Merge pull request #359 from SheffieldML/devel

Minor patch

GPy/__version__.py

@@ -1 +1 @@
-__version__ = "1.0.1"
+__version__ = "1.0.2"

GPy/core/gp.py

@@ -437,15 +437,22 @@ class GP(Model):
         warnings.warn("Wrong naming, use predict_wishart_embedding instead. Will be removed in future versions!", DeprecationWarning)
         return self.predict_wishart_embedding(Xnew, kern, mean, covariance)
 
-    def predict_magnification(self, Xnew, kern=None, mean=True, covariance=True):
+    def predict_magnification(self, Xnew, kern=None, mean=True, covariance=True, dimensions=None):
         """
         Predict the magnification factor as
 
         sqrt(det(G))
 
-        for each point N in Xnew
+        for each point N in Xnew.
+
+        :param bool mean: whether to include the mean of the wishart embedding.
+        :param bool covariance: whether to include the covariance of the wishart embedding.
+        :param array-like dimensions: which dimensions of the input space to use [defaults to self.get_most_significant_input_dimensions()[:2]]
         """
         G = self.predict_wishard_embedding(Xnew, kern, mean, covariance)
+        if dimensions is None:
+            dimensions = self.get_most_significant_input_dimensions()[:2]
+        G = G[:, dimensions][:,:,dimensions]
         from ..util.linalg import jitchol
         mag = np.empty(Xnew.shape[0])
         for n in range(Xnew.shape[0]):

GPy/models/state_space_model.py

@@ -15,19 +15,10 @@
 #
 
 import numpy as np
-from scipy import linalg
 from scipy import stats
-from ..core import Model
-from .. import kern
-#from GPy.plotting.matplot_dep.models_plots import gpplot
-#from GPy.plotting.matplot_dep.base_plots import x_frame1D
-#from GPy.plotting.matplot_dep import Tango
-#import pylab as pb
-from GPy.core.parameterization.param import Param
-
-import GPy
 from .. import likelihoods
-
+#from . import state_space_setup as ss_setup
+from ..core import Model
 from . import state_space_main as ssm
 from . import state_space_setup as ss_setup
 
@@ -37,12 +28,12 @@ class StateSpace(Model):
 
         if len(X.shape) == 1:
             X = np.atleast_2d(X).T
-        self.num_data, input_dim = X.shape
+        self.num_data, self.input_dim = X.shape
 
         if len(Y.shape) == 1:
             Y = np.atleast_2d(Y).T
 
-        assert input_dim==1, "State space methods are only for 1D data"
+        assert self.input_dim==1, "State space methods are only for 1D data"
 
         if len(Y.shape)==2:
             num_data_Y, self.output_dim = Y.shape
@@ -168,7 +159,7 @@ class StateSpace(Model):
     def log_likelihood(self):
         return self._log_marginal_likelihood
 
-    def _raw_predict(self, Xnew=None, Ynew=None, filteronly=False):
+    def _raw_predict(self, Xnew=None, Ynew=None, filteronly=False, **kw):
         """
         Performs the actual prediction for new X points.
         Inner function. It is called only from inside this class.
@@ -270,22 +261,23 @@ class StateSpace(Model):
         # Return the posterior of the state
         return (m, V)
 
-    def predict(self, Xnew=None, filteronly=False):
+    def predict(self, Xnew=None, filteronly=False, include_likelihood=True, **kw):
 
         # Run the Kalman filter to get the state
         (m, V) = self._raw_predict(Xnew,filteronly=filteronly)
 
         # Add the noise variance to the state variance
-        V += float(self.Gaussian_noise.variance)
+        if include_likelihood:
+            V += float(self.likelihood.variance)
 
         # Lower and upper bounds
-        lower = m - 2*np.sqrt(V)
-        upper = m + 2*np.sqrt(V)
+        #lower = m - 2*np.sqrt(V)
+        #upper = m + 2*np.sqrt(V)
 
         # Return mean and variance
-        return (m, V, lower, upper)
+        return m, V
 
-    def predict_quantiles(self, Xnew=None, quantiles=(2.5, 97.5)):
+    def predict_quantiles(self, Xnew=None, quantiles=(2.5, 97.5), **kw):
         mu, var = self._raw_predict(Xnew)
         #import pdb; pdb.set_trace()
         return  [stats.norm.ppf(q/100.)*np.sqrt(var + float(self.Gaussian_noise.variance)) + mu for q in quantiles]

