brainstate 0.1.10__py2.py3-none-any.whl → 0.2.1__py2.py3-none-any.whl

brainstate/graph/_node_test.py

@@ -0,0 +1,589 @@
+# Copyright 2024 BrainX Ecosystem Limited. All Rights Reserved.
+#
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at
+#
+#     http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing, software
+# distributed under the License is distributed on an "AS IS" BASIS,
+# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and
+# limitations under the License.
+# ==============================================================================
+
+import unittest
+
+import brainstate
+from brainstate._state import State
+from brainstate.graph._node import (
+    Node,
+    _node_flatten,
+    _node_set_key,
+    _node_pop_key,
+    _node_create_empty,
+    _node_clear
+)
+
+
+class TestNode(unittest.TestCase):
+    """Test suite for the Node class."""
+
+    def test_node_creation(self):
+        """Test basic node creation."""
+
+        class SimpleNode(Node):
+            def __init__(self, value):
+                self.value = value
+
+        node = SimpleNode(10)
+        self.assertEqual(node.value, 10)
+        self.assertIsInstance(node, Node)
+
+    def test_node_subclass_registration(self):
+        """Test that Node subclasses are automatically registered."""
+
+        class TestNode(Node):
+            pass
+
+        # The subclass should be registered automatically via __init_subclass__
+        node = TestNode()
+        self.assertIsInstance(node, Node)
+
+    def test_graph_invisible_attrs(self):
+        """Test that graph_invisible_attrs works correctly."""
+
+        class NodeWithInvisible(Node):
+            graph_invisible_attrs = ('_private', '_internal')
+
+            def __init__(self):
+                self.public = 1
+                self._private = 2
+                self._internal = 3
+
+        node = NodeWithInvisible()
+        flattened, static = _node_flatten(node)
+
+        # Check that only public attribute is in flattened
+        keys = [k for k, v in flattened]
+        self.assertIn('public', keys)
+        self.assertNotIn('_private', keys)
+        self.assertNotIn('_internal', keys)
+
+    def test_deepcopy_using_treefy(self):
+        """Test deep copying of nodes using treefy_split/merge."""
+
+        class NodeWithData(Node):
+            def __init__(self, data=None):
+                if data is not None:
+                    self.data = data
+                    self.nested = {'a': 1, 'b': [2, 3]}
+
+        original = NodeWithData([1, 2, 3])
+
+        # Use treefy_split and treefy_merge to copy
+        graphdef, state = brainstate.graph.treefy_split(original)
+        # Create a new instance using treefy_merge
+        copied = brainstate.graph.treefy_merge(graphdef, state)
+
+        # Check that it's a different object
+        self.assertIsNot(original, copied)
+
+        # Check that data is present
+        self.assertEqual(original.data, copied.data)
+
+        # Modify copied data shouldn't affect original
+        copied.data.append(4)
+        self.assertEqual(len(original.data), 3)
+        self.assertEqual(len(copied.data), 4)
+
+    def test_node_with_state(self):
+        """Test nodes containing State objects."""
+
+        class NodeWithState(Node):
+            def __init__(self):
+                self.value = State(10)
+                self.normal = 20
+
+        node = NodeWithState()
+        self.assertIsInstance(node.value, State)
+        self.assertEqual(node.value.value, 10)
+        self.assertEqual(node.normal, 20)
+
+    def test_complex_nested_structure(self):
+        """Test nodes with complex nested structures using treefy."""
+
+        class ComplexNode(Node):
+            def __init__(self):
+                self.list_data = [1, 2, [3, 4]]
+                self.dict_data = {'a': 1, 'b': {'c': 2}}
+                self.tuple_data = (1, 2, (3, 4))
+
+        node = ComplexNode()
+
+        # Test using treefy_split and merge
+        graphdef, state = brainstate.graph.treefy_split(node)
+        copied = brainstate.graph.treefy_merge(graphdef, state)
+
+        self.assertEqual(node.list_data, copied.list_data)
+        self.assertEqual(node.dict_data, copied.dict_data)
+        self.assertEqual(node.tuple_data, copied.tuple_data)
+
+
+class TestNodeHelperFunctions(unittest.TestCase):
+    """Test suite for node helper functions."""
+
+    def test_node_flatten(self):
+        """Test _node_flatten function."""
+
+        class TestNode(Node):
+            def __init__(self):
+                self.b = 2
+                self.a = 1
+                self.c = 3
+
+        node = TestNode()
+        flattened, static = _node_flatten(node)
+
+        # Check that attributes are sorted
+        keys = [k for k, v in flattened]
+        self.assertEqual(keys, ['a', 'b', 'c'])
+
+        # Check values
+        values = [v for k, v in flattened]
+        self.assertEqual(values, [1, 2, 3])
+
+        # Check static contains type
+        self.assertEqual(static, (TestNode,))
+
+    def test_node_flatten_with_invisible(self):
+        """Test _node_flatten with invisible attributes."""
