validmind 2.0.1__py3-none-any.whl → 2.0.7__py3-none-any.whl

validmind/unit_metrics/regression/KolmogorovSmirnovStatistic.py

@@ -0,0 +1,36 @@
+# Copyright © 2023-2024 ValidMind Inc. All rights reserved.
+# See the LICENSE file in the root of this repository for details.
+# SPDX-License-Identifier: AGPL-3.0 AND ValidMind Commercial
+
+from dataclasses import dataclass
+
+import numpy as np
+
+from validmind.vm_models import UnitMetric
+
+
+@dataclass
+class KolmogorovSmirnovStatistic(UnitMetric):
+    required_inputs = ["dataset", "model"]
+
+    def run(self):
+        y_true = self.inputs.dataset.y.flatten()
+        y_pred = self.inputs.dataset.y_pred(model_id=self.inputs.model.input_id)
+
+        # Sort true values and corresponding predicted values
+        idx_true = np.argsort(y_true)
+        idx_pred = np.argsort(y_pred)
+        y_true_sorted = y_true[idx_true]
+        y_pred_sorted = y_pred[idx_pred]
+
+        # Compute cumulative distribution functions (CDFs)
+        cdf_true = np.arange(1, len(y_true_sorted) + 1) / len(y_true_sorted)
+        cdf_pred = np.arange(1, len(y_pred_sorted) + 1) / len(y_pred_sorted)
+
+        # Compute absolute differences between CDFs
+        diff_cdf = np.abs(cdf_true - cdf_pred)
+
+        # Find maximum absolute difference
+        ks_statistic = np.max(diff_cdf)
+
+        return self.cache_results(metric_value=ks_statistic)

validmind/unit_metrics/regression/MeanAbsolutePercentageError.py

@@ -0,0 +1,22 @@
+# Copyright © 2023-2024 ValidMind Inc. All rights reserved.
+# See the LICENSE file in the root of this repository for details.
+# SPDX-License-Identifier: AGPL-3.0 AND ValidMind Commercial
+
+from dataclasses import dataclass
+
+import numpy as np
+
+from validmind.vm_models import UnitMetric
+
+
+@dataclass
+class MeanAbsolutePercentageError(UnitMetric):
+    required_inputs = ["dataset", "model"]
+
+    def run(self):
+        y_true = self.inputs.dataset.y
+        y_pred = self.inputs.dataset.y_pred(model_id=self.inputs.model.input_id)
+
+        value = np.mean(np.abs((y_true - y_pred) / y_true)) * 100
+
+        return self.cache_results(metric_value=value)

validmind/unit_metrics/regression/MeanBiasDeviation.py

@@ -0,0 +1,22 @@
+# Copyright © 2023-2024 ValidMind Inc. All rights reserved.
+# See the LICENSE file in the root of this repository for details.
+# SPDX-License-Identifier: AGPL-3.0 AND ValidMind Commercial
+
+from dataclasses import dataclass
+
+import numpy as np
+
+from validmind.vm_models import UnitMetric
+
+
+@dataclass
+class MeanBiasDeviation(UnitMetric):
+    required_inputs = ["dataset", "model"]
+
+    def run(self):
+        y_true = self.inputs.dataset.y
+        y_pred = self.inputs.dataset.y_pred(model_id=self.inputs.model.input_id)
+
+        value = np.mean(y_pred - y_true)
+
+        return self.cache_results(metric_value=value)

validmind/unit_metrics/regression/QuantileLoss.py

@@ -0,0 +1,25 @@
+# Copyright © 2023-2024 ValidMind Inc. All rights reserved.
+# See the LICENSE file in the root of this repository for details.
+# SPDX-License-Identifier: AGPL-3.0 AND ValidMind Commercial
+
+from dataclasses import dataclass
+
+import numpy as np
+
+from validmind.vm_models import UnitMetric
+
+
+@dataclass
+class QuantileLoss(UnitMetric):
+    required_inputs = ["dataset", "model"]
+
+    def run(self):
+        y_true = self.inputs.dataset.y
+        y_pred = self.inputs.dataset.y_pred(model_id=self.inputs.model.input_id)
+
+        error = y_true - y_pred
+        # Quantile value (between 0 and 1).
+        quantile = 0.5
+        value = np.mean(np.maximum(quantile * error, (quantile - 1) * error))
+
+        return self.cache_results(metric_value=value)

