evonet 0.1.0.dev26__tar.gz → 0.1.0.dev28__tar.gz

{evonet-0.1.0.dev26 → evonet-0.1.0.dev28}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: evonet
-Version: 0.1.0.dev26
+Version: 0.1.0.dev28
 Summary: Evolvable neural network core for integration with EvoLib
 Author-email: EvoLib <evolib@dismail.de>
 License: MIT License

{evonet-0.1.0.dev26 → evonet-0.1.0.dev28}/evonet/connection.py

@@ -7,7 +7,6 @@ Supports optional connection types for specialized behaviors (e.g. inhibitory,
 recurrent).
 """
 
-
 from collections import deque
 from typing import TYPE_CHECKING, Deque, Optional
 
@@ -38,18 +37,33 @@ class Connection:
         conn_type: ConnectionType = ConnectionType.STANDARD,
     ) -> None:
 
+        if delay < 0:
+            raise ValueError("delay must be >= 0")
+
+        if conn_type is not ConnectionType.RECURRENT:
+            delay = 0
+
+        # Normalize: recurrent implies delay >= 1
+        if conn_type is ConnectionType.RECURRENT and delay == 0:
+            delay = 1
+
         self.source = source
         self.target = target
         self.weight = weight
         self.delay = delay
-        self.type: ConnectionType = conn_type
+        self.type = conn_type
 
         self._history: Optional[Deque[float]] = None
-        if self.delay > 0:
+        if self.type is ConnectionType.RECURRENT:
             self._history = deque(maxlen=self.delay)
 
-        if self.delay < 0:
-            raise ValueError("delay must be >= 0")
+    def set_delay(self, delay: int) -> None:
+        if self.type is not ConnectionType.RECURRENT:
+            raise ValueError("set_delay is only valid for recurrent connections")
+        if delay <= 0:
+            delay = 1
+        self.delay = int(delay)
+        self._history = deque(maxlen=self.delay)
 
     def push_source_output(self, value: float) -> None:
         """

{evonet-0.1.0.dev26 → evonet-0.1.0.dev28}/evonet/core.py

@@ -351,6 +351,12 @@ class Nnet:
         if weight is None:
             weight = 0.0
 
+        # Normalize recurrent delays
+        if conn_type is ConnectionType.RECURRENT and delay <= 0:
+            delay = 1
+        if conn_type is not ConnectionType.RECURRENT:
+            delay = 0
+
         conn = Connection(
             source, target, weight=weight, conn_type=conn_type, delay=delay
         )
@@ -397,17 +403,12 @@ class Nnet:
         for i, n in enumerate(input_layer.neurons):
             n.input = float(input_values[i])
 
-        # Preload recurrent contributions from previous time step (last_output)
+        # Preload recurrent contributions from delay buffers.
         for layer in self.layers:
             for n in layer.neurons:
                 for c in n.incoming:
                     if c.type is ConnectionType.RECURRENT:
-                        if c.delay > 0:
-                            # Use delayed history if configured
-                            n.input += c.weight * c.delayed_source_output()
-                        else:
-                            # default: 1-step recurrence via last_output
-                            n.input += c.weight * c.source.last_output
+                        n.input += c.weight * c.delayed_source_output()
 
         # Feed-forward by layers: activate first, then propagate non-recurrent edges
         for layer in self.layers:
@@ -442,7 +443,7 @@ class Nnet:
 
         # Update recurrent delay buffers once per time step.
         for c in self.get_all_connections():
-            if c.type is ConnectionType.RECURRENT and c.delay > 0:
+            if c.type is ConnectionType.RECURRENT:
                 c.push_source_output(c.source.output)
 
         return [n.output for n in self.layers[-1].neurons]
@@ -478,6 +479,7 @@ class Nnet:
         colors_on: bool = True,
         thickness_on: bool = False,
         fillcolors_on: bool = False,
+        delay_on: bool = True,
     ) -> None:
         """
         Render a visual representation of the network using Graphviz.
@@ -489,6 +491,7 @@ class Nnet:
             colors_on (bool): Whether to color edges by sign.
             thickness_on (bool): Whether to scale edge thickness by weight.
             fillcolors_on (bool): Whether to color neurons by role.
+            delay_on (bool): Whether to show delay values on recurrent connections.
         """
 
         if not self.layers:
@@ -541,14 +544,24 @@ class Nnet:
         # Add edges
         for conn in self.get_all_connections():
             label = f"{conn.weight:.2f}" if labels_on else ""
+
+            # Optional delay annotation (only meaningful for recurrent connections)
+            if delay_on and labels_on and conn.type.name == "RECURRENT":
+                if label:
+                    label = f"{label}\\nd={conn.delay}"
+                else:
+                    label = f"d={conn.delay}"
+
             color = (
                 "green"
                 if colors_on and conn.weight >= 0
                 else "red" if colors_on else "black"
             )
+
             penwidth = (
                 str(max(1, min(5, abs(conn.weight * 5)))) if thickness_on else "1"
             )
+
             style = "dashed" if conn.type.name == "RECURRENT" else "solid"
 
             dot.edge(

{evonet-0.1.0.dev26 → evonet-0.1.0.dev28}/evonet.egg-info/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: evonet
-Version: 0.1.0.dev26
+Version: 0.1.0.dev28
 Summary: Evolvable neural network core for integration with EvoLib
 Author-email: EvoLib <evolib@dismail.de>
 License: MIT License

{evonet-0.1.0.dev26 → evonet-0.1.0.dev28}/pyproject.toml

@@ -4,7 +4,7 @@ build-backend = "setuptools.build_meta"
 
 [project]
 name = "evonet"
-version = "0.1.0.dev26"
+version = "0.1.0.dev28"
 description = "Evolvable neural network core for integration with EvoLib"
 authors = [
     { name = "EvoLib", email = "evolib@dismail.de" }
@@ -48,7 +48,7 @@ include = ["evonet*"]
 # Tool-Konfigurationen
 [tool.black]
 line-length = 88
-target-version = ["py311"]
+target-version = ["py312"]
 
 [tool.isort]
 profile = "black"

{evonet-0.1.0.dev26 → evonet-0.1.0.dev28}/tests/test_recurrent_dynamics.py

@@ -181,3 +181,12 @@ def test_reset_full_clears_delay_history() -> None:
     # After full reset, history is gone -> delay contribution must be 0.0
     y = net.calc([0.0])[0]
     assert y == pytest.approx(0.0)
+
+
+def test_recurrent_delay_zero_normalizes_to_one() -> None:
+    net, n_in, n_hid, _ = _mk_linear_ih_h_o()
+    c = net.add_connection(
+        n_hid, n_hid, weight=1.0, conn_type=ConnectionType.RECURRENT, delay=0
+    )
+    assert c is not None
+    assert c.delay == 1