From f16db004383a75b25b2d4093f07eec261e0368cb Mon Sep 17 00:00:00 2001
From: Zhenwen Dai <z.dai@sheffield.ac.uk>
Date: Wed, 2 Jul 2014 14:21:27 +0100
Subject: [PATCH] fix sparse_gplvm

---
 GPy/kern/_src/rbf.py       |  9 --------
 GPy/models/sparse_gplvm.py | 44 +++++++-------------------------------
 2 files changed, 8 insertions(+), 45 deletions(-)

diff --git a/GPy/kern/_src/rbf.py b/GPy/kern/_src/rbf.py
index 471c3305..90c9100b 100644
--- a/GPy/kern/_src/rbf.py
+++ b/GPy/kern/_src/rbf.py
@@ -43,15 +43,6 @@ class RBF(Stationary):
     def __setstate__(self, state):
         return super(RBF, self).__setstate__(state)
 
-#     def copy(self):
-#         k = super(RBF, self).copy()
-#         # Make sure the copy of the kernel instance has different instance of psicomp
-#         if k.useGPU:
-#             k.psicomp = PSICOMP_RBF_GPU()
-#         else:
-#             k.psicomp = PSICOMP_RBF()
-#         return k
-
     #---------------------------------------#
     #             PSI statistics            #
     #---------------------------------------#
diff --git a/GPy/models/sparse_gplvm.py b/GPy/models/sparse_gplvm.py
index 638da63e..b1b3c2db 100644
--- a/GPy/models/sparse_gplvm.py
+++ b/GPy/models/sparse_gplvm.py
@@ -11,7 +11,7 @@ from GPy.models.gplvm import GPLVM
 # from ..core import model
 # from ..util.linalg import pdinv, PCA
 
-class SparseGPLVM(SparseGPRegression, GPLVM):
+class SparseGPLVM(SparseGPRegression):
     """
     Sparse Gaussian Process Latent Variable Model
 
@@ -23,40 +23,12 @@ class SparseGPLVM(SparseGPRegression, GPLVM):
     :type init: 'PCA'|'random'
 
     """
-    def __init__(self, Y, input_dim, kernel=None, init='PCA', num_inducing=10):
-        X = self.initialise_latent(init, input_dim, Y)
+    def __init__(self, Y, input_dim, X=None, kernel=None, init='PCA', num_inducing=10):
+        if X is None:
+            from ..util.initialization import initialize_latent
+            X, fracs = initialize_latent(init, input_dim, Y)
         SparseGPRegression.__init__(self, X, Y, kernel=kernel, num_inducing=num_inducing)
-        self.ensure_default_constraints()
 
-    def _get_param_names(self):
-        return (sum([['X_%i_%i' % (n, q) for q in range(self.input_dim)] for n in range(self.num_data)], [])
-                + SparseGPRegression._get_param_names(self))
-
-    def _get_params(self):
-        return np.hstack((self.X.flatten(), SparseGPRegression._get_params(self)))
-
-    def _set_params(self, x):
-        self.X = x[:self.X.size].reshape(self.num_data, self.input_dim).copy()
-        SparseGPRegression._set_params(self, x[self.X.size:])
-
-    def log_likelihood(self):
-        return SparseGPRegression.log_likelihood(self)
-
-    def dL_dX(self):
-        dL_dX = self.kern.dKdiag_dX(self.dL_dpsi0, self.X)
-        dL_dX += self.kern.gradients_X(self.dL_dpsi1, self.X, self.Z)
-
-        return dL_dX
-
-    def _log_likelihood_gradients(self):
-        return np.hstack((self.dL_dX().flatten(), SparseGPRegression._log_likelihood_gradients(self)))
-
-    def plot(self):
-        GPLVM.plot(self)
-        # passing Z without a small amout of jitter will induce the white kernel where we don;t want it!
-        mu, var, upper, lower = SparseGPRegression.predict(self, self.Z + np.random.randn(*self.Z.shape) * 0.0001)
-        pb.plot(mu[:, 0] , mu[:, 1], 'ko')
-
-    def plot_latent(self, *args, **kwargs):
-        input_1, input_2 = GPLVM.plot_latent(*args, **kwargs)
-        pb.plot(m.Z[:, input_1], m.Z[:, input_2], '^w')
+    def parameters_changed(self):
+        super(SparseGPLVM, self).parameters_changed()
+        self.X.gradient = self.kern.gradients_X(self.grad_dict['dL_dKnm'], self.X, self.Z)