How to use the hub.load function in hub

def download():
    vol = hub.load(name='imagenet/image:val')[400:600]
    a = (vol.mean(axis=(1,2,3)) == 0).sum()
    print(vol.mean(axis=(1,2,3)) == 0)

def test_public_access_no_creds():
    x = hub.load('imagenet')
    assert x[0].mean() == 1

def upload_tfrecord(dataset_type, filepath, version, start_frame):
    storage = S3(bucket='waymo-dataset-upload')
    str_lasers_range_image = 'edward/{}-lasers-range-image:{}'.format(dataset_type, version)
    str_lasers_range_image_first = 'edward/{}-lasers-range-image-first:{}'.format(dataset_type, version)
    str_lasers_camera_proj = 'edward/{}-lasers-camera-proj:{}'.format(dataset_type, version)
    str_lasers_camera_proj_first = 'edward/{}-lasers-camera-proj-first:{}'.format(dataset_type, version)
    hub_lasers_range_image = hub.load(name=str_lasers_range_image, storage=storage)
    hub_lasers_camera_proj = hub.load(name=str_lasers_camera_proj, storage = storage)
    hub_lasers_range_image_first = hub.load(name=str_lasers_range_image_first, storage=storage)
    hub_lasers_camera_proj_first = hub.load(name=str_lasers_camera_proj_first, storage=storage)

    dataset = tf.data.TFRecordDataset(filepath)

    for batch in dataset.batch(1):
        def get_arr_image(range_image_compressed):
           data = zlib.decompress(range_image_compressed)
           mt = open_dataset.MatrixFloat()
           mt.ParseFromString(data)
           arr = np.reshape(np.array(mt.data), tuple(mt.shape.dims), order='C')
           return arr

        def get_arr_proj(camera_projection_compressed):
            data = zlib.decompress(camera_projection_compressed)
            mt = open_dataset.MatrixInt32()
            mt.ParseFromString(data)
            arr = np.reshape(np.array(mt.data), tuple(mt.shape.dims), order='C')

def upload_tfrecord(dataset_type, filepath, version, start_frame):
    storage = S3(bucket='waymo-dataset-upload')
    str_lasers_range_image = 'edward/{}-lasers-range-image:{}'.format(dataset_type, version)
    str_lasers_range_image_first = 'edward/{}-lasers-range-image-first:{}'.format(dataset_type, version)
    str_lasers_camera_proj = 'edward/{}-lasers-camera-proj:{}'.format(dataset_type, version)
    str_lasers_camera_proj_first = 'edward/{}-lasers-camera-proj-first:{}'.format(dataset_type, version)
    hub_lasers_range_image = hub.load(name=str_lasers_range_image, storage=storage)
    hub_lasers_camera_proj = hub.load(name=str_lasers_camera_proj, storage = storage)
    hub_lasers_range_image_first = hub.load(name=str_lasers_range_image_first, storage=storage)
    hub_lasers_camera_proj_first = hub.load(name=str_lasers_camera_proj_first, storage=storage)

    dataset = tf.data.TFRecordDataset(filepath)

    for batch in dataset.batch(1):
        def get_arr_image(range_image_compressed):
           data = zlib.decompress(range_image_compressed)
           mt = open_dataset.MatrixFloat()
           mt.ParseFromString(data)
           arr = np.reshape(np.array(mt.data), tuple(mt.shape.dims), order='C')
           return arr

        def get_arr_proj(camera_projection_compressed):
            data = zlib.decompress(camera_projection_compressed)
            mt = open_dataset.MatrixInt32()

def upload_tfrecord(dataset_type, filepath, version, start_frame):
    storage = S3(bucket='waymo-dataset-upload')
    label_name = 'edward/{}-labels:{}'.format(dataset_type, version)
    image_name = 'edward/{}-camera-images:{}'.format(dataset_type, version)
    # print('{} {}'.format(label_name, image_name))
    images_arr = hub.load(name=image_name, storage=storage)
    labels_arr = hub.load(name=label_name, storage=storage)
    frame_count = start_frame
    dataset = tf.data.TFRecordDataset(filepath)
    # print('Yeah {}'.format(frame_count))
    for batch in dataset.batch(16):
        # print('Cycle')
        t1 = clock()
        l = batch.shape[0]
        arr = np.zeros(shape=(l, 6, 1280, 1920, 3), dtype='uint8')
        lab = np.zeros(shape=(l, 2, 6, 30, 7), dtype='float64')
        for i in range(0, l):
            # print('Cycle2')
            data = batch[i]
            frame = open_dataset.Frame()
            frame.ParseFromString(bytearray(data.numpy()))
            for image in frame.images:

def upload_tfrecord(dataset_type, filepath, version, start_frame):
    storage = S3(bucket='waymo-dataset-upload')
    label_name = 'edward/{}-labels:{}'.format(dataset_type, version)
    image_name = 'edward/{}-camera-images:{}'.format(dataset_type, version)
    # print('{} {}'.format(label_name, image_name))
    images_arr = hub.load(name=image_name, storage=storage)
    labels_arr = hub.load(name=label_name, storage=storage)
    frame_count = start_frame
    dataset = tf.data.TFRecordDataset(filepath)
    # print('Yeah {}'.format(frame_count))
    for batch in dataset.batch(16):
        # print('Cycle')
        t1 = clock()
        l = batch.shape[0]
        arr = np.zeros(shape=(l, 6, 1280, 1920, 3), dtype='uint8')
        lab = np.zeros(shape=(l, 2, 6, 30, 7), dtype='float64')
        for i in range(0, l):
            # print('Cycle2')
            data = batch[i]
            frame = open_dataset.Frame()
            frame.ParseFromString(bytearray(data.numpy()))
            for image in frame.images:
                # print('Cycle3')

def upload_tfrecord(dataset_type, filepath, version, start_frame):
    storage = S3(bucket='waymo-dataset-upload')
    str_lasers_range_image = 'edward/{}-lasers-range-image:{}'.format(dataset_type, version)
    str_lasers_range_image_first = 'edward/{}-lasers-range-image-first:{}'.format(dataset_type, version)
    str_lasers_camera_proj = 'edward/{}-lasers-camera-proj:{}'.format(dataset_type, version)
    str_lasers_camera_proj_first = 'edward/{}-lasers-camera-proj-first:{}'.format(dataset_type, version)
    hub_lasers_range_image = hub.load(name=str_lasers_range_image, storage=storage)
    hub_lasers_camera_proj = hub.load(name=str_lasers_camera_proj, storage = storage)
    hub_lasers_range_image_first = hub.load(name=str_lasers_range_image_first, storage=storage)
    hub_lasers_camera_proj_first = hub.load(name=str_lasers_camera_proj_first, storage=storage)

    dataset = tf.data.TFRecordDataset(filepath)

    for batch in dataset.batch(1):
        def get_arr_image(range_image_compressed):
           data = zlib.decompress(range_image_compressed)
           mt = open_dataset.MatrixFloat()
           mt.ParseFromString(data)
           arr = np.reshape(np.array(mt.data), tuple(mt.shape.dims), order='C')
           return arr

        def get_arr_proj(camera_projection_compressed):
            data = zlib.decompress(camera_projection_compressed)
            mt = open_dataset.MatrixInt32()
            mt.ParseFromString(data)