From ea3bfbb59715a4801f844295d53c63bda17f0e47 Mon Sep 17 00:00:00 2001
From: James Hensman <james.hensman@gmail.com>
Date: Tue, 18 Aug 2015 08:00:47 +0100
Subject: [PATCH] minor bugfixes in plotting: quantiles are now computed using
 predict_kw correctly

---
 GPy/core/gp.py                           | 6 ++++--
 GPy/models/gp_var_gauss.py               | 3 ++-
 GPy/plotting/matplot_dep/models_plots.py | 3 ++-
 GPy/testing/inference_tests.py           | 2 +-
 4 files changed, 9 insertions(+), 5 deletions(-)

diff --git a/GPy/core/gp.py b/GPy/core/gp.py
index 12fb3d27..aeee34ac 100644
--- a/GPy/core/gp.py
+++ b/GPy/core/gp.py
@@ -249,7 +249,7 @@ class GP(Model):
         mean, var = self.likelihood.predictive_values(mu, var, full_cov, Y_metadata=Y_metadata)
         return mean, var
 
-    def predict_quantiles(self, X, quantiles=(2.5, 97.5), Y_metadata=None):
+    def predict_quantiles(self, X, quantiles=(2.5, 97.5), Y_metadata=None, kern=None):
         """
         Get the predictive quantiles around the prediction at X
 
@@ -257,10 +257,12 @@ class GP(Model):
         :type X: np.ndarray (Xnew x self.input_dim)
         :param quantiles: tuple of quantiles, default is (2.5, 97.5) which is the 95% interval
         :type quantiles: tuple
+        :param kern: optional kernel to use for prediction
+        :type predict_kw: dict
         :returns: list of quantiles for each X and predictive quantiles for interval combination
         :rtype: [np.ndarray (Xnew x self.output_dim), np.ndarray (Xnew x self.output_dim)]
         """
-        m, v = self._raw_predict(X,  full_cov=False)
+        m, v = self._raw_predict(X,  full_cov=False, kern=kern)
         if self.normalizer is not None:
             m, v = self.normalizer.inverse_mean(m), self.normalizer.inverse_variance(v)
         return self.likelihood.predictive_quantiles(m, v, quantiles, Y_metadata=Y_metadata)
diff --git a/GPy/models/gp_var_gauss.py b/GPy/models/gp_var_gauss.py
index 729b6bb8..9f2229f0 100644
--- a/GPy/models/gp_var_gauss.py
+++ b/GPy/models/gp_var_gauss.py
@@ -60,7 +60,8 @@ class GPVariationalGaussianApproximation(Model):
         var = np.diag(Sigma).reshape(-1,1)
 
         F, dF_dm, dF_dv, dF_dthetaL = self.likelihood.variational_expectations(self.Y, m, var, Y_metadata=self.Y_metadata)
-        self.likelihood.gradient = dF_dthetaL.sum(1).sum(1)
+        if dF_dthetaL is not None:
+            self.likelihood.gradient = dF_dthetaL.sum(1).sum(1)
         dF_da = np.dot(K, dF_dm)
         SigmaB = Sigma*self.beta
         dF_db = -np.diag(Sigma.dot(np.diag(dF_dv.flatten())).dot(SigmaB))*2
diff --git a/GPy/plotting/matplot_dep/models_plots.py b/GPy/plotting/matplot_dep/models_plots.py
index 78c80cb5..d0f6c952 100644
--- a/GPy/plotting/matplot_dep/models_plots.py
+++ b/GPy/plotting/matplot_dep/models_plots.py
@@ -110,7 +110,8 @@ def plot_fit(model, plot_limits=None, which_data_rows='all',
                 else:
                     Y_metadata['output_index'] = extra_data
             m, v = model.predict(Xgrid, full_cov=False, Y_metadata=Y_metadata, **predict_kw)
-            lower, upper = model.predict_quantiles(Xgrid, Y_metadata=Y_metadata)
+            fmu, fv = model._raw_predict(Xgrid, full_cov=False, **predict_kw)
+            lower, upper = model.likelihood.predictive_quantiles(fmu, fv, (2.5, 97.5), Y_metadata=Y_metadata)
 
 
         for d in which_data_ycols:
diff --git a/GPy/testing/inference_tests.py b/GPy/testing/inference_tests.py
index cd85235d..c1fce8b9 100644
--- a/GPy/testing/inference_tests.py
+++ b/GPy/testing/inference_tests.py
@@ -8,7 +8,7 @@ The test cases for various inference algorithms
 import unittest, itertools
 import numpy as np
 import GPy
-
+#np.seterr(invalid='raise')
 
 class InferenceXTestCase(unittest.TestCase):