viz

Abschlussprojekt/genome.py

@@ -80,6 +80,11 @@ class Genome:
         for _ in range(outputs):
             genome.add_node(node_type=NodeType.OUTPUT)
 
+        # Fully connect
+        for i in range(inputs):
+            for o in range(inputs, inputs + outputs):
+                genome.add_connection(i, o, weight=1)
+
         return genome
 
 
@@ -112,7 +117,7 @@ def _mutate_add_node(genome: Genome) -> None:
 
     # Find connection to split
     try:
-        connection = choice(list(genome.connections.values()))
+        connection = choice([node for node in genome.connections.values() if not node.disabled])
     except IndexError:
         return
     connection.disabled = True

Abschlussprojekt/requirements.txt

@@ -0,0 +1 @@
+pygraphviz

Abschlussprojekt/visualization.py

@@ -1,3 +1,5 @@
+import itertools
+
 import matplotlib.pyplot as plt
 import networkx as nx
 import numpy as np
@@ -5,8 +7,17 @@ import numpy as np
 from genome import Genome, NodeType, mutate
 
 
+def _find_layer(g: nx.DiGraph, hidden_node: int, inputs: list[int]) -> int:
+    paths = []
+    for input_node in inputs:
+        paths += list(nx.all_simple_paths(g, input_node, hidden_node))
+
+    path_lengths = [len(path) for path in paths]
+    return max(path_lengths)
+
+
 def genome(genome: Genome):
-    graph = nx.Graph()
+    graph = nx.DiGraph()
 
     # Add nodes
     for node in genome.nodes.keys():
@@ -20,33 +31,37 @@ def genome(genome: Genome):
         from_node, to_node = connection.nodes
         graph.add_edge(from_node, to_node, weight=connection.weight)
 
-    # Make sure that input and output nodes are fixed
-    pos = nx.spring_layout(graph)
-    x = [v[0] for v in pos.values()]
-    min_x, max_x = min(x), max(x)
-    y = [v[1] for v in pos.values()]
-    min_y, max_y = min(y), max(y)
+    inputs = [node.id for node in genome.nodes.values() if node.type == NodeType.INPUT]
+    hidden = [node.id for node in genome.nodes.values() if node.type == NodeType.HIDDEN]
+    outputs = [node.id for node in genome.nodes.values() if node.type == NodeType.OUTPUT]
 
-    inputs = [node for node in genome.nodes.values() if node.type == NodeType.INPUT]
-    outputs = [node for node in genome.nodes.values() if node.type == NodeType.OUTPUT]
+    for input_node in inputs:
+        graph.nodes[input_node]["layer"] = 0
 
-    for node, y in zip(inputs, np.linspace(min_y, max_y, len(inputs))):
-        pos[node.id] = np.array([min_x * 1.5, y])
+    max_layer = 1
+    for hidden_node in hidden:
+        layer = _find_layer(graph, hidden_node, inputs)
+        max_layer = max(layer, max_layer)
+        graph.nodes[hidden_node]["layer"] = layer
 
-    for node, y in zip(outputs, np.linspace(min_y, max_y, len(outputs))):
-        pos[node.id] = np.array([max_x * 1.5, y])
+    for output_node in outputs:
+        graph.nodes[output_node]["layer"] = max_layer + 1
 
     plt.subplot()
-    nx.draw_networkx(graph, pos)
+    pos = nx.multipartite_layout(graph, subset_key="layer")
+    nx.draw_networkx_nodes(graph, pos, nodelist=inputs, label=inputs, node_color="#ff0000")
+    nx.draw_networkx_nodes(graph, pos, nodelist=hidden, label=hidden, node_color="#00ff00")
+    nx.draw_networkx_nodes(graph, pos, nodelist=outputs, label=outputs, node_color="#0000ff")
+    nx.draw_networkx_edges(graph, pos)
 
 
 if __name__ == "__main__":
     g1 = Genome.new(3, 2)
-    g1.add_node()
-    g1.add_node()
-    g1.add_node()
     g1.add_connection(0, 4, 0.5)
-
     mutate(g1)
+    mutate(g1)
+    mutate(g1)
+
+    # mutate(g1)
     genome(g1)
     plt.show()