[AsadGudzevic94]

Description

With external ml_m predictions, the preliminary IPW step now takes the propensities of the units in train_inds_1 (m_hat["preds"][train_inds_1]), in the same order as d_train_1 and y_train_1.

New test in test_pq_external_predictions.py: with a propensity model that ignores its training data, external predictions of it give the same estimate as the model itself (to 1e-9). It fails on main for treatment=1.

Reference to Issues or PRs

Fixes #431

PR Checklist

• The title of the pull request summarizes the changes made.
• The PR contains a detailed description of all changes and additions.
• References to related issues or PRs are added.
• The code passes all (unit) tests.
• Enhancements or new feature are equipped with unit tests.
• The changes adhere to the PEP8 standards.

[SvenKlaassen]

Thanks @AsadGudzevic94,
I just adjusted the test slightly.