ddi-fw 0.0.181__tar.gz → 0.0.183__tar.gz

{ddi_fw-0.0.181 → ddi_fw-0.0.183}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: ddi_fw
-Version: 0.0.181
+Version: 0.0.183
 Summary: Do not use :)
 Author-email: Kıvanç Bayraktar <bayraktarkivanc@gmail.com>
 Maintainer-email: Kıvanç Bayraktar <bayraktarkivanc@gmail.com>

{ddi_fw-0.0.181 → ddi_fw-0.0.183}/pyproject.toml

@@ -6,7 +6,7 @@ build-backend = "setuptools.build_meta"
 
 [project]
 name = "ddi_fw"
-version = "0.0.181"
+version = "0.0.183"
 description = "Do not use :)"
 readme = "README.md"
 authors = [

{ddi_fw-0.0.181 → ddi_fw-0.0.183}/src/ddi_fw/datasets/core.py

@@ -127,6 +127,8 @@ class BaseDataset(BaseModel):
                 "X_train, y_train, X_test, and y_test are already provided. Skipping calculation.")
             return self.X_train, self.X_test, self.y_train, self.y_test, self.train_indexes, self.test_indexes, self.train_idx_arr, self.val_idx_arr
 
+        self.prep()
+
         if self.index_path is None:
             raise Exception(
                 "There is no index path. Please call split_dataset or provide indices.")
@@ -140,8 +142,6 @@ class BaseDataset(BaseModel):
         except FileNotFoundError as e:
             raise FileNotFoundError(f"Index files not found: {e.filename}")
 
-        self.prep()
-
         train = self.dataframe[self.dataframe.index.isin(train_idx_all)]
         test = self.dataframe[self.dataframe.index.isin(test_idx_all)]
 

{ddi_fw-0.0.181 → ddi_fw-0.0.183}/src/ddi_fw/ml/tensorflow_wrapper.py

@@ -19,6 +19,48 @@ from ddi_fw.ml.evaluation_helper import Metrics, evaluate
 import ddi_fw.utils as utils
 import os
 
+def convert_to_categorical(arr, num_classes):
+    """
+    This function takes an array of labels and converts them to one-hot encoding 
+    if they are not binary-encoded. If the array is already in a 
+    compatible format, it returns the original array.
+
+    Parameters:
+    - arr: numpy array with label data (could be binary-encoded or label-encoded)
+    - num_classes: number of classes to be used in one-hot encoding
+
+    Returns:
+    - The one-hot encoded array if the original array was binary or label encoded
+    - The original array if it doesn't require any conversion
+    """
+    
+    try:
+        # First, check if the array is binary-encoded
+        if not utils.is_binary_encoded(arr):
+            # If the arr labels are binary-encoded, convert them to one-hot encoding
+            return tf.keras.utils.to_categorical(np.argmax(arr, axis=1), num_classes=num_classes)
+        else:
+            print("No conversion needed, returning original array.")
+            return arr
+    except Exception as e:
+        # If binary encoding check raises an error, print it and continue to label encoding check
+        print(f"Error while checking binary encoding: {e}")
+    
+    try:
+        # Check if the array is label-encoded
+        if utils.is_label_encoded(arr):
+            # If the arr labels are label-encoded, convert them to one-hot encoding
+            return tf.keras.utils.to_categorical(arr, num_classes=num_classes)
+    except Exception as e:
+        # If label encoding check raises an error, print it
+        print(f"Error while checking label encoding: {e}")
+        # If the arr labels don't match any of the known encodings, raise an error
+        raise ValueError("Unknown label encoding format.")
+    
+    # If no conversion was needed, return the original array
+    
+    return arr
+
 
 class TFModelWrapper(ModelWrapper):
 
@@ -138,10 +180,10 @@ class TFModelWrapper(ModelWrapper):
                 self.best_model: Model = best_model
                 pred = self.predict()
                 actual = self.test_label
-                if not utils.is_binary_encoded(pred):
-                    pred = tf.keras.utils.to_categorical(np.argmax(pred,axis=1), num_classes=self.num_classes)
-                if not utils.is_binary_encoded(actual):
-                    actual = tf.keras.utils.to_categorical(actual, num_classes=self.num_classes)
+                # if not utils.is_binary_encoded(pred):
+                #     pred = tf.keras.utils.to_categorical(np.argmax(pred,axis=1), num_classes=self.num_classes)
+                pred= convert_to_categorical(pred, self.num_classes)
+                actual= convert_to_categorical(actual, self.num_classes)
 
                 logs, metrics = evaluate(
                     actual=actual, pred=pred, info=self.descriptor, print_detail=print_detail)
@@ -160,11 +202,12 @@ class TFModelWrapper(ModelWrapper):
             self.best_model = best_model
             pred = self.predict()
             actual = self.test_label
-            if not utils.is_binary_encoded(pred):
-                pred = tf.keras.utils.to_categorical(np.argmax(pred,axis=1), num_classes=self.num_classes)
-            if not utils.is_binary_encoded(actual):
-                actual = tf.keras.utils.to_categorical(actual, num_classes=self.num_classes)
-
+            # if not utils.is_binary_encoded(pred):
+            #     pred = tf.keras.utils.to_categorical(np.argmax(pred,axis=1), num_classes=self.num_classes)
+            # if not utils.is_binary_encoded(actual):
+            #     actual = tf.keras.utils.to_categorical(actual, num_classes=self.num_classes)
+            pred= convert_to_categorical(pred, self.num_classes)
+            actual= convert_to_categorical(actual, self.num_classes)
             logs, metrics = evaluate(
                 actual=actual, pred=pred, info=self.descriptor)
             metrics.format_float()

{ddi_fw-0.0.181 → ddi_fw-0.0.183}/src/ddi_fw/utils/__init__.py

@@ -4,4 +4,4 @@ from .py7zr_helper import Py7ZipHelper
 from .enums import UMLSCodeTypes, DrugBankTextDataTypes
 from .package_helper import get_import
 from .kaggle import create_kaggle_dataset
-from .categorical_data_encoding_checker import is_one_hot_encoded, is_binary_encoded, is_binary_vector
+from .categorical_data_encoding_checker import is_one_hot_encoded, is_binary_encoded, is_binary_vector,is_label_encoded

{ddi_fw-0.0.181 → ddi_fw-0.0.183}/src/ddi_fw/utils/categorical_data_encoding_checker.py

@@ -30,3 +30,15 @@ def is_binary_vector(arr):
     if arr.ndim != 1:
         raise ValueError("Input must be a 1D array.")
     return arr.ndim == 1 and np.all(np.isin(arr, [0, 1]))
+
+
+def is_label_encoded(arr):
+    # Check if the array is label encoded
+    # Ensure the input is a numpy ndarray and is 1D
+    if not isinstance(arr, np.ndarray):
+        raise ValueError("Input must be a NumPy ndarray.")
+    if arr.ndim != 1:
+        raise ValueError("Input must be a 1D array.")
+    
+    # Check if all values are non-negative integers (possible class labels)
+    return np.issubdtype(arr.dtype, np.integer) and np.all(arr >= 0)

{ddi_fw-0.0.181 → ddi_fw-0.0.183}/src/ddi_fw.egg-info/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: ddi_fw
-Version: 0.0.181
+Version: 0.0.183
 Summary: Do not use :)
 Author-email: Kıvanç Bayraktar <bayraktarkivanc@gmail.com>
 Maintainer-email: Kıvanç Bayraktar <bayraktarkivanc@gmail.com>