Harmonisation of names for attribute identifier

[floradanna]

In the database, the attribute for identifier is named:

• iri in the template_attributes table
• pid in the sample_attributes table

Since these two attribute names refer to the same concept (identifier), we could choose 1 name to be used in both tables.

[alaninmcr]

Are you sure they are the same thing? A PID must be permanent. There is no such requirement on IRI

[floradanna]

If I remember correctly, pid in sample_attributes table is not the seek_id of the attribute, it is a free text attribute (for now) where users can indicate if an attribute comes from a ontology or a schema by providing a link for it. The same applies to iri in template_attributes, I think. Please, correct me if I am wrong.

[stuzart]

the PID for the sample attribute is validated as being a valid IRI, based on it being a term from an ontology. the table field name shouldn't be changed to IRI as that is too vague

[floradanna]

I forgot that it gets validated. @vdkkia is iri in template_attributes also validated for being a valid IRI?

I don't have strong preference for names, but I would prefer to have consistency.

[alaninmcr]

@floradanna They only need to be consistent if they mean the same thing.

[vdkkia]

The iri in template_attributes is not getting validated now. Sample_types and templates have some concepts in common. As @stuzart said before, there must be a plan to consolidate them and take a unit thing out of them in the future. For now, we can add the same validation as sample_attributes to the template_attributes

[floradanna]

They mean the exact same thing. Names and functionalities can be harmonised and aligned when you guys see fit.

[floradanna]

the PID for the sample attribute is validated as being a valid IRI, based on it being a term from an ontology. the table field name shouldn't be changed to IRI as that is too vague

@stuzart I don't understand. the PID for the sample attribute is validated as being a valid IRI (like, "http://" or "ftp://", followed by a domain name or IP address, and a path to the resource ?), but it should not be labeled IRI?

What kind of validation occurs exactly: syntax, if it comes from OLS? or?

[stuzart]

the PID for the sample attribute is validated as being a valid IRI, based on it being a term from an ontology. the table field name shouldn't be changed to IRI as that is too vague

@stuzart I don't understand. the PID for the sample attribute is validated as being a valid IRI (like, "http://" or "ftp://", followed by a domain name or IP address, and a path to the resource ?), but it should not be labeled IRI?

What kind of validation occurs exactly: syntax, if it comes from OLS? or?

It is just validated that it's syntax is a valid IRI (or URI in this case). It doesn't need to be a resolvable URL, or necessarily come from OLS. ncbi:1234 would pass for example ( https://0mg.github.io/tools/uri/ ). Just what would be valid for an identifier from an ontology or for linked data. It's validated using a regular expression

[stuzart]

don't change the pid attribute to iri, it would be incorrect. It would be like change a 'title' attribute to just 'string'. pid is it's purpose, iri is just the syntax