validmind/unit_metrics/regression/sklearn/AdjustedRSquaredScore.py

@@ -0,0 +1,27 @@
+# Copyright © 2023-2024 ValidMind Inc. All rights reserved.
+# See the LICENSE file in the root of this repository for details.
+# SPDX-License-Identifier: AGPL-3.0 AND ValidMind Commercial
+
+from dataclasses import dataclass
+
+import sklearn.metrics as metrics
+
+from validmind.vm_models import UnitMetric
+
+
+@dataclass
+class AdjustedRSquaredScore(UnitMetric):
+    required_inputs = ["dataset", "model"]
+
+    def run(self):
+        y_true = self.inputs.dataset.y
+        y_pred = self.inputs.dataset.y_pred(model_id=self.inputs.model.input_id)
+
+        X_columns = self.inputs.dataset.get_features_columns()
+        row_count = len(y_true)
+        feature_count = len(X_columns)
+        value = 1 - (1 - metrics.r2_score(y_true, y_pred)) * (row_count - 1) / (
+            row_count - feature_count
+        )
+
+        return self.cache_results(metric_value=value)

validmind/unit_metrics/regression/sklearn/MeanAbsoluteError.py

@@ -0,0 +1,22 @@
+# Copyright © 2023-2024 ValidMind Inc. All rights reserved.
+# See the LICENSE file in the root of this repository for details.
+# SPDX-License-Identifier: AGPL-3.0 AND ValidMind Commercial
+
+from dataclasses import dataclass
+
+from sklearn.metrics import mean_absolute_error
+
+from validmind.vm_models import UnitMetric
+
+
+@dataclass
+class MeanAbsoluteError(UnitMetric):
+    required_inputs = ["dataset", "model"]
+
+    def run(self):
+        y_true = self.inputs.dataset.y
+        y_pred = self.inputs.dataset.y_pred(model_id=self.inputs.model.input_id)
+
+        value = mean_absolute_error(y_true, y_pred, **self.params)
+
+        return self.cache_results(metric_value=value)

validmind/unit_metrics/regression/sklearn/MeanSquaredError.py

@@ -0,0 +1,22 @@
+# Copyright © 2023-2024 ValidMind Inc. All rights reserved.
+# See the LICENSE file in the root of this repository for details.
+# SPDX-License-Identifier: AGPL-3.0 AND ValidMind Commercial
+
+from dataclasses import dataclass
+
+from sklearn.metrics import mean_squared_error
+
+from validmind.vm_models import UnitMetric
+
+
+@dataclass
+class MeanSquaredError(UnitMetric):
+    required_inputs = ["dataset", "model"]
+
+    def run(self):
+        y_true = self.inputs.dataset.y
+        y_pred = self.inputs.dataset.y_pred(model_id=self.inputs.model.input_id)
+
+        value = mean_squared_error(y_true, y_pred, **self.params)
+
+        return self.cache_results(metric_value=value)

validmind/unit_metrics/regression/sklearn/RSquaredScore.py

@@ -0,0 +1,22 @@
+# Copyright © 2023-2024 ValidMind Inc. All rights reserved.
+# See the LICENSE file in the root of this repository for details.
+# SPDX-License-Identifier: AGPL-3.0 AND ValidMind Commercial
+
+from dataclasses import dataclass
+
+import sklearn.metrics as metrics
+
+from validmind.vm_models import UnitMetric
+
+
+@dataclass
+class RSquaredScore(UnitMetric):
+    required_inputs = ["dataset", "model"]
+
+    def run(self):
+        y_true = self.inputs.dataset.y
+        y_pred = self.inputs.dataset.y_pred(model_id=self.inputs.model.input_id)
+
+        value = metrics.r2_score(y_true, y_pred)
+
+        return self.cache_results(metric_value=value)

