ValueError: The first argument to `Layer.call` must always be passed

Question

I am writing a custom layer inheriting tensorflow.keras.layers.Layer class as below.

class convolutional_block(tf.keras