✨ validate_or_build function introduced

[benrutter]

I'm pretty excited about this one, not sure how many people's workflows it'll fit with, but seems like a really great way to get off the ground with introducing testing for a whole range of datasets.

Functionality looks something like this:

import pandas as pd
import wimsey

from somewhere import storage_options

df = (
    pd.read_csv("some_csv.csv")
    .pipe(wimsey.validate_or_build, "s3://my-test-store/some_csv/tests.json")
    .assign(i_dont_know=lambda df: df["just_giving_an_example"] + df["i guess"])
    ...
)

If this is, say, a data engineering pipeline, the first time it runs, it'll build out some tests, and then it'll test against those each further time.