Refactor: Clean up unused variables and improve code style

This commit focuses on improving code quality by removing unused variables and adhering to more Pythonic style conventions.

Key changes include:

Removed numerous unused local variables across various modules, including model classes, pipeline steps, and utility functions.
Replaced lambda functions with standard def statements for better readability and clarity in keras_classifier_class.py and embeddings.py.
Updated boolean comparisons from == False to the more idiomatic not ... in main.py and notebooks.
Simplified boolean DataFrame indexing in plot_feature_categories.py and notebooks.
Removed unnecessary assignments where the return value of a function was not used.
These changes do not affect the functionality but enhance code maintainability and reduce clutter.

Author: SamoraHunter

ml_grid/model_classes/keras_classifier_class.py

@@ -120,20 +120,22 @@ def __init__(
             l1_reg=0.0,  # Register l1_reg with a default value
             l2_reg=0.0,  # Register l2_reg with a default value
         )
-        X_data = self.X
-        y_data = self.y
 
         # vals = np.linspace(2, 750, 6)
         vals = np.logspace(1, 2.0, 3)
 
-        floorer = lambda t: math.floor(t)
+        def floorer(t):
+            return math.floor(t)
+
         floored_width = np.array([floorer(xi) for xi in vals])
         floored_width = np.insert(floored_width, 0, 1, axis=None)
         floored_width
 
         vals = np.logspace(1, 2.0, 3)
 
-        floorer = lambda t: math.floor(t)
+        def floorer(t):
+            return math.floor(t)
+
         floored_depth = np.array([floorer(xi) for xi in vals])
         floored_depth = np.insert(floored_depth, 0, 1, axis=None)
         floored_depth

ml_grid/model_classes_time_series/CNNClassifier_module.py

@@ -29,9 +29,6 @@ def __init__(self, ml_grid_object: pipe):
         Args:
             ml_grid_object (pipe): An instance of the main data pipeline object.
         """
-        time_limit_param = ml_grid_object.global_params.time_limit_param
-
-        n_jobs_model_val = ml_grid_object.global_params.n_jobs_model_val
 
         random_state_val = ml_grid_object.global_params.random_state_val
 

ml_grid/model_classes_time_series/Catch22Classifer_module.py

@@ -31,7 +31,6 @@ def __init__(self, ml_grid_object: pipe):
             ml_grid_object (pipe): An instance of the main data pipeline object.
         """
 
-        verbose_param = ml_grid_object.verbose
         random_state_val = ml_grid_object.global_params.random_state_val
         n_jobs_model_val = ml_grid_object.global_params.n_jobs_model_val
 

ml_grid/model_classes_time_series/ResNetClassifier_module.py

@@ -31,10 +31,6 @@ def __init__(self, ml_grid_object: pipe):
             ml_grid_object (pipe): An instance of the main data pipeline object.
         """
 
-        random_state_val = ml_grid_object.global_params.random_state_val
-
-        n_jobs_model_val = ml_grid_object.global_params.n_jobs_model_val
-
         verbose_param = ml_grid_object.verbose
 
         param_space = ParamSpace(

ml_grid/model_classes_time_series/TemporalDictionaryEnsembleClassifier_module.py

@@ -32,8 +32,6 @@ def __init__(self, ml_grid_object: pipe):
             ml_grid_object (pipe): An instance of the main data pipeline object.
         """
 
-        verbose_param = ml_grid_object.verbose
-
         random_state_val = ml_grid_object.global_params.random_state_val
 
         time_limit_param = global_parameters.time_limit_param

ml_grid/pipeline/embeddings.py

@@ -152,7 +152,10 @@ def create_embedding_pipeline(
     elif method_lower == "select_kbest_mi":
         default_params = {"random_state": 42}
         default_params.update(kwargs)
-        score_func = lambda X, y: mutual_info_classif(X, y, **default_params)
+
+        def score_func(X, y):
+            return mutual_info_classif(X, y, **default_params)
+
         steps.append(("embed", SelectKBest(score_func=score_func, k=n_components)))
 
     else:

ml_grid/pipeline/grid_search_cross_validate.py

@@ -593,8 +593,6 @@ def __init__(
         if not failed and self.global_parameters.verbose >= 3:
             self.logger.debug("Fitting final model")
 
-        metric_list = self.metric_list
-
         if not failed and self.y_train.nunique() < 2:
             raise ValueError(
                 "Only one class present in y_train. ROC AUC score is not defined "
@@ -902,8 +900,6 @@ def __init__(
                     f"Cross-validation for {method_name} completed in {elapsed_time:.2f} seconds."
                 )
 
-        current_algorithm_scores = scores
-
         if self.global_parameters.verbose >= 4:
 
             debug_print_statements_class(scores).debug_print_scores()
@@ -975,7 +971,7 @@ def _adjust_knn_parameters(self, parameter_space: Union[Dict, List[Dict]]):
         Dynamically adjusts the 'n_neighbors' parameter for KNN-based models
         to prevent errors on small datasets during cross-validation.
         """
-        n_splits = self.cv.get_n_splits()
+        self.cv.get_n_splits()
 
         # Correctly calculate the training fold size
         dummy_indices = np.arange(len(self.X_train))

ml_grid/pipeline/main.py

@@ -361,7 +361,6 @@ def execute(self) -> Tuple[List[List[Any]], float]:
 
         self.model_error_list = []
         self.highest_score = 0
-        highest_score = 0  # for optimisation
 
         if self.multiprocess:
 
@@ -373,11 +372,11 @@ def multi_run_wrapper(args: Tuple) -> Any:
                 from multiprocessing import Pool
 
                 pool = Pool(8)
-                results = pool.map(multi_run_wrapper, self.arg_list)
+                pool.map(multi_run_wrapper, self.arg_list)
                 # print(results)
                 pool.close()  # exp
 
-        elif self.multiprocess == False:
+        elif not self.multiprocess:
             for k in range(0, len(self.arg_list)):
                 try:
                     self.logger.info(

ml_grid/results_processing/plot_feature_categories.py

@@ -187,8 +187,8 @@ def plot_category_impact_on_metric(
             if plot_data[category].nunique() < 2:
                 continue
 
-            mean_with = plot_data[plot_data[category] == True][metric].mean()
-            mean_without = plot_data[plot_data[category] == False][metric].mean()
+            mean_with = plot_data[plot_data[category]][metric].mean()
+            mean_without = plot_data[not plot_data[category]][metric].mean()
 
             impact = mean_with - mean_without
 

ml_grid/util/project_score_save.py

@@ -165,16 +165,12 @@ def update_score_log(
 
         X_train = ml_grid_object_iter.X_train
 
-        y_train = ml_grid_object_iter.y_train
-
         X_test = ml_grid_object_iter.X_test
 
         y_test = ml_grid_object_iter.y_test
 
         X_test_orig = ml_grid_object_iter.X_test_orig
 
-        y_test_orig = ml_grid_object_iter.y_test_orig
-
         param_space_index = ml_grid_object.param_space_index
 
         bayessearch = global_params.bayessearch