Demo of sampling data from NT3

SampleDataDemo.py

@@ -0,0 +1,61 @@
+### Modify NT3_baseline_keras2.py or similar
+# add a 'sample' parameter to gParameters to trigger
+
+import random
+
+def load_data(train_path, test_path, gParameters):
+
+    print('Loading data...')
+    df_train = (pd.read_csv(train_path,header=None).values).astype('float32')
+    df_test = (pd.read_csv(test_path,header=None).values).astype('float32')
+    print('done')
+
+    print('df_train shape:', df_train.shape)
+    print('df_test shape:', df_test.shape)
+
+    seqlen = df_train.shape[1]
+
+    df_y_train = df_train[:,0].astype('int')
+    df_y_test = df_test[:,0].astype('int')
+
+    Y_train = np_utils.to_categorical(df_y_train,gParameters['classes'])
+    Y_test = np_utils.to_categorical(df_y_test,gParameters['classes'])
+
+    df_x_train = df_train[:, 1:seqlen].astype(np.float32)
+    df_x_test = df_test[:, 1:seqlen].astype(np.float32)
+
+#        X_train = df_x_train.as_matrix()
+#        X_test = df_x_test.as_matrix()
+
+    X_train = df_x_train
+    X_test = df_x_test
+
+    scaler = MaxAbsScaler()
+    mat = np.concatenate((X_train, X_test), axis=0)
+    mat = scaler.fit_transform(mat)
+
+    X_train = mat[:X_train.shape[0], :]
+    X_test = mat[X_train.shape[0]:, :]
+    X_train = mat[:X_train.shape[0], :]
+    X_test = mat[X_train.shape[0]:, :]
+
+    sample = gParameters.get('sample', False)
+    if sample:
+        select = get_sample(X_train.shape[0], sample)
+        X_train = X_train[select]
+        Y_train = Y_train[select]
+
+    return X_train, Y_train, X_test, Y_test
+
+def get_sample(population, sample):
+    """ Choose a random sample from a population of the given size.
+        If the sample parameter is > 0,
+        draw a random sample of the size or proportion given.
+        If 0 < sample < 1, returns sample * population;
+        otherwise, sample should be an int < population"""
+
+    if 0.0 < sample < 1.0:
+        size = int(np.ceil(sample * population))
+    else:
+        size = int(sample)
+    return random.sample(range(int(population)), size)