[ENH] Test coverage for AEFCNNetwork Improved

aeon/networks/tests/test_ae_fcn.py

@@ -0,0 +1,288 @@
+"""Test for the AEFCNNetwork class."""
+
+import pytest
+
+from aeon.networks import AEFCNNetwork
+from aeon.utils.validation._dependencies import _check_soft_dependencies
+
+
+@pytest.mark.skipif(
+    not _check_soft_dependencies(["tensorflow"], severity="none"),
+    reason="Tensorflow soft dependency unavailable.",
+)
+def test_aefcn_default():
+    """Default testing for aefcn."""
+    model = AEFCNNetwork()
+    assert model.latent_space_dim == 128
+    assert model.temporal_latent_space is False
+    assert model.n_layers == 3
+    assert model.n_filters is None
+    assert model.kernel_size is None
+    assert model.activation == "relu"
+    assert model.padding == "same"
+    assert model.strides == 1
+    assert model.dilation_rate == 1
+    assert model.use_bias is True
+
+
+@pytest.mark.skipif(
+    not _check_soft_dependencies(["tensorflow"], severity="none"),
+    reason="Tensorflow soft dependency unavailable.",
+)
+@pytest.mark.parametrize("latent_space_dim", [64, 128, 256])
+def test_aefcn_latent_space(latent_space_dim):
+    """Test AEFCNNetwork with different latent space dimensions."""
+    import tensorflow as tf
+
+    aefcn = AEFCNNetwork(latent_space_dim=latent_space_dim)
+    encoder, decoder = aefcn.build_network((1000, 5))
+    assert isinstance(encoder, tf.keras.models.Model)
+    assert isinstance(decoder, tf.keras.models.Model)
+
+
+@pytest.mark.skipif(
+    not _check_soft_dependencies(["tensorflow"], severity="none"),
+    reason="Tensorflow soft dependency unavailable.",
+)
+@pytest.mark.parametrize(
+    "kernel_size, should_raise",
+    [
+        ([8, 5, 3], False),
+        (3, False),
+        ([5, 5], True),
+        ([3, 3, 3, 3], True),
+    ],
+)
+def test_aefcnnetwork_kernel_size(kernel_size, should_raise):
+    """Test AEFCNNetwork with different kernel sizes."""
+    import tensorflow as tf
+
+    if should_raise:
+        with pytest.raises(
+            ValueError,
+            match="Number of kernels .* should be the same as number of layers",
+        ):
+            AEFCNNetwork(kernel_size=kernel_size, n_layers=3).build_network((1000, 5))
+    else:
+        aefcn = AEFCNNetwork(kernel_size=kernel_size, n_layers=3)
+        encoder, decoder = aefcn.build_network((1000, 5))
+        assert isinstance(encoder, tf.keras.models.Model)
+        assert isinstance(decoder, tf.keras.models.Model)
+
+
+@pytest.mark.skipif(
+    not _check_soft_dependencies(["tensorflow"], severity="none"),
+    reason="Tensorflow soft dependency unavailable.",
+)
+@pytest.mark.parametrize(
+    "n_filters, should_raise",
+    [
+        ([128, 256, 128], False),
+        (32, False),
+        ([32, 64], True),
+        ([16, 32, 64, 128], True),
+    ],
+)
+def test_aefcnnetwork_n_filters(n_filters, should_raise):
+    """Test AEFCNNetwork with different number of filters."""
+    import tensorflow as tf
+
+    if should_raise:
+        with pytest.raises(
+            ValueError,
+            match="Number of filters .* should be the same as number of layers",
+        ):
+            AEFCNNetwork(n_filters=n_filters, n_layers=3).build_network((1000, 5))
+    else:
+        aefcn = AEFCNNetwork(n_filters=n_filters, n_layers=3)
+        encoder, decoder = aefcn.build_network((1000, 5))
+        assert isinstance(encoder, tf.keras.models.Model)
+        assert isinstance(decoder, tf.keras.models.Model)
+
+
+@pytest.mark.skipif(
+    not _check_soft_dependencies(["tensorflow"], severity="none"),
+    reason="Tensorflow soft dependency unavailable.",
+)
+@pytest.mark.parametrize(
+    "dilation_rate, should_raise",
+    [
+        ([1, 2, 1], False),
+        (2, False),
+        ([1, 2], True),
+        ([1, 2, 2, 1], True),
+    ],
+)
+def test_aefcnnetwork_dilation_rate(dilation_rate, should_raise):
+    """Test AEFCNNetwork with different dilation rates."""
+    import tensorflow as tf
+
+    if should_raise:
+        with pytest.raises(
+            ValueError,
+            match="Number of dilations .* should be the same as number of layers",
+        ):
+            AEFCNNetwork(dilation_rate=dilation_rate, n_layers=3).build_network(
+                (1000, 5)
+            )
+    else:
+        aefcn = AEFCNNetwork(dilation_rate=dilation_rate, n_layers=3)
+        encoder, decoder = aefcn.build_network((1000, 5))
+        assert isinstance(encoder, tf.keras.models.Model)
+        assert isinstance(decoder, tf.keras.models.Model)
+
+
+@pytest.mark.skipif(
+    not _check_soft_dependencies(["tensorflow"], severity="none"),
+    reason="Tensorflow soft dependency unavailable.",
+)
+@pytest.mark.parametrize(
+    "strides, should_raise",
+    [
+        ([1, 2, 1], False),
+        (2, False),
+        ([1, 2], True),
+        ([1, 2, 2, 1], True),
+    ],
+)
+def test_aefcnnetwork_strides(strides, should_raise):
+    """Test AEFCNNetwork with different strides."""
