From e48fd7bb5b6957f444a54b23ad3817da3ee80010 Mon Sep 17 00:00:00 2001
From: Drew Herren <drewherrenopensource@gmail.com>
Date: Thu, 20 Nov 2025 14:12:09 -0600
Subject: [PATCH 1/2] Update Python BCF to work without propensities

---
 stochtree/bcf.py | 35 +++++++++++++----------------------
 1 file changed, 13 insertions(+), 22 deletions(-)

diff --git a/stochtree/bcf.py b/stochtree/bcf.py
index 10bf3f3c..2e1b2ad2 100644
--- a/stochtree/bcf.py
+++ b/stochtree/bcf.py
@@ -932,16 +932,15 @@ def sample(
         if sample_sigma2_leaf_tau is not None:
             if not isinstance(sample_sigma2_leaf_tau, bool):
                 raise ValueError("sample_sigma2_leaf_tau must be a bool")
-        if propensity_covariate is not None:
-            if propensity_covariate not in [
-                "prognostic",
-                "treatment_effect",
-                "both",
-                "none",
-            ]:
-                raise ValueError(
-                    "propensity_covariate must be one of 'prognostic', 'treatment_effect', 'both', or 'none'"
-                )
+        if propensity_covariate not in [
+            "prognostic",
+            "treatment_effect",
+            "both",
+            "none",
+        ]:
+            raise ValueError(
+                "propensity_covariate must be one of 'prognostic', 'treatment_effect', 'both', or 'none'"
+            )
         if b_0 is not None:
             b_0 = check_scalar(
                 x=b_0,
@@ -1663,15 +1662,6 @@ def sample(
         ] = 0
 
         # Update covariates to include propensities if requested
-        if propensity_covariate not in [
-            "none",
-            "prognostic",
-            "treatment_effect",
-            "both",
-        ]:
-            raise ValueError(
-                "propensity_covariate must equal one of 'none', 'prognostic', 'treatment_effect', or 'both'"
-            )
         if propensity_covariate != "none":
             feature_types = np.append(
                 feature_types, np.repeat(0, propensity_train.shape[1])
@@ -1700,9 +1690,10 @@ def sample(
                 variable_weights_tau = np.append(
                     variable_weights_tau, np.repeat(1 / num_cov_orig, propensity_train.shape[1])
                 )
-        variable_weights_variance = np.append(
-            variable_weights_variance, np.repeat(0.0, propensity_train.shape[1])
-        )
+            # For now, propensities are not included in the variance forest
+            variable_weights_variance = np.append(
+                variable_weights_variance, np.repeat(0.0, propensity_train.shape[1])
+            )
 
         # Renormalize variable weights
         variable_weights_mu = variable_weights_mu / np.sum(variable_weights_mu)

From 0a4d8c868ffe34a40e391ce0062cb5477cc597d2 Mon Sep 17 00:00:00 2001
From: Drew Herren <drewherrenopensource@gmail.com>
Date: Thu, 20 Nov 2025 15:08:41 -0600
Subject: [PATCH 2/2] Added unit test to make sure BCF runs without propensity
 scores

---
 test/python/test_bcf.py | 13 +++++++++++++
 1 file changed, 13 insertions(+)

diff --git a/test/python/test_bcf.py b/test/python/test_bcf.py
index eca2a5ff..c5a1446f 100644
--- a/test/python/test_bcf.py
+++ b/test/python/test_bcf.py
@@ -194,6 +194,19 @@ def test_binary_bcf(self):
         # Check treatment effect prediction method
         tau_hat = bcf_model.predict(X=X_test, Z=Z_test, terms="cate")
 
+        # Check that we can run BCF without propensities
+        bcf_model = BCFModel()
+        general_params = {"propensity_covariate": "none"}
+        bcf_model.sample(
+            X_train=X_train,
+            Z_train=Z_train,
+            y_train=y_train,
+            num_gfr=num_gfr,
+            num_burnin=num_burnin,
+            num_mcmc=num_mcmc,
+            general_params=general_params,
+        )
+
     def test_continuous_univariate_bcf(self):
         # RNG
         random_seed = 101