slicing support for kernel input dimension

GPy/core/parameterization/array_core.py

@@ -16,7 +16,7 @@ class ObservableArray(np.ndarray, Observable):
     __array_priority__ = -1 # Never give back ObservableArray
     def __new__(cls, input_array):
         if not isinstance(input_array, ObservableArray):
-            obj = np.atleast_1d(input_array).view(cls)
+            obj = np.atleast_1d(np.require(input_array, dtype=np.float64, requirements=['W', 'C'])).view(cls)
         else: obj = input_array
         cls.__name__ = "ObservableArray\n     "
         return obj

GPy/core/parameterization/parameter_core.py

@@ -15,7 +15,6 @@ Observable Pattern for patameterization
 
 from transformations import Transformation, Logexp, NegativeLogexp, Logistic, __fixed__, FIXED, UNFIXED
 import numpy as np
-import itertools
 
 __updated__ = '2013-12-16'
 
@@ -43,6 +42,7 @@ class Observable(object):
     _updated = True
     def __init__(self, *args, **kwargs):
         self._observer_callables_ = []
+    
     def __del__(self, *args, **kwargs):
         del self._observer_callables_
         
@@ -551,8 +551,8 @@ class OptimizationHandlable(Constrainable, Observable):
         return p
 
     def _set_params_transformed(self, p):
-        if p is self._param_array_:
-            p = p.copy()
+        #if p is self._param_array_:
+        p = p.copy()
         if self._has_fixes(): self._param_array_[self._fixes_] = p
         else: self._param_array_[:] = p
         self.untransform()

GPy/core/parameterization/variational.py

@@ -66,10 +66,10 @@ class VariationalPosterior(Parameterized):
     def __init__(self, means=None, variances=None, name=None, **kw):
         super(VariationalPosterior, self).__init__(name=name, **kw)
         self.mean = Param("mean", means)
-        self.ndim = self.mean.ndim
-        self.shape = self.mean.shape
         self.variance = Param("variance", variances, Logexp())
         self.add_parameters(self.mean, self.variance)
+        self.ndim = self.mean.ndim
+        self.shape = self.mean.shape
         self.num_data, self.input_dim = self.mean.shape
         if self.has_uncertain_inputs():
             assert self.variance.shape == self.mean.shape, "need one variance per sample and dimenion"
@@ -77,6 +77,18 @@ class VariationalPosterior(Parameterized):
     def has_uncertain_inputs(self):
         return not self.variance is None
 
+    def __getitem__(self, s):
+        import copy
+        n = self.__new__(self.__class__)
+        dc = copy.copy(self.__dict__)
+        dc['mean'] = dc['mean'][s]
+        dc['variance'] = dc['variance'][s]
+        dc['shape'] = dc['mean'].shape
+        dc['ndim'] = dc['ndim']
+        dc['num_data'], dc['input_dim'] = self.mean.shape[0], self.mean.shape[1] if dc['ndim'] > 1 else 1
+        n.__dict__ = dc
+        return n
+    
 
 class NormalPosterior(VariationalPosterior):
     '''