Add VAMP, CKTests (MSM and VAMP)

.circleci/config.yml

@@ -16,6 +16,7 @@ jobs:
       - checkout
       - run: conda config --add channels conda-forge
       - run: conda config --set always_yes true
+      - run: conda config --set channel_priority strict
       - run: conda config --set quiet true
       - run: conda install conda-build conda-verify -c defaults
       - run: mkdir -p $CIRCLE_ARTIFACTS $CIRCLE_TEST_REPORTS

sktime/base.py

@@ -4,7 +4,9 @@
 from sklearn.base import _pprint as pprint_sklearn
 
 
-class _base_methods_mixin(object):
+class _base_methods_mixin(object, metaclass=abc.ABCMeta):
+    """ defines common methods used by both Estimator and Model classes.
+    """
 
     def __repr__(self):
         name = '{cls}-{id}:'.format(id=id(self), cls=self.__class__.__name__)
@@ -89,10 +91,19 @@ def copy(self):
 class Estimator(_base_methods_mixin):
 
     def __init__(self, model=None):
-        self._model = model if model is not None else self._create_model()
+        # we only need to create a default model in case the subclassing Estimator provides the partial_fit interface.
+        if hasattr(self.__class__, 'partial_fit') and model is None:
+            self._model = self._create_model()
+        # TODO: not tested (e.g. by partially fitted models.
+        elif model is not None:
+            self._model = model
 
     @abc.abstractmethod
     def fit(self, data):
+        """ performs a fit of this estimator with data. Creates a new model instance by default.
+        :param data:
+        :return: self
+        """
         pass
 
     def fetch_model(self) -> Model:
@@ -102,6 +113,11 @@ def fetch_model(self) -> Model:
     def _create_model(self):
         pass
 
+    def __getattribute__(self, item):
+        if item == 'fit':
+            self._model = self._create_model()
+        return super(_base_methods_mixin, self).__getattribute__(item)
+
 
 class Transformer(object):
 

sktime/decomposition/tica.py

@@ -266,19 +266,19 @@ class TICA(Estimator, Transformer):
     """
     def __init__(self, lagtime, epsilon=1e-6, reversible=True, dim=0.95,
                  scaling='kinetic_map', ncov=5):
-        super(TICA, self).__init__()
         # tica parameters
-        self._model.epsilon = epsilon
-        self._model.dim = dim
-        self._model.scaling = scaling
+        self.epsilon = epsilon
+        self.dim = dim
+        self.scaling = scaling
 
         # online cov parameters
         self.reversible = reversible
         self._covar = OnlineCovariance(lagtime=lagtime, compute_c00=True, compute_c0t=True, compute_ctt=False, remove_data_mean=True,
                                        reversible=self.reversible, bessels_correction=False, ncov=ncov)
+        super(TICA, self).__init__()
 
     def _create_model(self) -> TICAModel:
-        return TICAModel()
+        return TICAModel(scaling=self.scaling, dim=self.dim, epsilon=self.epsilon)
 
     def transform(self, data):
         r"""Projects the data onto the dominant independent components.