Merge pull request #142 from lcgraham/master

Update example documentation files

bet/calculateP/calculateP.py

@@ -6,9 +6,6 @@
 
 * :mod:`~bet.calculateP.prob_emulated` provides a skeleton class and calculates
     the probability for a set of emulation points.
-* :mod:`~bet.calculateP.calculateP.prob_samples_ex` calculates the exact
-    volumes of the interior voronoi cells and estimates the volumes of the
-    exterior voronoi cells by using a set of bounding points
 * :mod:`~bet.calculateP.calculateP.prob_samples_mc` estimates the volumes of
     the voronoi cells using MC integration
 """

bet/calculateP/voronoiHistogram.py

@@ -47,8 +47,8 @@ def center_and_layer1_points_binsize(center_pts_per_edge, center, r_size,
     rect_width = r_size*np.ones(sur_domain[:, 0].shape)
     rect_domain = np.column_stack([center - .5*rect_width,
         center + .5*rect_width])
-    if np.all(np.greater(r_size, rect_width)):
-        msg = "The hyperrectangle defined by this size is larger than the "
+    if np.any(np.greater(r_size, rect_width)):
+        msg = "The hyperrectangle defined by this size extends outside the "
         msg += "original domain."
         print msg
 
@@ -100,7 +100,7 @@ def center_and_layer1_points(center_pts_per_edge, center, r_ratio, sur_domain):
     """
     if np.all(np.greater(r_ratio, 1)):
         msg = "The hyperrectangle defined by this ratio is larger than the"
-        msg += "original domain."
+        msg += " original domain."
         print msg
 
     # determine r_size from the width of the surrounding domain
@@ -140,11 +140,11 @@ def edges_regular(center_pts_per_edge, rect_domain, sur_domain):
     """
     if np.any(np.greater_equal(sur_domain[:, 0], rect_domain[:, 0])):
         msg = "The hyperrectangle defined by this size is larger than the"
-        msg += "original domain."
+        msg += " original domain."
         print msg
     elif np.any(np.less_equal(sur_domain[:, 1], rect_domain[:, 1])):
         msg = "The hyperrectangle defined by this size is larger than the"
-        msg += "original domain."
+        msg += " original domain."
         print msg
 
     rect_edges = list()

bet/postProcess/plotP.py

@@ -6,10 +6,11 @@
 
 from bet.Comm import comm, MPI 
 import matplotlib.pyplot as plt
+#plt.rc('text', usetex=True)
+#plt.rc('font', family='serif')
 import numpy as np
 import copy, math
 
-
 def calculate_1D_marginal_probs(P_samples, samples, lam_domain, nbins=20):
 
     """
@@ -45,7 +46,8 @@ def calculate_1D_marginal_probs(P_samples, samples, lam_domain, nbins=20):
     # Calculate marginals
     marginals = {}
     for i in range(num_dim):
-        [marg, _] = np.histogram(samples[:,i], bins=bins[i], weights=P_samples)
+        [marg, _] = np.histogram(samples[:, i], bins=bins[i], 
+                                 weights=P_samples)
         marg_temp = np.copy(marg)
         comm.Allreduce([marg, MPI.DOUBLE], [marg_temp, MPI.DOUBLE], op=MPI.SUM)
         marginals[i] = marg_temp
@@ -88,8 +90,9 @@ def calculate_2D_marginal_probs(P_samples, samples, lam_domain, nbins=20):
     marginals = {}
     for i in range(num_dim):
         for j in range(i+1, num_dim):
-            (marg, _) = np.histogramdd(samples[:,[i,j]], bins = [bins[i], bins[j]], weights = P_samples)
-            marg=np.ascontiguousarray(marg)
+            (marg, _) = np.histogramdd(samples[:, [i, j]], bins=[bins[i],
+                                       bins[j]], weights=P_samples) 
+            marg = np.ascontiguousarray(marg)
             marg_temp = np.copy(marg)
             comm.Allreduce([marg, MPI.DOUBLE], [marg_temp, MPI.DOUBLE],
                     op=MPI.SUM) 
@@ -184,12 +187,12 @@ def plot_2D_marginal_probs(marginals, bins, lam_domain,
             ax = fig.add_subplot(111)
             boxSize = (bins[i][1]-bins[i][0])*(bins[j][1]-bins[j][0])
             quadmesh = ax.imshow(marginals[(i, j)].transpose()/boxSize,
-                    interpolation='bicubic', cmap=cm.jet, 
+                    interpolation='bicubic', cmap=cm.CMRmap_r, 
                     extent=[lam_domain[i][0], lam_domain[i][1],
                     lam_domain[j][0], lam_domain[j][1]], origin='lower',
                     vmax=marginals[(i, j)].max()/boxSize, vmin=0, aspect='auto')
             if type(lam_ref) != type(None):
-                ax.plot(lam_ref[i], lam_ref[j], 'ko', markersize=10)
+                ax.plot(lam_ref[i], lam_ref[j], 'wo', markersize=10)
             if lambda_label == None:
                 label1 = r'$\lambda_{' + str(i+1) + '}$'
                 label2 = r'$\lambda_{' + str(j+1) + '}$'
@@ -203,7 +206,8 @@ def plot_2D_marginal_probs(marginals, bins, lam_domain,
             fig.colorbar(quadmesh, ax=ax, label=label_cbar)
             plt.axis([lam_domain[i][0], lam_domain[i][1], lam_domain[j][0],
                 lam_domain[j][1]]) 
-            fig.savefig(filename + "_2D_" + str(i) + "_" + str(j) + ".eps", transparent=True)
+            fig.savefig(filename + "_2D_" + str(i) + "_" + str(j) + ".eps", 
+                        transparent=True)
             if interactive:
                 plt.show()
             else:
@@ -226,7 +230,8 @@ def plot_2D_marginal_probs(marginals, bins, lam_domain,
                 ax.set_zlabel(r'$P$')
                 plt.backgroundcolor = 'w'
                 fig.colorbar(surf, shrink=0.5, aspect=5, label=r'$P$')
-                fig.savefig(filename + "_surf_"+str(i)+"_"+str(j)+".eps", transparent=True)
+                fig.savefig(filename + "_surf_"+str(i)+"_"+str(j)+".eps",
+                            transparent=True) 
                 if interactive:
                     plt.show()
                 else:

