The attribute standard_relationship should be a collection

[andrew2net]

fixes #5

[ronaldtse]

@andrew2net can you add the spec fixture for the failure so we don't encounter it again? You can just copy that file from rawbib. Thanks.

[andrew2net]

@ronaldtse I can add fixtures. But we need to create another test for this cases because lutaml-model creates collection automatically even there is no collection: true option. So round-trip test doesn't fall without the option. But using Ieee::Idams::Publication's data is tricky because we need to handle 3 value types. For example without collection: true option abstract's value can be:

• nil if there is no any abstracts
• ArticleAbstract instance when there is only one abstract
• [ArticleAbstract] array of abstract instances if there are more than one abstract

With the option we always get an array and can handle it same and simplify code in the relaton-ieee.
I seems there are other attributes what need to be collection. I'll create new tests as soon as I find all of them.

[ronaldtse]

@andrew2net I think this is a problem in lutaml-model for being implicit or "not caring" the actual model.

This just means that ArticleAbstract is clearly a collection because it could be 0, 1 and *, i.e. [0..*].

I think when collection: true is not given, then we must enforce there is 1 and only 1 instance. Ping @HassanAkbar .

[andrew2net]

I think this is a problem in lutaml-model for being implicit or "not caring" the actual model.

@ronaldtse agree. If lutaml-model caries the actual model then the round-test will fall without collection option when an attribute is actually collection. So we won't need addition tests in this gem.

[ronaldtse]

Agree with @andrew2net , I've posted the issue at:

• Be strict about adherence to cardinality via collection: true lutaml/lutaml-model#185

[ronaldtse]

@andrew2net thank you, please feel free to merge!