+
+        class TestNode(Node):
+            graph_invisible_attrs = ('hidden',)
+
+            def __init__(self):
+                self.visible = 1
+                self.hidden = 2
+
+        node = TestNode()
+        flattened, static = _node_flatten(node)
+
+        keys = [k for k, v in flattened]
+        self.assertIn('visible', keys)
+        self.assertNotIn('hidden', keys)
+
+    def test_node_set_key_simple(self):
+        """Test _node_set_key with simple values."""
+
+        class TestNode(Node):
+            pass
+
+        node = TestNode()
+        _node_set_key(node, 'attr', 10)
+        self.assertEqual(node.attr, 10)
+
+        _node_set_key(node, 'attr', 20)
+        self.assertEqual(node.attr, 20)
+
+    def test_node_set_key_with_state(self):
+        """Test _node_set_key with State objects."""
+
+        class TestNode(Node):
+            def __init__(self):
+                self.state_attr = State(10)
+
+        node = TestNode()
+
+        # Test setting a regular value
+        _node_set_key(node, 'regular_attr', 30)
+        self.assertEqual(node.regular_attr, 30)
+
+        # Test updating with a TreefyState
+        # We'll use the real TreefyState from graph operations
+        graphdef, states = brainstate.graph.treefy_split(node)
+
+        # The states should contain our State object wrapped as TreefyState
+        # When setting with TreefyState, it should update the existing State
+        initial_state = node.state_attr
+
+        # Create a new node and try to set the TreefyState
+        new_node = TestNode()
+        for key, value in states.to_flat().items():
+            if 'state_attr' in key:
+                _node_set_key(new_node, 'state_attr', value)
+                # The State object should be updated via update_from_ref
+                self.assertIsInstance(new_node.state_attr, State)
+
+    def test_node_set_key_invalid_key(self):
+        """Test _node_set_key with invalid key."""
+
+        class TestNode(Node):
+            pass
+
+        node = TestNode()
+
+        with self.assertRaises(KeyError) as context:
+            _node_set_key(node, 123, 'value')
+        self.assertIn('Invalid key', str(context.exception))
+
+    def test_node_pop_key(self):
+        """Test _node_pop_key function."""
+
+        class TestNode(Node):
+            def __init__(self):
+                self.attr1 = 10
+                self.attr2 = 20
+
+        node = TestNode()
+
+        # Pop existing attribute
+        value = _node_pop_key(node, 'attr1')
+        self.assertEqual(value, 10)
+        self.assertFalse(hasattr(node, 'attr1'))
+        self.assertTrue(hasattr(node, 'attr2'))
+
+    def test_node_pop_key_invalid(self):
+        """Test _node_pop_key with invalid key."""
+
+        class TestNode(Node):
+            pass
+
+        node = TestNode()
+
+        # Invalid key type
+        with self.assertRaises(KeyError) as context:
+            _node_pop_key(node, 123)
+        self.assertIn('Invalid key', str(context.exception))
+
+        # Non-existent key
+        with self.assertRaises(KeyError):
+            _node_pop_key(node, 'nonexistent')
+
+    def test_node_create_empty(self):
+        """Test _node_create_empty function."""
+
+        class TestNode(Node):
+            def __init__(self, value=None):
+                self.value = value
+                self.initialized = True
+
+        # Create empty node
+        node = _node_create_empty((TestNode,))
+
+        # Check it's the right type
+        self.assertIsInstance(node, TestNode)
+
+        # Check __init__ was not called
+        self.assertFalse(hasattr(node, 'value'))
+        self.assertFalse(hasattr(node, 'initialized'))
+
+    def test_node_clear(self):
+        """Test _node_clear function."""
+
+        class TestNode(Node):
+            def __init__(self):
+                self.attr1 = 10
+                self.attr2 = 20
+                self.attr3 = [1, 2, 3]
+
+        node = TestNode()
+
+        # Verify attributes exist
+        self.assertTrue(hasattr(node, 'attr1'))
+        self.assertTrue(hasattr(node, 'attr2'))
+        self.assertTrue(hasattr(node, 'attr3'))
+
+        # Clear the node
+        _node_clear(node)
+
+        # Verify attributes are gone
+        self.assertFalse(hasattr(node, 'attr1'))
+        self.assertFalse(hasattr(node, 'attr2'))
+        self.assertFalse(hasattr(node, 'attr3'))
+
+        # Verify node still exists and is valid
+        self.assertIsInstance(node, TestNode)
+
+
+class TestNodeIntegration(unittest.TestCase):
+    """Integration tests for Node with the graph system."""
