add method for data splitting regression

Author: namsaraeva

src/sparcscore/ml/utils.py

@@ -103,3 +103,45 @@ def combine_datasets_balanced(list_of_datasets, class_labels, train_per_class, v
     val_dataset = torch.utils.data.ConcatDataset(val_dataset)
 
     return train_dataset, val_dataset, test_dataset
+
+
+def split_dataset_regression(dataset, train_size, test_size, val_size, seed=None):
+    """
+    Split a dataset into train, test, and validation set.
+
+    Parameters
+    ----------
+    dataset : torch.utils.data.Dataset
+        Dataset to be split.
+    train_size : int
+        Number of samples in the train set.
+    test_size : int
+        Number of samples in the test set.
+    val_size : int
+        Number of samples in the validation set.
+
+    Returns
+    -------
+    train : torch.utils.data.Dataset
+        Train dataset.
+    val : torch.utils.data.Dataset
+        Validation dataset.
+    test : torch.utils.data.Dataset
+        Test dataset.
+    """
+    residual_size = len(dataset) - train_size - test_size - val_size
+
+    if residual_size < 0:
+        raise ValueError(
+            f"Dataset with length {len(dataset)} is too small to be split into test set of size {test_size}, "
+            f"train set of size {train_size}, and validation set of size {val_size}. "
+        )
+    
+    if seed is not None:
+        gen = torch.Generator()
+        gen.manual_seed(seed)
+        train, test, val, _ = torch.utils.data.random_split(dataset, [train_size, test_size, val_size, residual_size], generator=gen)
+    else:
+        train, test, val, _ = torch.utils.data.random_split(dataset, [train_size, test_size, val_size, residual_size])
+
+    return train, val, test