Extend `typescript-enum-array` with option for zod-supported string union

[JamisonHolt]

This is namely a request for making the typescript-enum-array plugin have a zod-friendly way of declaring array exports in a way that is easily mappable to string union type declarations as well. Consider the following:

type MyGeneratedUnion = 'foo' | 'bar' | 'baz';
export const MY_GENERATED_UNION = ['foo' | 'bar' | 'baz'] as const; // using the constArrays option

The above works somewhat well with zod's z.enum type inference, where z.enum(MY_GENERATED_UNION).parse(someVar) will be type-inferred as 'foo' | 'bar' | 'baz'.

My main grievance is that any future devs inspecting a var will see the type as 'foo' | 'bar' | 'baz', which is harder to trace to the source. It's not clear at first glance that this is a generated string union from our GraphQL codegen. However, leaving constArrays disabled makes things worse:

Now we've lost even more information for the variable type annotation, making it harder to trace back to the source AND less useful for design patterns like switch-case exhaustiveness.

As a result, I'd like to propose a configuration option that would let us generate more zod-compatible, tuple declarations. For example:

I hope it's clear why this would be nice. In a large codebase, it can easily become harder for devs to track down where a type's definition comes from. Having the ability for zod to provide the named type makes tracking down the declaration much easier, especially for junior devs or anyone who's still ramping up.
I'm down to own this change, and here is a very rough idea of what it might look like: f60bd5f

[eddeee888]

Hiya! This is interesting!

I'm trying this example on https://zod-playground.vercel.app/:

type MyEnum = 'A' | 'B';

const MyEnumArray1: MyEnum[] = ['A', 'B'];
const test1 = z.enum(MyEnumArray1).parse('test')

const MyEnumArray2: [MyEnum, ...MyEnum[]] = ['A', 'B'];
const test2 = z.enum(MyEnumArray2).parse('test')

It seems both have the same result. Could you let me know what I'm doing wrong please?

Screen.Recording.2025-10-26.at.10.27.55.pm.mov