feature: extracted analysis config generation for bias

Author: aws-byeldos

src/sagemaker/clarify.py

@@ -971,7 +971,7 @@ def _run(
     def run_pre_training_bias(
         self,
         data_config,
-        data_bias_config,
+        bias_config,
         methods="all",
         wait=True,
         logs=True,
@@ -986,7 +986,7 @@ def run_pre_training_bias(
 
         Args:
             data_config (:class:`~sagemaker.clarify.DataConfig`): Config of the input/output data.
-            data_bias_config (:class:`~sagemaker.clarify.BiasConfig`): Config of sensitive groups.
+            bias_config (:class:`~sagemaker.clarify.BiasConfig`): Config of sensitive groups.
             methods (str or list[str]): Selects a subset of potential metrics:
                 ["`CI <https://docs.aws.amazon.com/sagemaker/latest/dg/clarify-bias-metric-class-imbalance.html>`_",
                 "`DPL <https://docs.aws.amazon.com/sagemaker/latest/dg/clarify-data-bias-metric-true-label-imbalance.html>`_",
@@ -1022,7 +1022,7 @@ def run_pre_training_bias(
         """  # noqa E501  # pylint: disable=c0301
         analysis_config = _AnalysisConfigGenerator.bias_pre_training(
             data_config,
-            data_bias_config,
+            bias_config,
             methods
         )
         if job_name is None:
@@ -1043,7 +1043,7 @@ def run_pre_training_bias(
     def run_post_training_bias(
         self,
         data_config,
-        data_bias_config,
+        bias_config,
         model_config,
         model_predicted_label_config,
         methods="all",
@@ -1063,7 +1063,7 @@ def run_post_training_bias(
 
         Args:
             data_config (:class:`~sagemaker.clarify.DataConfig`): Config of the input/output data.
-            data_bias_config (:class:`~sagemaker.clarify.BiasConfig`): Config of sensitive groups.
+            bias_config (:class:`~sagemaker.clarify.BiasConfig`): Config of sensitive groups.
             model_config (:class:`~sagemaker.clarify.ModelConfig`): Config of the model and its
                 endpoint to be created.
             model_predicted_label_config (:class:`~sagemaker.clarify.ModelPredictedLabelConfig`):
@@ -1106,7 +1106,7 @@ def run_post_training_bias(
         """  # noqa E501  # pylint: disable=c0301
         analysis_config = _AnalysisConfigGenerator.bias_post_training(
             data_config,
-            data_bias_config,
+            bias_config,
             model_predicted_label_config,
             methods,
             model_config
@@ -1200,23 +1200,14 @@ def run_bias(
                   the Trial Component will be unassociated.
                 * ``'TrialComponentDisplayName'`` is used for display in Amazon SageMaker Studio.
         """  # noqa E501  # pylint: disable=c0301
-        analysis_config = data_config.get_config()
-        analysis_config.update(bias_config.get_config())
-        analysis_config["predictor"] = model_config.get_predictor_config()
-        if model_predicted_label_config:
-            (
-                probability_threshold,
-                predictor_config,
-            ) = model_predicted_label_config.get_predictor_config()
-            if predictor_config:
-                analysis_config["predictor"].update(predictor_config)
-            if probability_threshold is not None:
-                analysis_config["probability_threshold"] = probability_threshold
-
-        analysis_config["methods"] = {
-            "pre_training_bias": {"methods": pre_training_methods},
-            "post_training_bias": {"methods": post_training_methods},
-        }
+        analysis_config = _AnalysisConfigGenerator.bias(
+            data_config,
+            bias_config,
+            model_config,
+            model_predicted_label_config,
+            pre_training_methods,
+            post_training_methods,
+        )
         if job_name is None:
             if self.job_name_prefix:
                 job_name = utils.name_from_base(self.job_name_prefix)
@@ -1375,22 +1366,22 @@ def explainability(
         return analysis_config
 
     @staticmethod
-    def bias_pre_training(data_config, data_bias_config, methods):
+    def bias_pre_training(data_config, bias_config, methods):
         analysis_config = data_config.get_config()
-        analysis_config.update(data_bias_config.get_config())
+        analysis_config.update(bias_config.get_config())
         analysis_config["methods"] = {"pre_training_bias": {"methods": methods}}
         return analysis_config
 
     @staticmethod
     def bias_post_training(
             data_config,
-            data_bias_config,
+            bias_config,
             model_predicted_label_config,
             methods,
             model_config
     ):
         analysis_config = data_config.get_config()
-        analysis_config.update(data_bias_config.get_config())
+        analysis_config.update(bias_config.get_config())
         analysis_config["methods"] = {"post_training_bias": {"methods": methods}}
         (
             probability_threshold,
@@ -1401,6 +1392,34 @@ def bias_post_training(
         _set(probability_threshold, "probability_threshold", analysis_config)
         return analysis_config
 
+    @staticmethod
+    def bias(
+        data_config,
+        bias_config,
+        model_config,
+        model_predicted_label_config,
+        pre_training_methods="all",
+        post_training_methods="all",
+    ):
+        analysis_config = data_config.get_config()
+        analysis_config.update(bias_config.get_config())
+        analysis_config["predictor"] = model_config.get_predictor_config()
+        if model_predicted_label_config:
+            (
+                probability_threshold,
+                predictor_config,
+            ) = model_predicted_label_config.get_predictor_config()
+            if predictor_config:
+                analysis_config["predictor"].update(predictor_config)
+            if probability_threshold is not None:
+                analysis_config["probability_threshold"] = probability_threshold
+
+        analysis_config["methods"] = {
+            "pre_training_bias": {"methods": pre_training_methods},
+            "post_training_bias": {"methods": post_training_methods},
+        }
+        return analysis_config
+
     @staticmethod
     def _common(analysis_config):
         analysis_config["methods"]["report"] = {

tests/unit/test_clarify.py

@@ -1347,3 +1347,33 @@ def test_analysis_config_generator_for_bias_post_training(data_config, data_bias
                               'model_name': 'xgboost-model',
                               'probability': 'pr'}}
     assert actual == expected
+
+
+def test_analysis_config_generator_for_bias(data_config, data_bias_config, model_config):
+    model_predicted_label_config = ModelPredictedLabelConfig(
+        probability="pr",
+        label_headers=["success"],
+    )
+    actual = _AnalysisConfigGenerator.bias(
+        data_config,
+        data_bias_config,
+        model_config,
+        model_predicted_label_config,
+        pre_training_methods="all",
+        post_training_methods="all",
+    )
+    expected = {'dataset_type': 'text/csv',
+                'facet': [{'name_or_index': 'F1'}],
+                           'group_variable': 'F2',
+                           'headers': ['Label', 'F1', 'F2', 'F3', 'F4'],
+                           'joinsource_name_or_index': 'F4',
+                           'label': 'Label',
+                           'label_values_or_threshold': [1],
+                           'methods': {'post_training_bias': {'methods': 'all'},
+                                       'pre_training_bias': {'methods': 'all'}},
+                           'predictor': {'initial_instance_count': 1,
+                                         'instance_type': 'ml.c5.xlarge',
+                                         'label_headers': ['success'],
+                                         'model_name': 'xgboost-model',
+                                         'probability': 'pr'}}
+    assert actual == expected