jsonld-context-autogen missing the base IRI

[tekrajchhetri]

@puja-trivedi @djarecka I noticed the context file is missing the default IRI.

Issue: Whenever you do the conversion of JSON-LD to ttl file, the file path is taken as base IRI, while it should be something like http://example.org/NIMP/<ID>..

In the previous version, it's not the case.

[tekrajchhetri]

Update:

The issue is in the JSON-LD file. So, there is no need to update the context file.

In current, i.e., rapid release file: We do not have any prefix and also no base IRI causing the aforementioned issue. I suggest to make use of prefix, explicitly rather than having base, as is the best practice.

      "was_derived_from": [
        "EC-VYUUTE137687",
        "EC-UYBXXW324656"
      ]

In old JSON-LD file: We have NIMP prefix so everything's fine.

"was_derived_from": [
        "NIMP:LA-GFHFDE415771LOMHPL202182",
        "NIMP:LA-LYZZRW951279LOMHPL202182",
        "NIMP:LA-QVRSQW524460LOMHPL202182",
        "NIMP:LA-BFDXZY775863LOMHPL202182",
        "NIMP:LA-VUVPPX234898LOMHPL202182",
        "NIMP:LA-HHEABG177343LOMHPL202182", 
      ],

[puja-trivedi]

@tekrajchhetri - I took a look at the library_generation model and the NIMP prefix is already set to http://example.org/NIMP/. See line#18 -> https://github.com/brain-bican/models/blob/main/linkml-schema/library_generation.yaml.

While we did not explicitly define the prefix in the spreadsheet; linkml automatically fills in any missing prefix (that is why it is not missing from the .yaml file).

[tekrajchhetri]

@puja-trivedi Yes, there exist NIMP prefix but in the JSON-LD files that you've shared with me, they're missing the NIMP prefix as I've pointed out in the earlier example (https://github.com/brain-bican/models/issues/88#issuecomment-2489311092). The same is shown in the image below, which is a snapshot of a JSON-LD file. Also, you can see from the image blow that not in all cases NIMP prefix is missing.

It could be the issue with the translator that you're using to generate JSON-LD files. Also a suggestion, I think it would be better to add some automated checks to ensure there's always a prefix associated.