Fix typos in supervised/ directory

supervised/algorithms/catboost.py

@@ -50,7 +50,7 @@ def catboost_eval_metric(ml_task, eval_metric):
             "rmse": "RMSE",
             "mse": "mse",
             "mae": "MAE",
-            "mape": "MAPE",
+            "map": "MAP",
             "r2": "R2",
             "spearman": "spearman",
             "pearson": "pearson",
@@ -73,7 +73,7 @@ def catboost_objective(ml_task, eval_metric):
             "spearman",
             "pearson",
             "user_defined_metric",
-        ]:  # cant optimize them directly
+        ]:  # can't optimize them directly
             objective = "RMSE"
     return objective
 
@@ -333,8 +333,8 @@ def get_metric_name(self):
             return "mse"
         elif metric == "MAE":
             return "mae"
-        elif metric == "MAPE":
-            return "mape"
+        elif metric == "MAP":
+            return "map"
         elif metric in ["F1", "TotalF1:average=Micro"]:
             return "f1"
         elif metric == "Accuracy":
@@ -404,7 +404,7 @@ class CBClassifier(ClassifierMixin, CatBoostAlgorithm):
 )
 
 regression_params = copy.deepcopy(classification_params)
-regression_params["loss_function"] = ["RMSE", "MAE", "MAPE"]
+regression_params["loss_function"] = ["RMSE", "MAE", "MAP"]
 
 regression_required_preprocessing = [
     "missing_values_inputation",

supervised/algorithms/lightgbm.py

@@ -61,7 +61,7 @@ def lightgbm_eval_metric(ml_task, automl_eval_metric):
             "rmse": "rmse",
             "mse": "l2",
             "mae": "l1",
-            "mape": "mape",
+            "map": "map",
             "r2": "custom",
             "spearman": "custom",
             "pearson": "custom",

supervised/algorithms/xgboost.py

@@ -326,7 +326,7 @@ def get_metric_name(self):
 
 xgb_regression_params = dict(xgb_bin_class_params)
 xgb_regression_params["objective"] = ["reg:squarederror"]
-# xgb_regression_params["eval_metric"] = ["rmse", "mae", "mape"]
+# xgb_regression_params["eval_metric"] = ["rmse", "mae", "map"]
 xgb_regression_params["max_depth"] = [4, 5, 6, 7, 8, 9]
 
 

supervised/automl.py

@@ -89,7 +89,7 @@ def __init__(
 
         Arguments:
             results_path (str): The path with results. If None, then the name of directory will be generated with the template: AutoML_{number},
-                where the number can be from 1 to 1,000 - depends which direcory name will be available.
+                where the number can be from 1 to 1,000 - depends which directory name will be available.
                 If the `results_path` will point to directory with AutoML results (`params.json` must be present),
                 then all models will be loaded.
 
@@ -154,13 +154,13 @@ def __init__(
 
                 - for binary classification: `logloss`, `auc`, `f1`, `average_precision`, `accuracy` - default is logloss (if left "auto")
                 - for mutliclass classification: `logloss`, `f1`, `accuracy` - default is `logloss` (if left "auto")
-                - for regression: `rmse`, `mse`, `mae`, `r2`, `mape`, `spearman`, `pearson` - default is `rmse` (if left "auto")
+                - for regression: `rmse`, `mse`, `mae`, `r2`, `map`, `spearman`, `pearson` - default is `rmse` (if left "auto")
 
             validation_strategy (dict): Dictionary with validation type. Right now train/test split and cross-validation are supported.
 
                 Example:
 
-                    Cross-validation exmaple:
+                    Cross-validation example:
                     {
                         "validation_type": "kfold",
                         "k_folds": 5,
@@ -169,6 +169,16 @@ def __init__(
                         "random_seed": 123
                     }
 
+                    Repeated k-fold cross-validation example:
+                    {
+                        "validation_type": "kfold",
+                        "k_folds": 5,
+                        "n_repetitions": 3,
+                        "shuffle": True,
+                        "stratify": True,
+                        "random_seed": 123
+                    }
+
                     Train/test example:
                     {
                         "validation_type": "split",
@@ -261,16 +271,16 @@ def __init__(
                 - `group_loss_ratio` - default metric.
 
 
-            fairness_threshold (float): The treshold value for fairness metric.
+            fairness_threshold (float): The threshold value for fairness metric.
                 The direction optimization (below or above threshold) of fairness metric is determined automatically.
 
