merge devel branch in

GPy/inference/latent_function_inference/__init__.py

@@ -38,6 +38,25 @@ class LatentFunctionInference(object):
         """
         pass
 
+class InferenceMethodList(LatentFunctionInference, list):
+
+    def on_optimization_start(self):
+        for inf in self:
+            inf.on_optimization_start()
+
+    def on_optimization_end(self):
+        for inf in self:
+            inf.on_optimization_end()
+    
+    def __getstate__(self):
+        state = []
+        for inf in self:
+            state.append(inf)
+        return state
+    
+    def __setstate__(self, state):
+        for inf in state:
+            self.append(inf)
 
 from exact_gaussian_inference import ExactGaussianInference
 from laplace import Laplace

GPy/inference/latent_function_inference/posterior.py

@@ -95,7 +95,7 @@ class Posterior(object):
         """
         if self._covariance is None:
             #LiK, _ = dtrtrs(self.woodbury_chol, self._K, lower=1)
-            self._covariance = self._K - (np.tensordot(np.dot(np.atleast_3d(self.woodbury_inv).T, self._K), self._K, [1,0]).T).squeeze()
+            self._covariance = (np.atleast_3d(self._K) - np.tensordot(np.dot(np.atleast_3d(self.woodbury_inv).T, self._K), self._K, [1,0]).T).squeeze()
             #self._covariance = self._K - self._K.dot(self.woodbury_inv).dot(self._K)
         return self._covariance
 

GPy/inference/latent_function_inference/var_dtc.py

@@ -202,6 +202,17 @@ class VarDTCMissingData(LatentFunctionInference):
     def set_limit(self, limit):
         self._Y.limit = limit
 
+    def __getstate__(self):
+        # has to be overridden, as Cacher objects cannot be pickled. 
+        return self._Y.limit, self._inan
+
+    def __setstate__(self, state):
+        # has to be overridden, as Cacher objects cannot be pickled. 
+        from ...util.caching import Cacher
+        self.limit = state[0]
+        self._inan = state[1]
+        self._Y = Cacher(self._subarray_computations, self.limit)
+
     def _subarray_computations(self, Y):
         if self._inan is None:
             inan = np.isnan(Y)
@@ -272,7 +283,11 @@ class VarDTCMissingData(LatentFunctionInference):
             else: beta = beta_all
 
             VVT_factor = (beta*y)
-            VVT_factor_all[v, ind].flat = VVT_factor.flat
+            try:
+                VVT_factor_all[v, ind].flat = VVT_factor.flat
+            except ValueError:
+                mult = np.ravel_multi_index((v.nonzero()[0][:,None],ind[None,:]), VVT_factor_all.shape)
+                VVT_factor_all.flat[mult] = VVT_factor
             output_dim = y.shape[1]
 
             psi0 = psi0_all[v]