improve substitute, update notebook output

examples/Introduction.ipynb

@@ -667,10 +667,7 @@
     {
      "data": {
       "text/plain": [
-       "[0:__fork__\"a\"  >3 >9,\n",
-       " 1:__fork__\"b\"  >4 >10,\n",
-       " 2:__fork__\"s\" <5 ,\n",
-       " 4:DFF\"cin\" <1 >0]"
+       "[0:__fork__\"a\" >3 >9, 1:__fork__\"b\" >4 >10, 2:__fork__\"s\" <5, 4:DFF\"cin\" <1 >0]"
       ]
      },
      "execution_count": 23,
@@ -1312,10 +1309,7 @@
     {
      "data": {
       "text/plain": [
-       "[0:__fork__\"a\"  >3 >9,\n",
-       " 1:__fork__\"b\"  >4 >10,\n",
-       " 2:__fork__\"s\" <5 ,\n",
-       " 4:DFF\"cin\" <1 >0]"
+       "[0:__fork__\"a\" >3 >9, 1:__fork__\"b\" >4 >10, 2:__fork__\"s\" <5, 4:DFF\"cin\" <1 >0]"
       ]
      },
      "execution_count": 37,
@@ -2045,10 +2039,10 @@
      "name": "stdout",
      "output_type": "stream",
      "text": [
-      "Waveforms              : 64293.5 kiB\n",
+      "Waveforms              : 64294.0 kiB\n",
       "State Allocation Table : 129.3 kiB\n",
       "Circuit Timing         : 1034.1 kiB\n",
-      "Circuit Netlist        : 750.8 kiB\n",
+      "Circuit Netlist        : 1126.1 kiB\n",
       "Sequential State       : 726.0 kiB\n"
      ]
     }

src/kyupy/circuit.py

@@ -343,7 +343,7 @@ class Circuit:
         the signal lines are connected appropriately. The number and arrangement
         of the input and output ports must match the pins of the replaced node.
         """
-        impl_in_lines = [n.outs[0] for n in impl.io_nodes if len(n.outs) > 0]
+        impl_in_lines = [n.outs[0] for n in impl.io_nodes if len(n.ins) == 0]
         impl_out_lines = [n.ins[0] for n in impl.io_nodes if len(n.ins) > 0]
         node_in_lines = list(node.ins) + [None] * (len(impl_in_lines)-len(node.ins))
         node_out_lines = list(node.outs) + [None] * (len(impl_out_lines)-len(node.outs))
@@ -354,24 +354,28 @@ class Circuit:
         node.remove()
         node_map = dict()
         ios = set(impl.io_nodes)
-        for n in impl.nodes:
+        for n in impl.nodes:  # add all nodes to main circuit
             if n not in ios:
                 node_map[n] = Node(self, node.name + '~' + n.name, n.kind)
-        for l in impl.lines:
-            if l in in_lines_map:
-                ll = in_lines_map[l]
-                if ll is not None:
+            elif len(n.outs) > 0 and len(n.ins) > 0:  # output is also read by impl. circuit, need to add a fork.
+                node_map[n] = Node(self, node.name + '~' + n.name)
+        for l in impl.lines:  # add all internal lines to main circuit
+            if l.reader in node_map and l.driver in node_map:
+                Line(self, (node_map[l.driver], l.driver_pin), (node_map[l.reader], l.reader_pin))
+        for l, ll in in_lines_map.items():  # connect inputs
+            if ll is None: continue
             ll.reader = node_map[l.reader]
             ll.reader_pin = l.reader_pin
             ll.reader.ins[ll.reader_pin] = ll
-            elif l in out_lines_map:
-                ll = out_lines_map[l]
-                if ll is not None:
+        for l, ll in out_lines_map.items():  # connect outputs
+            if ll is None: continue
+            if len(l.reader.outs) > 0:  # output is also read by impl. circuit, connect to fork.
+                ll.driver = node_map[l.reader]
+                ll.driver_pin = len(l.reader.outs)
+            else:
                 ll.driver = node_map[l.driver]
                 ll.driver_pin = l.driver_pin
             ll.driver.outs[ll.driver_pin] = ll
-            else:
-                Line(self, (node_map[l.driver], l.driver_pin), (node_map[l.reader], l.reader_pin))
 
     def copy(self):
         """Returns a deep copy of the circuit.