+
+    def test_node_with_nested_nodes(self):
+        """Test nodes containing other nodes."""
+
+        class ChildNode(Node):
+            def __init__(self, value=None):
+                if value is not None:
+                    self.value = value
+
+        class ParentNode(Node):
+            def __init__(self):
+                self.child1 = ChildNode(10)
+                self.child2 = ChildNode(20)
+                self.data = [1, 2, 3]
+
+        parent = ParentNode()
+
+        # Test using treefy_split and merge
+        graphdef, state = brainstate.graph.treefy_split(parent)
+        copied = brainstate.graph.treefy_merge(graphdef, state)
+
+        self.assertIsNot(parent.child1, copied.child1)
+        self.assertEqual(parent.child1.value, copied.child1.value)
+
+    def test_node_with_list_of_nodes(self):
+        """Test nodes containing lists of other nodes."""
+
+        class ItemNode(Node):
+            def __init__(self, id=None):
+                if id is not None:
+                    self.id = id
+
+        class ContainerNode(Node):
+            def __init__(self):
+                self.items = [ItemNode(i) for i in range(3)]
+
+        container = ContainerNode()
+        graphdef, state = brainstate.graph.treefy_split(container)
+        copied = brainstate.graph.treefy_merge(graphdef, state)
+
+        self.assertEqual(len(container.items), len(copied.items))
+        for orig, cp in zip(container.items, copied.items):
+            self.assertIsNot(orig, cp)
+            self.assertEqual(orig.id, cp.id)
+
+    def test_node_with_dict_of_nodes(self):
+        """Test nodes containing dictionaries of other nodes."""
+
+        class ValueNode(Node):
+            def __init__(self, value=None):
+                if value is not None:
+                    self.value = value
+
+        class DictNode(Node):
+            def __init__(self):
+                self.mapping = {
+                    'a': ValueNode(1),
+                    'b': ValueNode(2),
+                    'c': ValueNode(3)
+                }
+
+        node = DictNode()
+        graphdef, state = brainstate.graph.treefy_split(node)
+        copied = brainstate.graph.treefy_merge(graphdef, state)
+
+        self.assertEqual(set(node.mapping.keys()), set(copied.mapping.keys()))
+        for key in node.mapping:
+            self.assertIsNot(node.mapping[key], copied.mapping[key])
+            self.assertEqual(node.mapping[key].value, copied.mapping[key].value)
+
+
+class TestStateRetrieve(unittest.TestCase):
+    """Tests for state retrieval from nodes."""
+
+    def test_list_of_states_1(self):
+        """Test retrieving states from a list."""
+
+        class Model(brainstate.graph.Node):
+            def __init__(self):
+                self.a = [1, 2, 3]
+                self.b = [brainstate.State(1), brainstate.State(2), brainstate.State(3)]
+
+        m = Model()
+        graphdef, states = brainstate.graph.treefy_split(m)
+        print(states.to_flat())
+        self.assertTrue(len(states.to_flat()) == 3)
+
+    def test_list_of_states_2(self):
+        """Test retrieving states from nested lists."""
+
+        class Model(brainstate.graph.Node):
+            def __init__(self):
+                self.a = [1, 2, 3]
+                self.b = [brainstate.State(1), [brainstate.State(2), brainstate.State(3)]]
+
+        m = Model()
+        graphdef, states = brainstate.graph.treefy_split(m)
+        print(states.to_flat())
+        self.assertTrue(len(states.to_flat()) == 3)
+
+    def test_list_of_node_1(self):
+        """Test retrieving states from a list of nodes."""
+
+        class Model(brainstate.graph.Node):
+            def __init__(self):
+                self.a = [1, 2, 3]
+                self.b = [brainstate.nn.Linear(1, 2), brainstate.nn.Linear(2, 3)]
+
+        m = Model()
+        graphdef, states = brainstate.graph.treefy_split(m)
+        print(states.to_flat())
+        self.assertTrue(len(states.to_flat()) == 2)
+
+    def test_list_of_node_2(self):
+        """Test retrieving states from nested structures of nodes."""
+
+        class Model(brainstate.graph.Node):
+            def __init__(self):
+                self.a = [1, 2, 3]
+                self.b = [brainstate.nn.Linear(1, 2), [brainstate.nn.Linear(2, 3)],
+                          (brainstate.nn.Linear(3, 4), brainstate.nn.Linear(4, 5))]
+
+        m = Model()
+        graphdef, states = brainstate.graph.treefy_split(m)
+        print(states.to_flat())
+        self.assertTrue(len(states.to_flat()) == 4)
+
+    def test_mixed_states_and_nodes(self):
+        """Test nodes with mixed states and sub-nodes."""
