added log transform layer

phygnn/layers/custom_layers.py

@@ -955,3 +955,44 @@ def call(self, x):
         """
 
         return tf.math.maximum(tf.math.sigmoid(x), x + 0.5)
+
+
+class LogTransform(tf.keras.layers.Layer):
+    """Log transform or inverse transform of data
+
+    ``y = log(x + adder)`` or ``y = exp(x) - adder`` for inverse
+    """
+
+    def __init__(self, name=None, adder=0, inverse=False):
+        """
+        Parameters
+        ----------
+        name : str | None
+            Name of the tensorflow layer
+        adder : float
+            Adder for ``y = log(x + adder)``
+        inverse : bool
+            Option to perform the inverse operation e.g. ``y = exp(x) - adder``
+        """
+
+        super().__init__(name=name)
+        self.adder = adder
+        self.inverse = inverse
+
+    def call(self, x):
+        """Operates on x with (inverse) log transform
+
+        Parameters
+        ----------
+        x : tf.Tensor
+            Input tensor
+
+        Returns
+        -------
+        y : tf.Tensor
+            Output ``y = log(x + adder)`` or ``y = exp(x) - adder`` if inverse
+        """
+        if not self.inverse:
+            return tf.math.log(x + self.adder)
+        else:
+            return tf.math.exp(x) - self.adder

tests/test_layers.py

@@ -17,6 +17,7 @@
     FunctionalLayer,
     GaussianAveragePooling2D,
     SigLin,
+    LogTransform,
 )
 from phygnn.layers.handlers import HiddenLayers, Layers
 from phygnn import TfModel
@@ -510,3 +511,21 @@ def test_siglin():
     assert x.shape == y.shape
     assert (y > 0).all()
     assert np.allclose(y[mid:], x[mid:] + 0.5)
+
+
+def test_logtransform():
+    """Test the log transform layer"""
+    n_points = 1000
+    lt = LogTransform(adder=0)
+    x = np.linspace(0, 10, n_points + 1)
+    y = lt(x).numpy()
+    assert x.shape == y.shape
+    assert y[0] == -np.inf
+    lt = LogTransform(adder=1)
+    ilt = LogTransform(adder=1, inverse=True)
+    x = np.linspace(0, 10, n_points + 1)
+    y = lt(x).numpy()
+    xinv = ilt(y).numpy()
+    assert not np.isnan(y).any()
+    assert np.allclose(y, np.log(x + 1))
+    assert np.allclose(x, xinv)