Preprocessing docstring and cli_test update

zookeeper/cli_test.py

@@ -24,7 +24,8 @@ def outputs(self, data):
 
 @registry.register_preprocess("mnist")
 class default(ImageClassification):
-    pass
+    def inputs(self, data):
+        return super().inputs(data) / 255
 
 
 @registry.register_preprocess("mnist")

zookeeper/preprocessing.py

@@ -2,6 +2,17 @@
 
 
 class Preprocessing(abc.ABC):
+    """ An abstract Preprocessing class that contains the abstract contains the
+    methods required for Preprocessing.
+
+    We define a neural network to be a mapping between an input and an output.
+    hence we define to abstract methods, an input (e.g the image for image
+    classification), and an output (e.g the class label for image classification).
+    We also define decoders, which allows use to customize the decoding, and kwargs,
+    which allows one to pass information from the preprocessing to the model (e.g
+    input size.)
+    """
+
     decoders = None
     kwargs = {}