vitest `test` / `it`, and `describe` should follow the same pattern as `@playwright/test` for test context extensions

Clear and concise description of the problem

As a developer I'd love if the API for test context (fixtures) mirrored that of playwright test

This would be better for the following reasons:

It makes context modifications easier to share
Its not reliant on setting up beforeEach before a test to modify its context. You could easily share customized test / describe and/or it blocks. Using beforeEach is clunky and not easily sharable.
Its (perhaps subjectively) more ergonomic and discoverable. Certainly works better for typing with TypeScript, which for large teams (especially enterprises) this is a huge win.

Suggested solution

Allow for test context to be set on the test function level, via test /it, and/or describe. e.g.

// my-test.ts
import { describe as base } from 'vitest;
import type { Todo } from './todo';
import * as faker from '@faker-js/faker';

// Declare the types of your fixtures.
interface TodoFixtures {
  todos: Todo[];
};

// Extend base test by providing "todo"
// This new "describe" can be used in multiple test files, and each of them will get the fixtures.
export const describe = base.extend<TodoFixtures>({
  todo: async (use) => {
    let i = 0;
    const todos = [];
    for (i; i < 25; i++) {
           todos.push({
              due: faker.date.between('2020-01-01T00:00:00.000Z', '2030-01-01T00:00:00.000Z'),
              assigned: faker.name.findName(),
              text: faker.lorem.lines()
          })
     }
     use(todos)
  },
});

Alternative

No response

Additional context

This is largely brought about because on my team of 24+ engineers we would like to be able to have standard patterns for faking data. One of the requirements when it comes to sharing API data is having fake datasets that can be re-used for testing and tested against schema so that can test that our validators work correctly. Its also the same fake data factory datasets used for our API test responses, and for data transformers, among other things. We really started to scale the @playwright/test approach well and our team is finding this API to be rather clunky by comparison, and having to declare a module augmentation adds more overhead and is prone to error.

Validations

[X] Follow our Code of Conduct
[X] Read the Contributing Guidelines.
[X] Read the docs.
[X] Check that there isn't already an issue that request the same feature to avoid creating a duplicate.

Jul 17 '22 22:07 ScottAwesome

I have had the same thought recently, and went ahead and made an implementation a few days ago, of "vitest-fixture" a package "on top of" Vitest, that provides an API very similar to the Playwright fixture feature:

https://www.npmjs.com/package/vitest-fixture

Please let me know if this would be useful for your use-case, and let me know if you have improvement suggestions!

Jul 19 '22 12:07 larsthorup

I'd love to see this feature land in Vitest. @larsthorup 's package is good. But I think this would be a fantastic addition to vitest.

Let's discuss the API for this. Playwright's API is really battle tested and it evolved over the versions of playwright (was Folio previously). I've been watching the project evolve, and I think they landed in a nice API.

vitest-fixture matches Playwright's API very closely so we could bake it in vitest core.

Thoughts?

Oct 21 '22 12:10 haikyuu

I'd love to see this too. Makes typing and composing context much simpler, and I would hope an implementation would carry over Playwright's feature to smartly initialize fixtures based on usage.