_adjustment_set_from_formula now uses a set rather than a sorted list

Author: jmafoster1

causal_testing/estimation/abstract_regression_estimator.py

@@ -101,7 +101,7 @@ def _adjustment_set_from_formula(self):
             tree = ast.parse(code, mode="eval")
             adjustment_set = adjustment_set.union(self._get_adjusted_variables(tree))
 
-        self.adjustment_set = sorted(list(adjustment_set))
+        self.adjustment_set = adjustment_set
 
     def _setup_covariates(self, df: pd.DataFrame) -> pd.Series:
         """

causal_testing/estimation/linear_regression_estimator.py

@@ -33,7 +33,7 @@ def gp_formula(
         seed: int = 0,
     ):
         # pylint: disable=too-many-arguments
-        """
+        r"""
         Use Genetic Programming (GP) to infer the regression equation from the data.
 
         :param df: The data to use.

tests/estimation_tests/test_linear_regression_estimator.py

@@ -71,7 +71,7 @@ def test_complex_formula_adjustment_set(self):
         linear_regression_estimator = LinearRegressionEstimator(
             treatment_variable="cut", outcome_variable="price", formula=formula
         )
-        self.assertEqual(linear_regression_estimator.adjustment_set, ["C", "color", "depth"])
+        self.assertEqual(linear_regression_estimator.adjustment_set, {"C", "color", "depth"})
 
     def test_complex_formula_adjustment_set_no_c(self):
         """
@@ -81,7 +81,7 @@ def test_complex_formula_adjustment_set_no_c(self):
         linear_regression_estimator = LinearRegressionEstimator(
             treatment_variable="cut", outcome_variable="price", formula=formula
         )
-        self.assertEqual(linear_regression_estimator.adjustment_set, ["caret", "color"])
+        self.assertEqual(linear_regression_estimator.adjustment_set, {"caret", "color"})
 
     def test_complex_formula_adjustment_set_no_dependent(self):
         """