                 Default values:
 
                 - for `demographic_parity_difference` the metric value should be below 0.1,
                 - for `demographic_parity_ratio` the metric value should be above 0.8,
                 - for `equalized_odds_difference` the metric value should be below 0.1,
-                - for `equalized_odds_ratio` the metric value shoule be above 0.8.
-                - for `group_loss_ratio` the metric value shoule be above 0.8.
+                - for `equalized_odds_ratio` the metric value should be above 0.8.
+                - for `group_loss_ratio` the metric value should be above 0.8.
 
                 For `group_loss_difference` the default threshold value can't be set because it depends on the dataset.
                 If `group_loss_difference` metric is used and `fairness_threshold` is not specified manually, then an exception will be raised.

supervised/base_automl.py

@@ -337,7 +337,9 @@ def create_dir(self, model_path):
 
     def _expected_learners_cnt(self):
         try:
-            repeats = self._validation_strategy.get("repeats", 1)
+            repeats = self._validation_strategy.get(
+                "repeats", self._validation_strategy.get("n_repetitions", 1)
+            )
             folds = self._validation_strategy.get("k_folds", 1)
             return repeats * folds
         except Exception as e:
@@ -949,7 +951,7 @@ def _set_adjusted_validation(self):
 
             self.verbose_print(f"Validation strategy: {k_folds}-fold CV {','.join(cv)}")
         else:
-            # cant stack models for train/test split
+            # can't stack models for train/test split
             self._stack_models = False
             self.verbose_print("Disable stacking for split validation")
 
@@ -961,7 +963,7 @@ def _apply_constraints_stack_models(self):
                 self.verbose_print("Disable stacking for split validation")
             self._stack_models = False
             self._boost_on_errors = False
-        if "repeats" in self._validation_strategy:
+        if "repeats" in self._validation_strategy or "n_repetitions" in self._validation_strategy:
             if self._stack_models:
                 self.verbose_print("Disable stacking for repeated validation")
             self._stack_models = False
@@ -972,7 +974,7 @@ def _apply_constraints_stack_models(self):
             self.tuner._stack_models = self._stack_models
             self.tuner._boost_on_errors = self._boost_on_errors
 
-        # update Time Controler
+        # update Time Controller
         if self._time_ctrl is not None:
             self._time_ctrl._is_stacking = self._stack_models
 
@@ -1037,10 +1039,10 @@ def _fit(self, X, y, sample_weight=None, cv=None, sensitive_features=None):
         self._adjust_validation = False
         self._apply_constraints()
         if not self._adjust_validation:
-            # if there is no validation adjustement
+            # if there is no validation adjustment
             # then we can apply stack_models constraints immediately
-            # if there is validation adjustement
-            # then we will apply contraints after the adjustement
+            # if there is validation adjustment
+            # then we will apply constraints after the adjustment
             self._apply_constraints_stack_models()
 
         try:
@@ -1078,7 +1080,7 @@ def _fit(self, X, y, sample_weight=None, cv=None, sensitive_features=None):
 
             # Automatic Exloratory Data Analysis
             # I disabled EDA, because it won't be supported
-            # I recomend use pandas_profiling or Sweetviz
+            # I recommend use pandas_profiling or Sweetviz
             # if self._explain_level == 2:
             #     EDA.compute(X, y, os.path.join(self._results_path, "EDA"))
 
@@ -2062,7 +2064,7 @@ def _validate_eval_metric(self):
             "mse",
             "mae",
             "r2",
-            "mape",
+            "map",
             "spearman",
             "pearson",
         ]:

supervised/ensemble.py

@@ -167,13 +167,13 @@ def get_oof_matrix(self, models):
                 m for m in models if m.is_fast_enough(self._max_single_prediction_time)
             ]:
                 raise NotTrainedException(
-                    "Can't contruct ensemble with prediction time smaller than limit."
+                    "Can't construct ensemble with prediction time smaller than limit."
                 )
 
         # check if we can construct fair ensemble
         if self._fairness_metric is not None:
             if not [m for m in models if m.is_fair()]:
-                raise NotTrainedException("Can't contruct fair ensemble.")
+                raise NotTrainedException("Can't construct fair ensemble.")
 
