@Test
public void canSetLayersOnAnUninitializedNetwork() {
assertEquals(0, network.getLayers().size());
Set<Layer> layers = new HashSet<Layer>();
layers.add(new Layer());
layers.add(new Layer());
layers.add(new Layer());
network.setLayers(layers);
assertEquals(layers, network.getLayers());
}
@Test
public void initializedNetworkHasLayers() {
List<Layer> layers = new ArrayList<Layer>();
layers.add(new Layer());
layers.add(new Layer());
layers.add(new Layer());
NeuralNetworkImpl network = new NeuralNetworkImpl(layers);
assertEquals(layers, network.getLayers());
}
@Test
public void addLayersAddsLayersToNullLayerList() {
network.setLayers(null);
List<Layer> layers = new ArrayList<Layer>();
layers.add(new Layer());
layers.add(new Layer());
network.addLayers(layers);
assertEquals(2, network.getLayers().size());
}
@Test
public void addLayersWithANullListDoesntRemoveOldLayers() {
List<Layer> layers = new ArrayList<Layer>();
layers.add(new Layer());
layers.add(new Layer());
layers.add(new Layer());
NeuralNetworkImpl network = new NeuralNetworkImpl(layers);
network.addLayers(null);
assertEquals(3, network.getLayers().size());
}
@Test
public void addLayersAddsUniqueLayers() {
List<Layer> layers = new ArrayList<Layer>();
layers.add(new Layer());
layers.add(new Layer());
layers.add(new Layer());
NeuralNetworkImpl network = new NeuralNetworkImpl(layers);
layers = new ArrayList<Layer>();
layers.add(new Layer());
layers.add(new Layer());
network.addLayers(layers);
assertEquals(5, network.getLayers().size());
}
@Test
public void addLayerDoesntAddExistingLayer() {
List<Layer> layers = new ArrayList<Layer>();
Layer oldLayer = new Layer();
layers.add(oldLayer);
layers.add(new Layer());
layers.add(new Layer());
NeuralNetworkImpl network = new NeuralNetworkImpl(layers);
layers = new ArrayList<Layer>();
layers.add(oldLayer);
layers.add(new Layer());
network.addLayers(layers);
assertEquals(4, network.getLayers().size());
}
@Test
public void uninitializedNetworkHasEmptyLayers() {
assertEquals(0, network.getLayers().size());
}