pymc-labs/CausalPy

Fix failing doctests

Closed this issue · 1 comments

We've currently got 12 failing doctests.

Initially I was unable to replicate the failing tests running make doctest locally. But after rebuilding the environment from scratch I did manage to replicate the failing tests. So the suspicion is that something has changed in a dependency.

I've not dug into the details too much, but possibly related to the expected outputs. Maybe a type hinting issue?

061     >>> model.fit(X, y)
Expected:
    Inference...
Got:
    <BLANKLINE>
    <BLANKLINE>
    Inference data with groups:
        > posterior
        > posterior_predictive
        > sample_stats
        > prior
        > prior_predictive
        > observed_data
        > constant_data

Tagging @jpreszler

On rebuilding the environment, I'm getting pytensor build issues. So I believe that we had previously gotten away with a pip based method to build the environment, but since moving to pymc>=5.14.0 we aren't getting away with this any more.

So my prediction is that #304 is a blocker for tests/doctests to pass locally, and #281 may be a blocker for remote tests to pass. Prediction was incorrect.