bet/postProcess/postTools.py

@@ -75,7 +75,7 @@ def sample_highest_prob(top_percentile, P_samples, samples, lam_vol=None,
                 lam_vol, data)
 
     P_sum = np.cumsum(P_samples)
-    num_samples = np.sum(P_sum <= top_percentile)
+    num_samples = np.sum(np.logical_and(0.0 < P_sum, P_sum <= top_percentile))
     P_samples = P_samples[0:num_samples]
     samples = samples[0:num_samples, :]
     if type(lam_vol) != type(None):
@@ -164,7 +164,8 @@ def save_parallel_probs_mat(P_samples, samples, file_prefix, compress=False):
     """
     file_dict = {"P_samples": P_samples,
                "samples": samples}
-    sio.savemat(file_prefix + str(comm.rank), file_dict, do_compression=compress)
+    sio.savemat(file_prefix + str(comm.rank), file_dict,
+                do_compression=compress)
 
 def collect_parallel_probs_mat(file_prefix, num_files, save=False,
        compress=False):

doc/examples/example_rst_files/Q_2D.rst

@@ -20,6 +20,16 @@ defined and the approximation of the volume of :math:`\mathcal{V}_j`. See
 this example is essentially the same as :ref:`q1D` we will only highlight the
 differences between the two.
 
+
+.. note::
+
+    * In line 36 reduce number of emulated points to 1e4 so that it will run
+        quickly.
+    * reduce number of ``ref_nums`` to a single number so that we only
+        examine a single set of QoI
+    * add plotting of marginals using
+        :mod:`bet.postProcess.plotP.plot_2D_marginal_probs`
+
 Define the filename to save :math:`\hat{\rho}_{\Lambda, j}` to::
 
         filename = 'P_q'+str(station_nums[0]+1)+'_q'+str(station_nums[1]+1)

doc/examples/example_rst_files/fromfile2D.rst

@@ -14,7 +14,13 @@ We will walk through the following `example
 <https://github.com/UT-CHG/BET/blob/master/examples/fromFile_ADCIRCMap/fromFile2D.py>`_ that uses a linear interpolant of
 the QoI map :math:`Q(\lambda) = (q_1(\lambda), q_6(\lambda))` for a
 2-dimensional data space. The parameter space in this example is also
-2-dimensional. 
+2-dimensional.
+
+.. note::
+
+    * In the lines 56, 57 change chain length and num chains to
+      reduce the total number of forward solves.
+    * Saves to ``sandbox2d.mat``.
 
 The modules required by this example are::
 

doc/examples/example_rst_files/linearMapUniformSampling.rst

@@ -5,7 +5,7 @@
 Example: Linear Map with Uniform Sampling
 ===========================
 
-We will wailk through the following `example
+We will walk through the following `example
 <https://github.com/UT-CHG/BET/blob/master/examples/linearMap/linearMapUniformSampling.py>`_. This example
 generates uniform samples in a 3D input parameter space and evaluates a 
 linear map to a 2D data space. The stochastic inverse problem solved here

doc/examples/example_rst_files/nonlinearMapUniformSampling.rst

@@ -4,7 +4,7 @@
 Example: Nonlinear Map with Uniform Sampling
 ============================================
 
-We will wailk through the following `example
+We will walk through the following `example
 <https://github.com/UT-CHG/BET/blob/master/examples/nonlinearMap/nonlinearMapUniformSampling.py>`_. 
 This example generates uniform samples in a 2D input parameter space 
 and evaluates a nonlinear map to either a 1D or 2D output data space.

doc/examples/example_rst_files/validation.rst

@@ -4,7 +4,7 @@
 Example: Validation
 ============================================
 
-We will wailk through the following `example
+We will walk through the following `example
 <https://github.com/UT-CHG/BET/blob/master/examples/validationExample/linearMap.py>`_. 
 This 2D linear example shows that geometrically distinct QoI can
 recreate a probability measure on the input parameter space

doc/examples/overview.rst → doc/examples/examples_overview.rst

@@ -1,3 +1,4 @@
+.. _examples:
 
 =======================================
 Some References and Examples

setup.py

@@ -1,6 +1,6 @@
 #!/usr/bin/env python
 
-# Copyright (C) 2014-2015 Lindley Graham and Steven Mattis
+# Copyright (C) 2014-2015 BET Development Team
 
 '''
 The python script for building the BET package and subpackages.
@@ -11,7 +11,7 @@
     from distutils.core import setup
 
 setup(name='bet',
-      version='0.2.0',
+      version='1.0.0',
       description='Butler, Estep, Tavener method',
       author='Steven Mattis',
       author_email='[email protected]',