+
+        class Model(brainstate.graph.Node):
+            def __init__(self):
+                self.state1 = brainstate.State(1.0)
+                self.state2 = brainstate.State(2.0)
+                self.linear = brainstate.nn.Linear(5, 10)
+                self.data = [1, 2, 3]
+
+        m = Model()
+        graphdef, states = brainstate.graph.treefy_split(m)
+
+        # Should have states from both direct State objects and Linear layer
+        flat_states = states.to_flat()
+        self.assertGreaterEqual(len(flat_states), 2)  # At least the two direct states
+
+
+class TestEdgeCases(unittest.TestCase):
+    """Test edge cases and error conditions."""
+
+    def test_empty_node(self):
+        """Test node with no attributes."""
+
+        class EmptyNode(Node):
+            pass
+
+        node = EmptyNode()
+        flattened, static = _node_flatten(node)
+
+        self.assertEqual(len(flattened), 0)
+        self.assertEqual(static, (EmptyNode,))
+
+    def test_node_with_none_values(self):
+        """Test node with None values."""
+
+        class NoneNode(Node):
+            def __init__(self):
+                self.none_val = None
+                self.real_val = 10
+
+        node = NoneNode()
+        flattened, static = _node_flatten(node)
+
+        values_dict = dict(flattened)
+        self.assertIsNone(values_dict['none_val'])
+        self.assertEqual(values_dict['real_val'], 10)
+
+    def test_node_with_special_attributes(self):
+        """Test node with special Python attributes."""
+
+        class SpecialNode(Node):
+            def __init__(self):
+                self.__dict__['special'] = 'value'
+                self.normal = 'normal'
+
+        node = SpecialNode()
+        self.assertEqual(node.special, 'value')
+        self.assertEqual(node.normal, 'normal')
+
+    def test_circular_reference(self):
+        """Test handling of circular references."""
+
+        class CircularNode(Node):
+            pass
+
+        node1 = CircularNode()
+        node2 = CircularNode()
+        node1.ref = node2
+        node2.ref = node1
+
+        # This should not cause infinite recursion with treefy
+        try:
+            graphdef, state = brainstate.graph.treefy_split(node1)
+            copied = brainstate.graph.treefy_merge(graphdef, state)
+            # Check that circular reference is preserved
+            self.assertIs(copied.ref.ref, copied)
+        except RecursionError:
+            self.fail("Treefy failed with circular reference")
+
+    def test_node_inheritance(self):
+        """Test node inheritance hierarchy."""
+
+        class BaseNode(Node):
+            def __init__(self):
+                self.base_attr = 'base'
+
+        class DerivedNode(BaseNode):
+            def __init__(self):
+                super().__init__()
+                self.derived_attr = 'derived'
+
+        node = DerivedNode()
+        self.assertEqual(node.base_attr, 'base')
+        self.assertEqual(node.derived_attr, 'derived')
+
+        flattened, static = _node_flatten(node)
+        keys = [k for k, v in flattened]
+        self.assertIn('base_attr', keys)
+        self.assertIn('derived_attr', keys)
+
+    def test_node_with_property(self):
+        """Test node with property decorators."""
+
+        class PropertyNode(Node):
+            def __init__(self):
+                self._value = 10
+
+            @property
+            def value(self):
+                return self._value
+
+            @value.setter
+            def value(self, val):
+                self._value = val
+
+        node = PropertyNode()
+        self.assertEqual(node.value, 10)
+
+        node.value = 20
+        self.assertEqual(node.value, 20)
+
+        # Only _value should appear in flattened
+        flattened, static = _node_flatten(node)
+        keys = [k for k, v in flattened]
+        self.assertIn('_value', keys)
+
+    def test_multiple_inheritance(self):
+        """Test node with multiple inheritance."""
+
+        class Mixin:
+            def mixin_method(self):
+                return 'mixin'
+
+        class MultiNode(Node, Mixin):
+            def __init__(self):
+                self.data = 'data'
+
+        node = MultiNode()
+        self.assertEqual(node.mixin_method(), 'mixin')
+        self.assertEqual(node.data, 'data')
+
+        # Test that it still works as a Node with treefy
+        graphdef, state = brainstate.graph.treefy_split(node)
+        copied = brainstate.graph.treefy_merge(graphdef, state)
+        self.assertIsNot(node, copied)
+        self.assertEqual(copied.data, 'data')
+
+
+if __name__ == '__main__':
+    unittest.main()