GPy/plotting/__init__.py

@@ -52,6 +52,17 @@ def inject_plotting():
         GP.plot_f = gpy_plot.gp_plots.plot_f
         GP.plot_magnification = gpy_plot.latent_plots.plot_magnification
 
+        from ..models import StateSpace
+        StateSpace.plot_data = gpy_plot.data_plots.plot_data
+        StateSpace.plot_data_error = gpy_plot.data_plots.plot_data_error
+        StateSpace.plot_errorbars_trainset = gpy_plot.data_plots.plot_errorbars_trainset
+        StateSpace.plot_mean = gpy_plot.gp_plots.plot_mean
+        StateSpace.plot_confidence = gpy_plot.gp_plots.plot_confidence
+        StateSpace.plot_density = gpy_plot.gp_plots.plot_density
+        StateSpace.plot_samples = gpy_plot.gp_plots.plot_samples
+        StateSpace.plot = gpy_plot.gp_plots.plot
+        StateSpace.plot_f = gpy_plot.gp_plots.plot_f
+
         from ..core import SparseGP
         SparseGP.plot_inducing = gpy_plot.data_plots.plot_inducing
 

GPy/plotting/gpy_plot/plot_util.py

@@ -190,6 +190,7 @@ def scatter_label_generator(labels, X, visible_dims, marker=None):
             x = X[index, input_1]
             y = X[index, input_2]
             z = X[index, input_3]
+
         yield x, y, z, this_label, index, m
 
 def subsample_X(X, labels, num_samples=1000):

GPy/plotting/plotly_dep/plot_definitions.py

@@ -131,14 +131,15 @@ class PlotlyPlots(AbstractPlottingLibrary):
             #not matplotlib marker
             pass
         marker_kwargs = marker_kwargs or {}
-        marker_kwargs.setdefault('symbol', marker)
+        if 'symbol' not in marker_kwargs:
+            marker_kwargs['symbol'] = marker
         if Z is not None:
             return Scatter3d(x=X, y=Y, z=Z, mode='markers',
                              showlegend=label is not None,
                              marker=Marker(color=color, colorscale=cmap, **marker_kwargs),
                              name=label, **kwargs)
         return Scatter(x=X, y=Y, mode='markers', showlegend=label is not None,
-                       marker=Marker(color=color, colorscale=cmap, **marker_kwargs or {}),
+                       marker=Marker(color=color, colorscale=cmap, **marker_kwargs),
                        name=label, **kwargs)
 
     def plot(self, ax, X, Y, Z=None, color=None, label=None, line_kwargs=None, **kwargs):

GPy/testing/model_tests.py

@@ -730,6 +730,7 @@ class GradientTests(np.testing.TestCase):
         self.assertTrue( np.allclose(var1, var2) )
 
     def test_gp_VGPC(self):
+        np.random.seed(10)
         num_obs = 25
         X = np.random.randint(0, 140, num_obs)
         X = X[:, None]
@@ -737,6 +738,7 @@ class GradientTests(np.testing.TestCase):
         kern = GPy.kern.Bias(1) + GPy.kern.RBF(1)
         lik = GPy.likelihoods.Gaussian()
         m = GPy.models.GPVariationalGaussianApproximation(X, Y, kernel=kern, likelihood=lik)
+        m.randomize()
         self.assertTrue(m.checkgrad())
 
     def test_ssgplvm(self):
@@ -744,12 +746,14 @@ class GradientTests(np.testing.TestCase):
         from GPy.models import SSGPLVM
         from GPy.examples.dimensionality_reduction import _simulate_matern
 
+        np.random.seed(10)
         D1, D2, D3, N, num_inducing, Q = 13, 5, 8, 45, 3, 9
         _, _, Ylist = _simulate_matern(D1, D2, D3, N, num_inducing, False)
         Y = Ylist[0]
         k = kern.Linear(Q, ARD=True)  # + kern.white(Q, _np.exp(-2)) # + kern.bias(Q)
         # k = kern.RBF(Q, ARD=True, lengthscale=10.)
         m = SSGPLVM(Y, Q, init="rand", num_inducing=num_inducing, kernel=k, group_spike=True)
+        m.randomize()
         self.assertTrue(m.checkgrad())
 
 if __name__ == "__main__":

GPy/testing/plotting_tests.py

@@ -89,6 +89,9 @@ def _image_directories():
         cbook.mkdirs(result_dir)
     return baseline_dir, result_dir
 