         oofs = {}
         sensitive_features = None
@@ -336,10 +336,10 @@ def fit(self, oofs, y, sample_weight=None, sensitive_features=None):
         self.sensitive_features = sensitive_features
         start_time = time.time()
         selected_algs_cnt = 0  # number of selected algorithms
-        self.best_algs = []  # selected algoritms indices from each loop
+        self.best_algs = []  # selected algorithms indices from each loop
 
         total_prediction_time = 0
-        best_sum = None  # sum of best algorihtms
+        best_sum = None  # sum of best algorithms
         for j in range(len(oofs)):  # iterate over all solutions
             min_score = self.metric.get_maximum()
             best_model = None
@@ -379,7 +379,7 @@ def fit(self, oofs, y, sample_weight=None, sensitive_features=None):
                 self.best_loss = min_score
                 selected_algs_cnt = j
 
-            self.best_algs.append(best_model)  # save the best algoritm
+            self.best_algs.append(best_model)  # save the best algorithm
             # update best_sum value
             best_sum = (
                 oofs[best_model] if best_sum is None else best_sum + oofs[best_model]
@@ -531,7 +531,7 @@ def save(self, results_path, model_subpath):
 
         LearningCurves.plot_for_ensemble(self._scores, self.metric.name, model_path)
 
-        # call additional metics just to be sure they are computed
+        # call additional metrics just to be sure they are computed
         self._additional_metrics = self.get_additional_metrics()
 
         AdditionalMetrics.save(

supervised/fairness/metrics.py

@@ -292,7 +292,7 @@ def regression(
                 mean_squared_error(t, p, sample_weight=sample_weight)
             ),
             "R2": r2_score,
-            "MAPE": mean_absolute_percentage_error,
+            "MAP": mean_absolute_percentage_error,
             "SPEARMAN": spearman,
             "PEARSON": pearson,
         }

supervised/fairness/report.py

@@ -100,7 +100,7 @@ def save_classification(
                 if v.get("privileged_value") is not None:
                     fout.write(f'Privileged value is {v["privileged_value"]}.\n')
 
-            # add fairness cerificate
+            # add fairness certificate
             FairnessReport.write_certificate_section(
                 fout,
                 FairnessReport.certificate_info(

supervised/model_framework.py

@@ -91,7 +91,7 @@ def get_train_time(self):
     def predictions(
         self,
         learner,
-        preproces,
+        preprocess,
         X_train,
         y_train,
         sample_weight,
@@ -106,17 +106,17 @@ def predictions(
         y_validation_true = y_validation
         y_validation_predicted = learner.predict(X_validation)
 
-        y_train_true = preproces.inverse_scale_target(y_train_true)
-        y_train_predicted = preproces.inverse_scale_target(y_train_predicted)
-        y_validation_true = preproces.inverse_scale_target(y_validation_true)
-        y_validation_predicted = preproces.inverse_scale_target(y_validation_predicted)
+        y_train_true = preprocess.inverse_scale_target(y_train_true)
+        y_train_predicted = preprocess.inverse_scale_target(y_train_predicted)
+        y_validation_true = preprocess.inverse_scale_target(y_validation_true)
+        y_validation_predicted = preprocess.inverse_scale_target(y_validation_predicted)
 
         y_validation_columns = []
         if self._ml_task == MULTICLASS_CLASSIFICATION:
-            # y_train_true = preproces.inverse_categorical_target(y_train_true)
-            # y_validation_true = preproces.inverse_categorical_target(y_validation_true)
+            # y_train_true = preprocess.inverse_categorical_target(y_train_true)
+            # y_validation_true = preprocess.inverse_categorical_target(y_validation_true)
             # get columns, omit the last one (it is label)
-            y_validation_columns = preproces.prepare_target_labels(
+            y_validation_columns = preprocess.prepare_target_labels(
                 y_validation_predicted
             ).columns.tolist()[:-1]
         elif self._ml_task == BINARY_CLASSIFICATION:
@@ -643,7 +643,7 @@ def save(self, results_path, model_subpath):
             trees_in_iteration=self.additional_params.get("trees_in_step"),
         )
 
-        # call additional metics just to be sure they are computed
+        # call additional metrics just to be sure they are computed
         self._additional_metrics = self.get_additional_metrics()
 
