From 08d511b94c7b0c39b915f36d52095516709dd3a8 Mon Sep 17 00:00:00 2001
From: howsoRes <144272317+howsoRes@users.noreply.github.com>
Date: Thu, 31 Oct 2024 13:09:32 -0400
Subject: [PATCH] 22064: Adds rebalance_features to analyze(), MINOR (#317)

---
 howso/client/base.py    | 13 +++++++++++++
 howso/engine/trainee.py |  6 ++++++
 version.json            |  2 +-
 3 files changed, 20 insertions(+), 1 deletion(-)

diff --git a/howso/client/base.py b/howso/client/base.py
index 883562eb..50f5dcee 100644
--- a/howso/client/base.py
+++ b/howso/client/base.py
@@ -3639,6 +3639,7 @@ def analyze(
         num_analysis_samples: t.Optional[int] = None,
         num_samples: t.Optional[int] = None,
         p_values: t.Optional[Collection[float]] = None,
+        rebalance_features: t.Optional[t.Collection[str]] = None,
         targeted_model: t.Optional[TargetedModel] = None,
         use_case_weights: t.Optional[bool] = None,
         use_deviations: t.Optional[bool] = None,
@@ -3683,6 +3684,10 @@ def analyze(
             The number of samples used in calculating feature residuals.
         p_values : Collection of float, optional
             The p value hyperparameters to analyze with.
+        rebalance_features : Collection[str], optional
+            The list of features whose values to use to rebalance case
+            weighting of the data and to store into weight_feature.
+            Cannot be used with ablation.
         targeted_model : {"omni_targeted", "single_targeted", "targetless"}, optional
             Type of hyperparameter targeting.
             Valid options include:
@@ -3711,6 +3716,7 @@ def analyze(
 
         util.validate_list_shape(context_features, 1, "context_features", "str")
         util.validate_list_shape(action_features, 1, "action_features", "str")
+        util.validate_list_shape(rebalance_features, 1, "rebalance_features", "str")
         util.validate_list_shape(p_values, 1, "p_values", "int")
         util.validate_list_shape(k_values, 1, "k_values", "float")
         util.validate_list_shape(dt_values, 1, "dt_values", "float")
@@ -3761,6 +3767,7 @@ def analyze(
             num_samples=num_samples,
             analysis_sub_model_size=analysis_sub_model_size,
             p_values=p_values,
+            rebalance_features=rebalance_features,
             targeted_model=targeted_model,
             use_deviations=use_deviations,
             weight_feature=weight_feature,
@@ -3829,6 +3836,7 @@ def set_auto_analyze_params(
         num_analysis_samples: t.Optional[int] = None,
         num_samples: t.Optional[int] = None,
         p_values: t.Optional[Collection[float]] = None,
+        rebalance_features: t.Optional[t.Collection[str]] = None,
         targeted_model: t.Optional[TargetedModel] = None,
         use_deviations: t.Optional[bool] = None,
         use_case_weights: t.Optional[bool] = None,
@@ -3873,6 +3881,10 @@ def set_auto_analyze_params(
             When True, bypasses calculation of feature weights.
         bypass_hyperparameter_analysis : bool, optional
             When True, bypasses hyperparameter analysis.
+        rebalance_features : Collection[str], optional
+            The list of features whose values to use to rebalance case
+            weighting of the data and to store into weight_feature.
+            Cannot be used with ablation.
         targeted_model : Literal["omni_targeted", "single_targeted", "targetless"], optional
             Type of hyperparameter targeting.
             Valid options include:
@@ -3982,6 +3994,7 @@ def set_auto_analyze_params(
             "bypass_hyperparameter_analysis": bypass_hyperparameter_analysis,
             "bypass_calculate_feature_residuals": bypass_calculate_feature_residuals,
             "bypass_calculate_feature_weights": bypass_calculate_feature_weights,
+            "rebalance_features": rebalance_features,
             "targeted_model": targeted_model,
             "num_analysis_samples": num_analysis_samples,
             "analysis_sub_model_size": analysis_sub_model_size,
diff --git a/howso/engine/trainee.py b/howso/engine/trainee.py
index 228362e6..a3e33da9 100644
--- a/howso/engine/trainee.py
+++ b/howso/engine/trainee.py
@@ -960,6 +960,7 @@ def analyze(
         num_samples: t.Optional[int] = None,
         analysis_sub_model_size: t.Optional[int] = None,
         p_values: t.Optional[Collection[float]] = None,
+        rebalance_features: t.Optional[t.Collection[str]] = None,
         targeted_model: t.Optional[TargetedModel] = None,
         use_case_weights: t.Optional[bool] = None,
         use_deviations: t.Optional[bool] = None,
@@ -1001,6 +1002,10 @@ def analyze(
             randomly held-out and not included in calculations.
         p_values : Collection of float, optional
             The p value hyperparameters to analyze with.
+        rebalance_features : Collection[str], optional
+            The list of features whose values to use to rebalance case
+            weighting of the data and to store into weight_feature.
+            Cannot be used with ablation.
         targeted_model : {"omni_targeted", "single_targeted", "targetless"}, optional
             Type of hyperparameter targeting.
             Valid options include:
@@ -1043,6 +1048,7 @@ def analyze(
                 num_samples=num_samples,
                 analysis_sub_model_size=analysis_sub_model_size,
                 p_values=p_values,
+                rebalance_features=rebalance_features,
                 targeted_model=targeted_model,
                 use_deviations=use_deviations,
                 weight_feature=weight_feature,
diff --git a/version.json b/version.json
index 28bd6c79..29c43470 100644
--- a/version.json
+++ b/version.json
@@ -1,5 +1,5 @@
 {
   "dependencies": {
-    "howso-engine": "89.0.2"
+    "howso-engine": "89.1.0"
   }
 }