Hi there! After writing approximately the same bunch of test functions at two jobs, I've extracted a little pytest library for doing expectation testing the way that I like to, and published in the usual places. So, I thought I'd tell y'all about it.

Essentially, it allows lots of variations of this kind of test:

def test_compute(resources):  
    input = resources.load_json("input")  
    output = compute(input)  
    resources.expect_json(output, "output")  

The input data and expected results will be located in predictable places relative to the test function, and in a format that is nice to work with as a developer. No magic pickles or multiple expectations squashed into a single file that you're afraid to edit.

It works well with text data, json and pydantic models and others may appear over time. Test file path creation is completely configurable, but I think the defaults are reasonable.

I'd love for it to see some usage and to get some feedback about how to make it better.

That's it!