validmind/unit_metrics/regression/sklearn/RootMeanSquaredError.py

@@ -0,0 +1,23 @@
+# Copyright © 2023-2024 ValidMind Inc. All rights reserved.
+# See the LICENSE file in the root of this repository for details.
+# SPDX-License-Identifier: AGPL-3.0 AND ValidMind Commercial
+
+from dataclasses import dataclass
+
+import numpy as np
+from sklearn.metrics import mean_squared_error
+
+from validmind.vm_models import UnitMetric
+
+
+@dataclass
+class RootMeanSquaredError(UnitMetric):
+    required_inputs = ["dataset", "model"]
+
+    def run(self):
+        y_true = self.inputs.dataset.y
+        y_pred = self.inputs.dataset.y_pred(model_id=self.inputs.model.input_id)
+
+        value = np.sqrt(mean_squared_error(y_true, y_pred, **self.params))
+
+        return self.cache_results(metric_value=value)

validmind/unit_metrics/sklearn/classification/Accuracy.py

@@ -11,6 +11,8 @@ from validmind.vm_models import UnitMetric
 
 @dataclass
 class Accuracy(UnitMetric):
+    required_inputs = ["dataset", "model"]
+
     def run(self):
         y_true = self.inputs.dataset.y
         y_pred = self.inputs.dataset.y_pred(model_id=self.inputs.model.input_id)

validmind/unit_metrics/sklearn/classification/F1.py

@@ -11,6 +11,8 @@ from validmind.vm_models import UnitMetric
 
 @dataclass
 class F1(UnitMetric):
+    required_inputs = ["dataset", "model"]
+
     def run(self):
         y_true = self.inputs.dataset.y
         y_pred = self.inputs.dataset.y_pred(model_id=self.inputs.model.input_id)

validmind/unit_metrics/sklearn/classification/Precision.py

@@ -11,6 +11,8 @@ from validmind.vm_models import UnitMetric
 
 @dataclass
 class Precision(UnitMetric):
+    required_inputs = ["dataset", "model"]
+
     def run(self):
         y_true = self.inputs.dataset.y
         y_pred = self.inputs.dataset.y_pred(model_id=self.inputs.model.input_id)

validmind/unit_metrics/sklearn/classification/ROC_AUC.py

@@ -11,6 +11,8 @@ from validmind.vm_models import UnitMetric
 
 @dataclass
 class ROC_AUC(UnitMetric):
+    required_inputs = ["dataset", "model"]
+
     def run(self):
         y_true = self.inputs.dataset.y
         y_pred = self.inputs.dataset.y_pred(model_id=self.inputs.model.input_id)

validmind/unit_metrics/sklearn/classification/Recall.py

@@ -11,6 +11,8 @@ from validmind.vm_models import UnitMetric
 
 @dataclass
 class Recall(UnitMetric):
+    required_inputs = ["dataset", "model"]
+
     def run(self):
         y_true = self.inputs.dataset.y
         y_pred = self.inputs.dataset.y_pred(model_id=self.inputs.model.input_id)

validmind/utils.py

@@ -85,6 +85,8 @@ def nan_to_none(obj):
 
 class NumpyEncoder(json.JSONEncoder):
     def default(self, obj):
+        if isinstance(obj, pd.Interval):
+            return f"[{obj.left}, {obj.right}]"
         if isinstance(obj, np.integer):
             return int(obj)
         if isinstance(obj, np.floating):
@@ -253,7 +255,21 @@ def clean_docstring(docstring: str) -> str:
     # Join paragraphs with double newlines for markdown
     description = "\n\n".join(paragraphs)
 
-    return description
+    lines = description.split("\n")
+    in_bullet_list = False
+    for i, line in enumerate([line for line in lines]):
+        if line.strip().startswith("-") and not in_bullet_list:
+            if lines[i - 1] != "":
+                lines[i] = "\n" + line
+
+            in_bullet_list = True
+            continue
+        elif line.strip().startswith("-") and in_bullet_list:
+            continue
+        elif line.strip() == "" and in_bullet_list:
+            in_bullet_list = False
+
+    return "\n".join(lines)
 
 
 def format_number(number):