Add CIFAR_10 dataset loading and available for benchmarking (#121)

* Add CIFAR_10 dataset loading and available for benchmarking

* Remove line according to PEP8

Author: inteldimitrius

configs/sklearn/performance/tsne.json

@@ -23,7 +23,16 @@
                         "x": "data/mnist_x_test.npy",
                         "y": "data/mnist_y_test.npy"
                     }
-                }
+                },
+		{
+		    "source": "npy",
+		    "name": "cifar_10",
+		    "training":
+		    {
+			"x": "data/cifar_10_x_train.npy",
+			"y": "data/cifar_10_y_train.npy"
+		    }
+		}
             ],
             "workload-size": "medium"
         }

datasets/load_datasets.py

@@ -27,7 +27,7 @@
                                     fraud, gisette, hepmass_150K,
                                     higgs, higgs_one_m, higgs_150K, ijcnn, klaverjas,
                                     santander, skin_segmentation, susy)
-from .loader_multiclass import (connect, covertype, covtype, letters, mlsr,
+from .loader_multiclass import (cifar_10, connect, covertype, covtype, letters, mlsr,
                                 mnist, msrank, plasticc, sensit)
 from .loader_regression import (abalone, california_housing, fried, higgs_10500K,
                                 medical_charges_nominal, mortgage_first_q,
@@ -47,6 +47,7 @@
     "census": census,
     "cifar_binary": cifar_binary,
     "cifar_cluster": cifar_cluster,
+    "cifar_10": cifar_10,
     "codrnanorm": codrnanorm,
     "connect": connect,
     "covertype": covertype,

datasets/loader_multiclass.py

@@ -28,6 +28,42 @@
 from .loader_utils import count_lines, read_libsvm_msrank, retrieve
 
 
+def cifar_10(dataset_dir: Path) -> bool:
+    """
+    Source:
+    University of Toronto
+    Collected by Alex Krizhevsky, Vinod Nair, and Geoffrey Hinton
+    https://www.cs.toronto.edu/~kriz/cifar.html
+
+    Classification task. n_classes = 10
+    cifar_10 x train dataset  (54000, 3072)
+    cifar_10 y train dataset  (54000, 1)
+    cifar_10 x test  dataset  (6000, 3072)
+    cifar_10 y test  dataset  (6000, 1)
+
+    """
+    dataset_name = 'cifar_10'
+    os.makedirs(dataset_dir, exist_ok=True)
+
+    X, y = fetch_openml(data_id=40927, return_X_y=True,
+                        as_frame=False, data_home=dataset_dir)
+
+    X = pd.DataFrame(X)
+    y = pd.DataFrame(y)
+    y = y.astype(int)
+
+    logging.info(f'{dataset_name} is loaded, started parsing...')
+
+    x_train, x_test, y_train, y_test = train_test_split(
+        X, y, test_size=0.1, random_state=42)
+    for data, name in zip((x_train, x_test, y_train, y_test),
+                          ('x_train', 'x_test', 'y_train', 'y_test')):
+        filename = f'{dataset_name}_{name}.npy'
+        np.save(os.path.join(dataset_dir, filename), data)
+    logging.info(f'dataset {dataset_name} is ready.')
+    return True
+
+
 def connect(dataset_dir: Path) -> bool:
     """
     Source: