Remove printing for useless nodes when generating python code or latex formula

src/tensorneat/common/graph.py

@@ -94,6 +94,27 @@ def topological_sort_python(
 
     return topo_order, topo_layer
 
+def find_useful_nodes(
+    nodes: Union[Set[int], List[int]],
+    conns: Union[Set[Tuple[int, int]], List[Tuple[int, int]]],
+    output_idx: Set[int],
+) -> Set[int]:
+    """
+    Find all useful nodes (really contribute to outputs)
+    """
+    useful_nodes = set()
+    useful_nodes = useful_nodes | output_idx
+    while True:
+        aux = set()
+        for in_, out in conns:
+            if out in useful_nodes and in_ not in useful_nodes:
+                aux.add(in_)
+        if len(aux) == 0:  # no new nodes
+            break
+        else:
+            useful_nodes = useful_nodes | aux
+    # print(f"All nodes cnt={len(nodes)}, useful nodes cnt={len(useful_nodes)}")
+    return useful_nodes            
         
 @jit
 def check_cycles(nodes: Array, conns: Array, from_idx, to_idx) -> Array:

src/tensorneat/common/sympy_tools.py

@@ -43,11 +43,14 @@ def replace_variable_names(expression, mode):
     return expression_str
 
 
-def to_latex_code(symbols, args_symbols, input_symbols, nodes_exprs, output_exprs, forward_func, topo_order):
+def to_latex_code(symbols, args_symbols, input_symbols, nodes_exprs, output_exprs, forward_func, topo_order, useful_nodes):
     input_cnt, hidden_cnt, output_cnt, norm_symbols = analysis_nodes_exprs(nodes_exprs)
     res = "\\begin{align}\n"
     
     for i in topo_order[input_cnt: ]:
+        # do not add node that does not contribute to output; useful nodes may be nan and cause error
+        if i not in useful_nodes:
+            continue
         symbol = symbols[i]
         expr = nodes_exprs[symbol].subs(args_symbols).subs(norm_symbols)
         rounded_expr = round_expr(expr, 2)
@@ -58,7 +61,7 @@ def to_latex_code(symbols, args_symbols, input_symbols, nodes_exprs, output_expr
     return res
 
 
-def to_python_code(symbols, args_symbols, input_symbols, nodes_exprs, output_exprs, forward_func, topo_order):
+def to_python_code(symbols, args_symbols, input_symbols, nodes_exprs, output_exprs, forward_func, topo_order, useful_nodes):
     input_cnt, hidden_cnt, output_cnt, norm_symbols = analysis_nodes_exprs(nodes_exprs)
     res = ""
     if hidden_cnt > 0:
@@ -66,6 +69,9 @@ def to_python_code(symbols, args_symbols, input_symbols, nodes_exprs, output_exp
     res += f"o = np.zeros({output_cnt})\n"
 
     for i in topo_order[input_cnt: ]:
+        # do not add node that does not contribute to output; useful nodes may be nan and cause error
+        if i not in useful_nodes:
+            continue
         symbol = symbols[i]
         expr = nodes_exprs[symbol].subs(args_symbols).subs(norm_symbols)
         rounded_expr = round_expr(expr, 6)

src/tensorneat/genome/default.py

@@ -13,6 +13,7 @@ from .utils import unflatten_conns, extract_gene_attrs, extract_gene_attrs
 from tensorneat.common import (
     topological_sort,
     topological_sort_python,
+    find_useful_nodes,
     I_INF,
     attach_with_inf,
     ACT,
@@ -131,6 +132,11 @@ class DefaultGenome(BaseGenome):
         )
         network["topo_order"] = topo_order
         network["topo_layers"] = topo_layers
+        network["useful_nodes"] = find_useful_nodes(
+            set(network["nodes"]), 
+            set(network["conns"]), 
+            set(self.output_idx)
+        )
         return network
 
     def sympy_func(
@@ -251,7 +257,8 @@ class DefaultGenome(BaseGenome):
             nodes_exprs,
             output_exprs,
             forward_func,
-            network["topo_order"]
+            network["topo_order"],
+            network["useful_nodes"]
         )
 
     def visualize(

src/tensorneat/genome/utils.py

@@ -98,7 +98,7 @@ def re_cound_idx(nodes, conns, input_idx, output_idx):
     for i, key in enumerate(nodes[:, 0]):
         if np.isnan(key):
             continue
-        if np.in1d(key, input_idx + output_idx):
+        if np.isin(key, input_idx + output_idx):
             continue
         old2new[int(key)] = next_key
         next_key += 1