var_cutoff can be disabled in covariance koopman models. (#255)

Author: clonker

deeptime/decomposition/_koopman.py

@@ -344,15 +344,16 @@ def var_cutoff(self) -> Optional[float]:
         precedence over the :meth:`dim` parameter.
 
         :getter: Yields the current variance cutoff.
-        :setter: Sets a new variance cutoff
+        :setter: Sets a new variance cutoff or disables variance cutoff by setting the value to `None`.
         :type: float or None
         """
         return self._var_cutoff
 
     @var_cutoff.setter
-    def var_cutoff(self, value):
-        assert 0 < value <= 1., "Invalid dimension parameter, if it is given in terms of a variance cutoff, " \
-                                "it can only be in the interval (0, 1]."
+    def var_cutoff(self, value: Optional[float]):
+        assert value is None or 0 < value <= 1., \
+            "Invalid dimension parameter, if it is given in terms of a variance cutoff, " \
+            "it can only be in the interval (0, 1]."
         self._var_cutoff = value
         self._update_output_dimension()
 

tests/decomposition/test_tica.py

@@ -8,6 +8,7 @@
 
 import numpy as np
 import pytest
+from numpy.testing import assert_, assert_equal
 
 from deeptime.covariance import Covariance
 from deeptime.data import ellipsoids
@@ -16,6 +17,22 @@
 from deeptime.numeric import ZeroRankError
 
 
+def test_update_projection_dimension():
+    # tests for https://github.com/deeptime-ml/deeptime/issues/254
+    data = np.random.normal(size=(1000, 50))
+    model = TICA(lagtime=1, var_cutoff=.1).fit_fetch(data)
+    assert_equal(model.var_cutoff, .1)
+    assert_(model.transform(data).shape[1] <= 10)
+    model.var_cutoff = None
+    assert_equal(model.var_cutoff, None)
+    model.dim = 5
+    assert_equal(model.dim, 5)
+    assert_(model.transform(data).shape[1] <= 5)
+    model.dim = 1
+    assert_equal(model.dim, 1)
+    assert_equal(model.transform(data).shape[1], 1)
+
+
 def test_fit_reset():
     lag = 100
     np.random.seed(0)