Exclude unnecessary inputs from in useless_composite rewrite

Author: ricardoV94

pytensor/tensor/rewriting/elemwise.py

@@ -1094,21 +1094,30 @@ def print_profile(stream, prof, level=0):
 @register_canonicalize
 @node_rewriter([Elemwise])
 def local_useless_composite(fgraph, node):
-    """For elemwise Composite that have multiple outputs, remove the
-    outputs that are not used.
-
-    """
+    """Remove inputs and outputs of Composite Ops that are not used anywhere."""
     if not isinstance(node.op, Elemwise) or not isinstance(
         node.op.scalar_op, aes.Composite
     ):
         return
     comp = node.op.scalar_op
-    idx = [i for i, o_extern in enumerate(node.outputs) if fgraph.clients[o_extern]]
-    if len(idx) < len(node.outputs):
-        new_outputs = [comp.outputs[i] for i in idx]
-        c = aes.Composite(inputs=comp.inputs, outputs=new_outputs)
-        e = Elemwise(scalar_op=c)(*node.inputs, return_list=True)
-        return dict(zip([node.outputs[i] for i in idx], e))
+    used_outputs_idxs = [i for i, o_extern in enumerate(node.outputs) if fgraph.clients[o_extern]]
+    used_inner_outputs = [comp.outputs[i] for i in used_outputs_idxs]
+    comp_fgraph = FunctionGraph(
+        inputs=comp.inputs, outputs=used_inner_outputs, clone=False
+    )
+    used_inputs_idxs = [
+        i
+        for i, i_intern in enumerate(comp_fgraph.inputs)
+        if comp_fgraph.clients[i_intern]
+    ]
+    used_inner_inputs = [comp.inputs[i] for i in used_inputs_idxs]
+    if len(used_inner_inputs) < len(node.inputs) or len(used_inner_outputs) < len(
+        node.outputs
+    ):
+        used_inputs = [node.inputs[i] for i in used_inputs_idxs]
+        c = aes.Composite(inputs=used_inner_inputs, outputs=used_inner_outputs)
+        e = Elemwise(scalar_op=c)(*used_inputs, return_list=True)
+        return dict(zip([node.outputs[i] for i in used_outputs_idxs], e))
 
 
 @node_rewriter([CAReduce])

tests/tensor/rewriting/test_elemwise.py

@@ -1441,22 +1441,35 @@ def test_nested_composite(self):
 
     def test_local_useless_composite(self):
         x = aes.float32()
-        c = aes.Composite([x], [x + 1, x - 1])
-        X = matrix()
-        o = Elemwise(scalar_op=c)(X)
+        y = aes.float32()
+        c = aes.Composite([x, y], [x + 1, y - 1])
+        X = matrix("X")
+        Y = matrix("Y")
+        o1, o2 = Elemwise(scalar_op=c)(X, Y)
         mode = get_default_mode().including("local_useless_composite")
 
-        f = function([X], o[0], mode=mode)
+        f = function([X, Y], [o1, o2], mode=mode)
         topo = f.maker.fgraph.toposort()
         assert len(topo) == 1
+        assert len(topo[0].inputs) == 2
+        assert len(topo[0].outputs) == 2
+        res1, res2 = f([[1.0]], [[1.0]])
+        utt.assert_allclose(res1, [[2.0]])
+        utt.assert_allclose(res2, [[0.0]])
+
+        f = function([X, Y], o1, mode=mode)
+        topo = f.maker.fgraph.toposort()
+        assert len(topo) == 1
+        assert len(topo[0].inputs) == 1
         assert len(topo[0].outputs) == 1
-        utt.assert_allclose(f([[1.0]]), [[2.0]])
+        utt.assert_allclose(f([[1.0]], [[np.nan]]), [[2.0]])
 
-        f = function([X], o[1], mode=mode)
+        f = function([X, Y], o2, mode=mode)
         topo = f.maker.fgraph.toposort()
         assert len(topo) == 1
+        assert len(topo[0].inputs) == 1
         assert len(topo[0].outputs) == 1
-        utt.assert_allclose(f([[1.0]]), [[0.0]])
+        utt.assert_allclose(f([[np.nan]], [[1.0]]), [[0.0]])
 
 
 def test_local_useless_dimshuffle_makevector():