+    import tensorflow as tf
+
+    if should_raise:
+        with pytest.raises(
+            ValueError,
+            match="Number of strides .* should be the same as number of layers",
+        ):
+            AEFCNNetwork(strides=strides, n_layers=3).build_network((1000, 5))
+    else:
+        aefcn = AEFCNNetwork(strides=strides, n_layers=3)
+        encoder, decoder = aefcn.build_network((1000, 5))
+        assert isinstance(encoder, tf.keras.models.Model)
+        assert isinstance(decoder, tf.keras.models.Model)
+
+
+@pytest.mark.skipif(
+    not _check_soft_dependencies(["tensorflow"], severity="none"),
+    reason="Tensorflow soft dependency unavailable.",
+)
+@pytest.mark.parametrize(
+    "padding, should_raise",
+    [
+        (["same", "valid", "same"], False),
+        ("same", False),
+        (["same", "valid"], True),
+        (
+            ["same", "valid", "same", "valid"],
+            True,
+        ),
+    ],
+)
+def test_aefcnnetwork_padding(padding, should_raise):
+    """Test AEFCNNetwork with different paddings."""
+    import tensorflow as tf
+
+    if should_raise:
+        with pytest.raises(
+            ValueError,
+            match="Number of paddings .* should be the same as number of layers",
+        ):
+            AEFCNNetwork(padding=padding, n_layers=3).build_network((1000, 5))
+    else:
+        aefcn = AEFCNNetwork(padding=padding, n_layers=3)
+        encoder, decoder = aefcn.build_network((1000, 5))
+        assert isinstance(encoder, tf.keras.models.Model)
+        assert isinstance(decoder, tf.keras.models.Model)
+
+
+@pytest.mark.skipif(
+    not _check_soft_dependencies(["tensorflow"], severity="none"),
+    reason="Tensorflow soft dependency unavailable.",
+)
+@pytest.mark.parametrize(
+    "activation, should_raise",
+    [
+        (["relu", "sigmoid", "tanh"], False),
+        ("sigmoid", False),
+        (["relu", "sigmoid"], True),
+        (
+            ["relu", "sigmoid", "tanh", "softmax"],
+            True,
+        ),
+    ],
+)
+def test_aefcnnetwork_activation(activation, should_raise):
+    """Test AEFCNNetwork with different activations."""
+    import tensorflow as tf
+
+    if should_raise:
+        with pytest.raises(
+            ValueError,
+            match="Number of activations .* should be the same as number of layers",
+        ):
+            AEFCNNetwork(activation=activation, n_layers=3).build_network((1000, 5))
+    else:
+        aefcn = AEFCNNetwork(activation=activation, n_layers=3)
+        encoder, decoder = aefcn.build_network((1000, 5))
+        assert isinstance(encoder, tf.keras.models.Model)
+        assert isinstance(decoder, tf.keras.models.Model)
+
+
+@pytest.mark.skipif(
+    not _check_soft_dependencies(["tensorflow"], severity="none"),
+    reason="Tensorflow soft dependency unavailable.",
+)
+@pytest.mark.parametrize(
+    "use_bias, should_raise",
+    [
+        ([True, False, True], False),
+        (True, False),
+        ([True, False], True),
+        ([True, False, True, False], True),
+    ],
+)
+def test_aefcnnetwork_use_bias(use_bias, should_raise):
+    """Test AEFCNNetwork with different use_bias values."""
+    import tensorflow as tf
+
+    if should_raise:
+        with pytest.raises(
+            ValueError,
+            match="Number of biases .* should be the same as number of layers",
+        ):
+            AEFCNNetwork(use_bias=use_bias, n_layers=3).build_network((1000, 5))
+    else:
+        aefcn = AEFCNNetwork(use_bias=use_bias, n_layers=3)
+        encoder, decoder = aefcn.build_network((1000, 5))
+        assert isinstance(encoder, tf.keras.models.Model)
+        assert isinstance(decoder, tf.keras.models.Model)
+
+
+@pytest.mark.skipif(
+    not _check_soft_dependencies(["tensorflow"], severity="none"),
+    reason="Tensorflow soft dependency unavailable.",
+)
+@pytest.mark.parametrize("temporal_latent_space", [True, False])
+def test_aefcnnetwork_temporal_latent_space(temporal_latent_space):
+    """Test for temporal latent space."""
+    import tensorflow as tf
+
+    input_shape = (1000, 5)
+
+    aefcn = AEFCNNetwork(
+        latent_space_dim=128, temporal_latent_space=temporal_latent_space
+    )
+
+    encoder, decoder = aefcn.build_network(input_shape)
+
+    assert isinstance(encoder, tf.keras.models.Model)
+    assert isinstance(decoder, tf.keras.models.Model)
+
+    if temporal_latent_space:
+        assert any(
+            isinstance(layer, tf.keras.layers.Conv1D) for layer in encoder.layers
+        ), "Expected Conv1D layer in encoder but not found."
+    else:
+        assert any(
+            isinstance(layer, tf.keras.layers.Dense) for layer in decoder.layers
+        ), "Expected Dense layer in decoder but not found."