Add files via upload

data_loaders/ames.py

@@ -0,0 +1,27 @@
+import numpy as np
+import pandas as pd
+import os
+import ast
+
+from data_loaders.dataset import DataLoader
+
+class Ames(DataLoader):
+    def __init__(self):
+        # We're using the Mordred fingerprint
+        # This dataset was generated using Mordred.py
+        dataset_dir = os.path.dirname(os.path.abspath(__file__)) + "/../Fingerprints/Ames_Mordred.csv"
+
+        data = pd.read_csv(dataset_dir)
+        fingerprint_arrays = data["fingerprint_Mordred"].apply(lambda array_str: np.array(ast.literal_eval(array_str))).tolist()
+        self.x_values = np.stack(fingerprint_arrays)
+        self.y_values = data["Y"].to_numpy()
+        self.name = "Ames"
+
+    def size(self):
+        return len(self.x_values)
+
+    def x(self, dataset_slice_indices: np.ndarray) -> np.ndarray:
+        return self.x_values[dataset_slice_indices]
+
+    def y(self, dataset_slice_indices):
+        return self.y_values[dataset_slice_indices]

data_loaders/dataset.py

@@ -0,0 +1,21 @@
+import numpy as np
+
+# This class will be used to load the dataset and perform feature extraction
+# When we do bayesian optimization, we'll call the dataloader to get the x and y values for each dataset slice
+class DataLoader:
+    def __init__(self):
+        self.name = ""
+        # You should perform feature extraction and data cleaning here
+        pass
+
+    def size(self) -> int:
+        pass
+
+    # for a given slice of the dataset, we'll return the x values
+    # these slices are indexes of the original dataset. e.g. if you pass in [1, 2, 9], The features at index 1, 2, and 9 will be returned
+    def x(self, dataset_slice: np.ndarray) -> np.ndarray:
+        pass
+
+    # for a given slice of the dataset, we'll return the y values
+    def y(self, dataset_slice: np.ndarray) -> np.ndarray:
+        pass

data_loaders/halflife.py

@@ -0,0 +1,25 @@
+import numpy as np
+import pandas as pd
+from data_loaders.dataset import DataLoader
+from transformer.fingerprints import FingerprintsTransformer    
+import os
+
+class halflife(DataLoader):
+    def __init__(self):
+        # path = r"datasets\cleaned_datasets\halflife_dataset.csv"
+        path = os.path.dirname(os.path.abspath(__file__)) + "/../datasets/cleaned_datasets/halflife_dataset.csv"
+        self.data = pd.read_csv(path)
+        transformer = FingerprintsTransformer(self.data, "Drug", "ECFP")
+
+        self.name = "halflife"
+        self.x_values = transformer.to_np()
+        self.y_values = self.data["Y"].to_numpy()
+
+    def size(self):
+        return len(self.x_values)
+
+    def x(self, dataset_slice_indices: np.ndarray) -> np.ndarray:
+        return self.x_values[dataset_slice_indices]
+
+    def y(self, dataset_slice_indices):
+        return self.y_values[dataset_slice_indices]

data_loaders/ld50.py

@@ -0,0 +1,33 @@
+import os
+import numpy as np
+import pandas as pd
+import ast
+
+from data_loaders.dataset import DataLoader
+from functools import lru_cache
+
+class LD50(DataLoader):
+    fingerprint = 'ECFP'
+
+    def __init__(self):
+        path = os.path.join(
+            os.path.dirname(os.path.abspath(__file__)),
+            "../datasets/fingerprint_datasets/LD50_Zhu_ECFP.h5"
+        )
+        self.data = pd.read_hdf(path)
+        self.name = "ld50"
+        self.x_values = np.stack(self.data['fingerprint_ECFP'])
+        self.y_values = self.data["Y"].to_numpy()
+
+    def size(self):
+        return len(self.x_values)
+
+    def x(self, dataset_slice: slice | np.ndarray = None) -> np.ndarray:
+        if dataset_slice is None:
+            return self.x_values
+        return self.x_values[dataset_slice]
+
+    def y(self, dataset_slice: slice | np.ndarray = None) -> np.ndarray:
+        if dataset_slice is None:
+            return self.y_values
+        return self.y_values[dataset_slice]

data_loaders/tox21.py

@@ -0,0 +1,26 @@
+from tdc.utils import retrieve_label_name_list
+from tdc.single_pred import Tox
+import numpy as np
+from transformer.fingerprints import FingerprintsTransformer
+from data_loaders.dataset import DataLoader
+
+# https://tdcommons.ai/single_pred_tasks/tox/#tox21
+class Tox21(DataLoader):
+    def __init__(self):
+        label_list = retrieve_label_name_list('Tox21')
+        data = Tox(name = 'Tox21', label_name = label_list[0]).get_data()
+
+        transformer = FingerprintsTransformer(data, "Drug", "ECFP")
+
+        self.name = "Tox21"
+        self.x_values = transformer.to_np()
+        self.y_values = data["Y"].to_numpy()
+
+    def size(self):
+        return len(self.x_values)
+
+    def x(self, dataset_slice_indices: np.ndarray) -> np.ndarray:
+        return self.x_values[dataset_slice_indices]
+
+    def y(self, dataset_slice_indices):
+        return self.y_values[dataset_slice_indices]