         AdditionalMetrics.save(
@@ -721,6 +721,7 @@ def load(results_path, model_subpath, lazy_load=True):
         for learner_desc, learner_subpath in zip(
             json_desc.get("learners"), json_desc.get("saved")
         ):
+            learner_subpath = os.path.normpath(learner_subpath)
             learner_path = os.path.join(results_path, learner_subpath)
             l = AlgorithmFactory.load(learner_desc, learner_path, lazy_load)
             mf.learners += [l]

supervised/preprocessing/goldenfeatures_transformer.py

@@ -140,9 +140,9 @@ def fit(self, X, y):
             )
             return
         if X.shape[1] == 0:
-            self._error = f"Golden Features not created. No continous features. Input data shape: {X.shape}, {y.shape}"
+            self._error = f"Golden Features not created. No continuous features. Input data shape: {X.shape}, {y.shape}"
             self.save()
-            raise AutoMLException("Golden Features not created. No continous features.")
+            raise AutoMLException("Golden Features not created. No continuous features.")
 
         start_time = time.time()
         combinations = itertools.combinations(X.columns, r=2)

supervised/preprocessing/kmeans_transformer.py

@@ -36,8 +36,8 @@ def fit(self, X, y):
             )
             return
         if X.shape[1] == 0:
-            self._error = f"KMeans not created. No continous features. Input data shape: {X.shape}, {y.shape}"
-            raise AutoMLException("KMeans Features not created. No continous features.")
+            self._error = f"KMeans not created. No continuous features. Input data shape: {X.shape}, {y.shape}"
+            raise AutoMLException("KMeans Features not created. No continuous features.")
 
         start_time = time.time()
 

supervised/preprocessing/preprocessing.py

@@ -281,14 +281,14 @@ def fit_and_transform(self, X_train, y_train, sample_weight=None):
             drop_cols = [c for c in self._drop_features if c in available_cols]
             if len(drop_cols) == X_train.shape[1]:
                 raise AutoMLException(
-                    "All features are droppped! Your data looks like random data."
+                    "All features are dropped! Your data looks like random data."
                 )
             if drop_cols:
                 X_train.drop(drop_cols, axis=1, inplace=True)
             self._drop_features = drop_cols
 
         if X_train is not None:
-            # there can be catagorical columns (in CatBoost) which cant be clipped
+            # there can be categorical columns (in CatBoost) which can't be clipped
             numeric_cols = X_train.select_dtypes(include="number").columns.tolist()
             X_train[numeric_cols] = X_train[numeric_cols].clip(
                 lower=np.finfo(np.float32).min + 1000,
@@ -374,7 +374,7 @@ def transform(self, X_validation, y_validation, sample_weight_validation=None):
             # we should notice user about it!
             # warnings should go to the separate file ...
             # warnings.warn(
-            #    "There are columns {} with missing values which didnt have missing values in train dataset.".format(
+            #    "There are columns {} with missing values which didn't have missing values in train dataset.".format(
             #        list(
             #            X_validation.columns[np.where(np.sum(pd.isnull(X_validation)))]
             #        )
@@ -415,7 +415,7 @@ def transform(self, X_validation, y_validation, sample_weight_validation=None):
             X_validation.drop(self._drop_features, axis=1, inplace=True)
 
         if X_validation is not None:
-            # there can be catagorical columns (in CatBoost) which cant be clipped
+            # there can be categorical columns (in CatBoost) which can't be clipped
             numeric_cols = X_validation.select_dtypes(include="number").columns.tolist()
             X_validation[numeric_cols] = X_validation[numeric_cols].clip(
                 lower=np.finfo(np.float32).min + 1000,

supervised/preprocessing/preprocessing_missing.py

@@ -59,7 +59,7 @@ def transform(self, X):
         X = self._transform_na_fill(X)
         # this is additional run through columns,
         # in case of transforming data with new columns with missing values
-        # X = self._make_sure_na_filled(X) # disbaled for now
+        # X = self._make_sure_na_filled(X) # disabled for now
         return X
 
     def _transform_na_fill(self, X):

supervised/preprocessing/preprocessing_utils.py

@@ -10,7 +10,7 @@ class PreprocessingUtilsException(Exception):
 
 class PreprocessingUtils(object):
     CATEGORICAL = "categorical"
-    CONTINOUS = "continous"
+    CONTINUOUS = "continuous"
     DISCRETE = "discrete"
     DATETIME = "datetime"
     TEXT = "text"
@@ -26,7 +26,7 @@ def get_type(x):
 
         data_type = PreprocessingUtils.CATEGORICAL
         if col_type.startswith("float"):
-            data_type = PreprocessingUtils.CONTINOUS
+            data_type = PreprocessingUtils.CONTINUOUS
         elif col_type.startswith("int") or col_type.startswith("uint"):
             data_type = PreprocessingUtils.DISCRETE
         elif col_type.startswith("datetime"):