Type model and API rework

[kMutagene]

This PR:

• Re-models Cvterm and CvUnit as records
• uses unified naming of CvTerm/CvUnit record fields across all functions:
  • Accession instead of id or AnnotationId
  • Value instead of name
  • RefUri (instead of ref)
• adds many convenience and accession functions on the modules/types that are actually used in the functions. Example: tryCvParam, a function of signature IParamBase -> CvParam option will not be on the CvParam type anymore, as functions prefixed with CvParam are expected to have actual CvParams as arguments.

[omaus]

Value instead of name

is "Value" not already taken by the real value?

I mean, you have a CvParam with

• TAN "TO:000000"
• Ref "TO"
• Name "Total traffic load"
• Value of string "563 KB"

[kMutagene]

Value instead of name

is "Value" not already taken by the real value?

I mean, you have a CvParam with

• TAN "TO:000000"
• Ref "TO"
• Name "Total traffic load"
• Value of string "563 KB"

I am talking about CvTerm, not CvParam.

[kMutagene]

We can however decide to use Name if that is more favorable. the thing is we had always 2 choices, and they were used inconsistently across the library, as CvTerm was described as:

/// Represents a term from a controlled vocabulary (Cv)
/// in the form of: id|accession ; name|value ; refUri
// ?Maybe []

[kMutagene]

Also, to be clear, what you are referring to as value is the ParamValue.

[omaus]

We can however decide to use Name if that is more favorable. the thing is we had always 2 choices, and they were used inconsistently across the library, as CvTerm was described as:

/// Represents a term from a controlled vocabulary (Cv)
/// in the form of: id|accession ; name|value ; refUri
// ?Maybe []

Yes, and I didn't like it.

Wouldn't it be more consistent when using "name"? Because if "value" in CvParam means a totally other thing than "value" in CvTerm, wouldn't this confuse people?

[kMutagene]

Is there a case where the Name/Value field represents something else than the name of an ontology term? if not I think we can use Name everywhere. Otherwise, I think Value is more generic and therefore better.

[omaus]

Is there a case where the Name/Value field represents something else than the name of an ontology term? if not I think we can use Name everywhere. If not, I think Value is more generic and therefore better.

OBO format goes with "name" for Term names which makes it more clear imho.
Besides that, I don't understand your double "if not" case. 😅

[kMutagene]

ready for review @omaus ;)