Allow enum name payload

[EtienneWallet]

Context

When supplying the payload of a an enum which might contains fields, the user has to provide the discriminant.
This doesn't seem very user friendly as in a smart-contract the user defines the variants by names and the names hold much more meaning.

Proposed changes

I propose to allow the direct use of variant names when setting the payload of an EnumValue.
The logic is to detect when a name was provided instead of a discriminant and convert it before resuming operations.
For that, a dictionary names_to_discriminants was introduced to allow the user to provide the conversion means from variant names to variant discriminants

For example, if the user provides a list payload, this would allow to go from [44, 7, 8] to ["VariantD", 7, 8]

[github-actions]

Coverage report

Click to see where and how coverage changed

File	Statements	Missing	Coverage	Coverage (new stmts)	Lines missing
multiversx_sdk/abi
abi.py
constants.py
enum_value.py					85, 110, 120
Project Total

This report was generated by python-coverage-comment-action

[popenta]

Can we maybe skip this check? I think it will help allow for the old flow of initialization.
e.g. someone manually initializes an EnumValue(fields_provider=...) and then calls set_payload()

[EtienneWallet]

The issue I have with removing this check is that we would have to create three duplicates checks each time we want to use names_to_discriminants within the function set_payload.
I would argue that this brings unnecessary verbosity and is not coherent with how fields_provider is treated

A third way maybe: I haven't consider how all the classes are configured but could we create a class method constructor EnumValue.from_definition(cls, def : EnumDefinition) -> EnumValue ? This would give the user a simple way to create the EnumValue without having to worry about fields_provider or names_to_discriminants.

[popenta]

We want to allow the old flow, so I think removing the check here is still the way to go. Regarding duplicating the code, that can be a extracted in a private method that will be called wherever we use self.names_to_discriminants. Maybe something like:

def _self.ensure_names_to_discriminants_is_set(self):
    if self.names_to_discriminants is None:
            raise ValueError("...")

The idea about EnumValue.from_definition(...) is great and we should do that for StructValue, as well. Let's do it in a separate PR.

[EtienneWallet]

The switch to lazy check has been made as you wanted
instead of having a function that only does the check, I preferred to create a public function to convert a variant name to its discriminant. This way, we are sure that the check is done and the conversion is included
If we ever need to use names_to_discriminants differently, we could always separate the check and the conversion

Don't hesitate to tell me if this doesn't suit you, I'll change it.

[popenta]

Looks good to me!