from sklearn import neighbors
from sklearn.decomposition import PCA, IncrementalPCA, KernelPCA
import make_dataset
if __name__ == '__main__':
test_file_names = os.listdir('/media/dick/Storage/test')
N_test = len(test_file_names)
print(str(N_test) + ' test files.')
train_file_names = os.listdir('/media/dick/Storage64GB_2/train')
N_train = len(train_file_names)
print(str(N_train) + ' train files.')
array_train, array_list_train = make_dataset.read_retina_images(which_set='train')
array_test, array_list_test = make_dataset.read_retina_images(which_set='test')
# train_labels = np.load('/media/dick/Storage64GB_2/train_labels.npy')
# print(train_labels.shape)
train_imgs, train_labels = make_dataset.read_retina_images(which_set='train')
test_imgs, test_labels = make_dataset.read_retina_images(which_set='test')
N_train, m_train, n_train, chan_train = train_imgs.shape
N_test, m_test, n_test, chan_test = test_imgs.shape
# cv2.imshow('image', train_imgs[0])
# cv2.waitKey(0)
# cv2.destroyAllWindows()
print(train_imgs.reshape((N_train, m_train*n_train))[:5000].shape)
from wabbit_wappa import *
import make_dataset
def make_features(data):
length = 64
features = []
for i in range(length):
features.append(str(data[i]))
return features
if __name__ == '__main__':
test_file_names = os.listdir('/media/dick/Storage1TB/test')
train_imgs, train_labels = make_dataset.read_retina_images(which_set='train')
test_imgs, test_labels = make_dataset.read_retina_images(which_set='test')
N_train, m_train, n_train, chan_train = train_imgs.shape
N_test, m_test, n_test, chan_test = test_imgs.shape
print('# of training examples: ' + str(N_train) + '\n# of testing examples: ' + str(N_test))
# of training examples: 1580670
# of testing examples: 2410920
y_train = train_labels
print(len(y_train))
# Create/Open the memory mapped variables
X_train = np.memmap('/media/dick/Storage1TB/transformed/train_ipca.mmap',
mode='r', shape=(N_train, 64), dtype='float')
num_train = int(round(0.8*N_train))