  
    "@type": "Project",  
    "wasAssociatedWith": {  
      "@id": "NIH",  
      "@type": "Organization",  
      "hadRole": "Funding"  
    }

• [Bug]: Identifier for the agent should be a uuid Inria-Empenn/BEP028_BIDSprov#9 : the agent's identifier has been replaced by a uuid instead of "@id": "exampleAgentID". Activities are directed to the agent by the agent id and not by its RRID.

Moreover, the keywords "AssociatedWith" and "GeneratedWith" need to have "with" as a prefix to draw the graphs automatically. This prefix is not indicated in the standard (https://docs.google.com/document/d/1vw3VNDof5cecv2PkFp7Lw_pNUTUo8-m8V4SIdtGJVKs/edit#heading=h.h15lzmuljsky).

[cmaumet]

Thank you! I have updated the title of this PR to be more explicit and updated sllightly the text in the first comment to add links to each issue explicitly. Can you try and solve the conflicts? (Note: most of the conflicts are in results and should be solved by running the exporter again).

[cmaumet]

Thanks for this! I'll review #82 first and then will come back to this one.

But before I forget: could you create an issue about your last comment:

Moreover, the keywords "AssociatedWith" and "GeneratedWith" need to have "with" as a prefix to draw the graphs automatically. This prefix is not indicated in the standard (https://docs.google.com/document/d/1vw3VNDof5cecv2PkFp7Lw_pNUTUo8-m8V4SIdtGJVKs/edit#heading=h.h15lzmuljsky).

We should be able to correct for that by updating the context but let's do it in another iteration/PR.

[thomasbtnfr]

The merge is now ok.

I also put back the removal of the max length of labels for those that are not mapped as it was done with Hermann and Cyril.
For example :
the label in the jsonld of cfg_basicio.file_dir.file_ops.file_move (38 characters) now stays the same and does not become file_dir.file_ops.file_move (the limit was set to 30 characters).

[cmaumet]

Thanks @thomasbtnfr! I'll review this now.

[cmaumet]

@thomasbtnfr: Have you tried here to use directly a UUID (without the conversion to string)? I would like to check how the namespace is handled by the uuid library (rather than having to add manually the urn namespace before).

[cmaumet]

(If we add the urn namespace manually then we will have to declare it somewhere, in the context and in the RDF python library)

[thomasbtnfr]

Related to this issue : Inria-Empenn#13

[satra]

urn does not need to be defined in the context, it's handled appropriately as an uri by most libraries i believe.

https://api.dandiarchive.org/api/dandisets/000350/versions/0.221219.1506/assets/6b78d427-92e0-4a8e-bc64-481cbddea1e2/ (look at the publisher section). also we don't define it anywhere in our context: https://github.com/dandi/schema/blob/master/releases/0.6.3/context.json

[cmaumet]

Thanks @satra! We opened Inria-Empenn#13 to work on this in a coming PR. I'll link your comment in there!