feat: Add {file, name, dir, ext} to classNameSlug resolution

[mikestopcontinues]

Fixes #650
Fixes #571

Motivation

If you want to know where a particular class is coming from, you need more than a hash or even the original title of the class. How many .inner or .title classes have you seen in a large codebase, right?

Summary

This PR adds four new vars to classNameSlug resolution. It also adds the original two hash and title to the object, with the idea that a future breaking change might sandwich the function parameters down to just this one.

While just file would have been sufficient for function style, I added dir, name, ext for ease of use with string-style. Here's what you would expect in each:

{
  hash: 'qdegrke',
  title: 'displayName',
  file: 'this/is/your/module/Component/SubComp.tsx',
  name: 'SubComp',
  ext: 'tsx',
  dir: 'Component',
}

Test plan

Added tests.

[mikestopcontinues]

I don't think the circleci issue has to do with my code. I couldn't get it to run properly on my machine either, but yarn watch and yarn test both work.

[Anber]

Hi @mikestopcontinues! Thank you for your PR.

I don't think the circleci issue has to do with my code. I couldn't get it to run properly on my machine either, but yarn watch and yarn test both work.

There is a problem with types in new code:

src/visitors/GenerateClassNames.ts(150,59): error TS7053: Element implicitly has an 'any' type because expression of type 'string' can't be used to index type 'ClassNameSlugVars'.
  No index signature with a parameter of type 'string' was found on type 'ClassNameSlugVars'.

You can either relax the definition of ClassNameSlugVars by changing it to Record<string, string> or replace v in slugVars with some more strict type guard (like const isSlugVar = (v: string): v is keyof ClassNameSlugVars => v in slugVars).

All checks can be run by yarn prepare

[callstack-bot]

Hey @mikestopcontinues, thank you for your pull request 🤗.
The coverage report for this branch can be viewed here.

[mikestopcontinues]

@Anber Sorry about that. I just assumed what the circleci issue was. I added one additional test for non-existing replacement patterns to make sure the logic remained consistent.