Merge remote-tracking branch 'upstream/master'

Updating to upstream codebase

GANDLF/anonymize/__init__.py

@@ -43,7 +43,7 @@ def run_anonymizer(
                 input_path,
                 output_path,
                 anonymization_actions={},
-                deletePrivateTags=parameters["delete_private_tags"],
+                delete_private_tags=parameters["delete_private_tags"],
             )
     elif parameters["modality"] in ["histo", "path"]:
         # anonymize_slide(

GANDLF/cli/__init__.py

@@ -6,6 +6,7 @@
 from .recover_config import recover_config
 from .post_training_model_optimization import post_training_model_optimization
 from .generate_metrics import generate_metrics_dict
+from .data_split_saver import split_data_and_save_csvs
 
 from datetime import date
 

GANDLF/cli/data_split_saver.py

@@ -0,0 +1,49 @@
+from typing import Union
+import os
+
+import pandas as pd
+from GANDLF.utils import get_dataframe, split_data
+
+
+def split_data_and_save_csvs(
+    input_data: Union[pd.DataFrame, str], output_dir: str, parameters: dict
+) -> None:
+    """
+    Split the data into training, validation, and testing sets and save them as csvs in the output directory
+
+    Args:
+        input_data (Union[pd.Dataframe, str]): The input data to be split and saved.
+        output_dir (str): The output directory to save the split data.
+        parameters (dict): The parameters dictionary.
+    """
+
+    full_data = get_dataframe(input_data)
+
+    dataframe_split = split_data(full_data, parameters)
+
+    for (
+        testing_and_valid_indeces,
+        trainingData,
+        validationData,
+        testingData,
+    ) in dataframe_split:
+        # training and validation dataframes use the same index, since they are based on the validation split
+        training_data_path = os.path.join(
+            output_dir, f"training_{testing_and_valid_indeces[1]}.csv"
+        )
+        validation_data_path = os.path.join(
+            output_dir, f"validation_{testing_and_valid_indeces[1]}.csv"
+        )
+        # testing dataframes use the first index
+        testing_data_path = os.path.join(
+            output_dir, f"testing_{testing_and_valid_indeces[0]}.csv"
+        )
+
+        for data, path in zip(
+            [trainingData, validationData, testingData],
+            [training_data_path, validation_data_path, testing_data_path],
+        ):
+            # check if the data is not None and the path does not exist
+            if not os.path.exists(path):
+                if data is not None:
+                    data.to_csv(path, index=False)

GANDLF/cli/main_run.py

@@ -1,4 +1,3 @@
-import os, pickle
 from typing import Optional
 from pathlib import Path
 
@@ -9,7 +8,6 @@
     populate_header_in_parameters,
     parseTrainingCSV,
     parseTestingCSV,
-    set_determinism,
 )
 
 

GANDLF/compute/training_loop.py

@@ -237,9 +237,9 @@ def training_loop(
     params["validation_data"] = validation_data
     params["testing_data"] = testing_data
     testingDataDefined = True
-    if params["testing_data"] is None:
-        # testing_data = validation_data
-        testingDataDefined = False
+    if not isinstance(testing_data, pd.DataFrame):
+        if params["testing_data"] is None:
+            testingDataDefined = False
 
     # Setup a few variables for tracking
     best_loss = 1e7

GANDLF/config_manager.py

@@ -630,6 +630,12 @@ def _parseConfig(
         "nested_training" in params
     ), "The parameter 'nested_training' needs to be defined"
     # initialize defaults for nested training
+    params["nested_training"]["stratified"] = params["nested_training"].get(
+        "stratified", False
+    )
+    params["nested_training"]["stratified"] = params["nested_training"].get(
+        "proportional", params["nested_training"]["stratified"]
+    )
     params["nested_training"]["testing"] = params["nested_training"].get("testing", -5)
     params["nested_training"]["validation"] = params["nested_training"].get(
         "validation", -5

GANDLF/losses/regression.py

@@ -1,7 +1,7 @@
 from typing import Optional
 import torch
 import torch.nn.functional as F
-from torch.nn import MSELoss, CrossEntropyLoss, L1Loss
+from torch.nn import CrossEntropyLoss
 from GANDLF.utils import one_hot