initital attempt to write unit test for _parameter_estimates

Author: am-kaiser

tests/qoi/test_marginal_cdf_extrapolation_with_importance_sampling.py

@@ -0,0 +1,42 @@
+"""
+Plan:
+    - Unit test:
+        - _parameter_estimates: input and output weights are attached to the same samples
+    - Integration test:
+        - successful run of minimal version of the qoi using a deterministic model with importance sampling
+"""
+
+# %%
+import torch
+from botorch.models.deterministic import GenericDeterministicModel
+
+from axtreme.qoi.marginal_cdf_extrapolation import MarginalCDFExtrapolation
+
+
+def test_parameter_estimates_consistency_of_weights(gp_passthrough_1p: GenericDeterministicModel):
+    importance_samples = torch.Tensor([1, 2, 3, 4])
+    importance_weights = torch.Tensor([0.1, 0.2, 0.3, 0.4])
+
+    samples = [importance_samples, importance_weights]
+
+    env_sample = torch.tensor([[[0], [1], [2]]], dtype=torch.float64)
+    qoi_estimator = MarginalCDFExtrapolation(env_iterable=env_sample, period_len=3)
+
+    posterior_samples, importance_weights_qoi = qoi_estimator._parameter_estimates(gp_passthrough_1p)
+
+    print(importance_samples, importance_weights)
+    print(posterior_samples, importance_weights_qoi)
+
+
+# %%
+if __name__ == "__main__":
+    import sys
+    from pathlib import Path
+
+    root_dir = Path("../../")
+    sys.path.append(str(root_dir))
+    # from conftest import gp_passthrough_1p
+
+    test_parameter_estimates_consistency_of_weights(gp_passthrough_1p)
+
+# %%