removed marginal and derivative from posterior object

GPy/core/gp.py

@@ -60,10 +60,10 @@ class GP(Model):
         self.parameters_changed()
 
     def parameters_changed(self):
-        self.posterior = self.inference_method.inference(self.kern, self.X, self.likelihood, self.Y)
+        self.posterior, self._log_marginal_likelihood, self._dL_dK = self.inference_method.inference(self.kern, self.X, self.likelihood, self.Y)
 
     def log_likelihood(self):
-        return self.posterior.log_marginal
+        return self._log_marginal_likelihood
 
     def dL_dtheta_K(self):
         return self.kern.dK_dtheta(self.posterior.dL_dK, self.X)

GPy/core/sparse_gp.py

@@ -6,7 +6,6 @@ import pylab as pb
 from gp import GP
 from parameterization.param import Param
 from ..inference.latent_function_inference import varDTC
-from posterior import Posterior
 
 class SparseGP(GP):
     """

GPy/inference/latent_function_inference/exact_gaussian_inference.py

@@ -53,6 +53,6 @@ class ExactGaussianInference(object):
 
         likelihood.update_gradients(np.diag(dL_dK))
 
-        return Posterior(log_marginal, dL_dK, LW, alpha, K)
+        return Posterior(LW, alpha, K), log_marginal, dL_dK
 
 

GPy/inference/latent_function_inference/posterior.py

@@ -14,10 +14,8 @@ class Posterior(object):
     schemes and the model classes.
 
     """
-    def __init__(self, log_marginal, dL_dK, woodbury_chol=None, woodbury_vector=None, K=None, mean=None, cov=None, K_chol=None):
+    def __init__(self, woodbury_chol=None, woodbury_vector=None, K=None, mean=None, cov=None, K_chol=None):
         """
-        log_marginal: log p(Y|X)
-        dL_dK: d/dK log p(Y|X)
         woodbury_chol : a lower triangular matrix L that satisfies posterior_covariance = K - K L^{-T} L^{-1} K
         woodbury_vector : a matrix (or vector, as Nx1 matrix) M which satisfies posterior_mean = K M
         K : the proir covariance (required for lazy computation of various quantities)
@@ -26,8 +24,6 @@ class Posterior(object):
 
         Not all of the above need to be supplied! You *must* supply:
 
-          log_marginal
-          dL_dK
           K (for lazy computation)
 
        You may supply either:
@@ -46,8 +42,6 @@ class Posterior(object):
         From the supplied quantities, all of the others will be computed on demand (lazy computation)
         """
         #obligatory
-        self.log_marginal = log_marginal
-        self.dL_dK = dL_dK
         self._K = K
 
         if ((woodbury_chol is not None) and (woodbury_vector is not None) and (K is not None)) or ((mean is not None) and (cov is not None) and (K is not None)):

GPy/inference/latent_function_inference/varDTC.py

@@ -2,10 +2,11 @@
 # Licensed under the BSD 3-clause license (see LICENSE.txt)
 
 from posterior import Posterior
-from .../util.linalg import pdinv, dpotrs, tdot
+from ...util.linalg import pdinv, dpotrs, tdot
+import numpy as np
 log_2_pi = np.log(2*np.pi)
 
-class DTCVar(object):
+class VarDTC(object):
     """
     An object for inference when the likelihood is Gaussian, but we want to do sparse inference.
 
@@ -19,7 +20,7 @@ class DTCVar(object):
         self._YYTfactor_cache = caching.cache()
         self.const_jitter = 1e-6
 
-     def get_YYTfactor(self, Y):
+    def get_YYTfactor(self, Y):
         """
         find a matrix L which satisfies LLT = YYT.