+baseline_dir, result_dir = _image_directories()
+if not os.path.exists(baseline_dir):
+    raise SkipTest("Not installed from source, baseline not available. Install from source to test plotting")
 
 def _sequenceEqual(a, b):
     assert len(a) == len(b), "Sequences not same length"
@@ -99,7 +102,6 @@ def _notFound(path):
     raise IOError('File {} not in baseline')
 
 def _image_comparison(baseline_images, extensions=['pdf','svg','png'], tol=11):
-    baseline_dir, result_dir = _image_directories()
     for num, base in zip(plt.get_fignums(), baseline_images):
         for ext in extensions:
             fig = plt.figure(num)

MANIFEST.in

@@ -17,5 +17,5 @@ recursive-include GPy *.h
 recursive-include GPy *.pyx
 
 # Testing
-include GPy/testing/baseline/*.png
+#include GPy/testing/baseline/*.png
 #include GPy/testing/pickle_test.pickle

README.md

@@ -9,6 +9,8 @@ The Gaussian processes framework in Python.
 * Travis-CI [unit-tests](https://travis-ci.org/SheffieldML/GPy)
 * [![licence](https://img.shields.io/badge/licence-BSD-blue.svg)](http://opensource.org/licenses/BSD-3-Clause)
 
+[![develstat](https://travis-ci.org/SheffieldML/GPy.svg?branch=devel)](https://travis-ci.org/SheffieldML/GPy) [![covdevel](http://codecov.io/github/SheffieldML/GPy/coverage.svg?branch=devel)](http://codecov.io/github/SheffieldML/GPy?branch=devel) [![docdevel](https://readthedocs.org/projects/gpy/badge/?version=devel)](http://gpy.readthedocs.org/en/devel/) [![Research software impact](http://depsy.org/api/package/pypi/GPy/badge.svg)](http://depsy.org/package/python/GPy)
+
 ## Updated Structure
 
 We have pulled the core parameterization out of GPy. It is a package called [paramz](https://github.com/sods/paramz) and is the pure gradient based model optimization.
@@ -27,12 +29,6 @@ A warning: This usually works, but sometimes `distutils/setuptools` opens a
 whole can of worms here, specially when compiled extensions are involved.
 If that is the case, it is best to clean the repo and reinstall.
 
-## Continuous integration
-
-|      | Travis-CI | Codecov | RTFD |
-| ---: | :--: | :---: | :---: |
-| **devel:**  | [![develstat](https://travis-ci.org/SheffieldML/GPy.svg?branch=devel)](https://travis-ci.org/SheffieldML/GPy) | [![covdevel](http://codecov.io/github/SheffieldML/GPy/coverage.svg?branch=devel)](http://codecov.io/github/SheffieldML/GPy?branch=devel) | [![docdevel](https://readthedocs.org/projects/gpy/badge/?version=devel)](http://gpy.readthedocs.org/en/devel/) | 
-
 ## Supported Platforms:
 
 [<img src="https://www.python.org/static/community_logos/python-logo-generic.svg" height=40px>](https://www.python.org/)
@@ -40,7 +36,7 @@ If that is the case, it is best to clean the repo and reinstall.
 [<img src="https://upload.wikimedia.org/wikipedia/commons/8/8e/OS_X-Logo.svg" height=40px>](http://www.apple.com/osx/)
 [<img src="https://upload.wikimedia.org/wikipedia/commons/3/35/Tux.svg" height=40px>](https://en.wikipedia.org/wiki/List_of_Linux_distributions)
 
-Python 2.7, 3.3 and higher
+Python 2.7, 3.4 and higher
 
 ## Citation
 

setup.cfg

@@ -1,5 +1,5 @@
 [bumpversion]
-current_version = 1.0.1
+current_version = 1.0.2
 tag = False
 commit = True
 
@@ -11,3 +11,6 @@ universal = 1
 [upload_docs]
 upload-dir = doc/build/html
 
+[metadata]
+description-file = README.rst
+

setup.py

@@ -182,6 +182,8 @@ if not os.path.exists(user_file):
     if os.path.exists(old_user_file):
         # Move it to new location:
         print("GPy: Found old config file, moving to new location {}".format(user_file))
+        if not os.path.exists(os.path.dirname(user_file)):
+            os.makedirs(os.path.dirname(user_file))
         os.rename(old_user_file, user_file)
     else:
         # No config file exists, save informative stub to user config folder: