records: update json schema to include container images

[nancyhamdan]

fixes #3386
updated records' JSON schema to include container images metadata fields

[tiborsimko]

It may be useful to specify that the image is used for analysing the data, for example:

• Description of the container image that is recommended for analysing these data

(Alternatively, if we could have several images linked here, one recommended for analysis, one how the data was born, we could even introduce a new property called "purpose" or some such... But at this moment it's probably an overkill, since we plan to store only the images useful for data analyses. Unless there could be several images in the future for different purposes... CC @katilp )

[katilp]

Yes, for now, we do not have plans to link to other images than that to be used for analysing the data.

(I could only imagine a separate image for MC to generate a new similar MC dataset with some changes in the input but we are not providing those images and I'd rather not link them to the records in any case)

[tiborsimko]

Let's start again by the same word as the field is called (recid) so that readers can form better mental association:

• Record ID of the container image (if it exists as another Open Data record)

[tiborsimko]

It may be useful to provide example values in the documentation:

• Registry type where the image can be found (e.g. dockerhub, cerngitlab)

[tiborsimko]

Again here:

• URI where the container image is located

However, we may perhaps be better off using a common term FQIN here to denote fully-qualified image name concept, i.e. the field would be called "fqin" instead of "uri" and the description would explain it as follows:

• Fully Qualified Image Name (FQIN) where the container image is located (repository/name:tag)

We can discuss live also with @katilp before merging

[tiborsimko]

P.S. BTW it the "fqin" field name risks looking a bit weird, we could use some longer name variant such as "fully_qualified_name", or just "name", and document that the name should be FQIN in the schema.

[katilp]

Good points! I would go for "name" and documenting it in the schema as @tiborsimko proposed

[tiborsimko]

Thanks, rebased against latest QA.

After merge, please prepare another PR with a small example file containing the data as well as the output format Jinja snippet displaying it, as we had talked about IRL.