Merge branch 'master' into parallel_like1

dpmm/dpmm.py

@@ -125,13 +125,13 @@ def update_latent_data(self):
         # Pseudo-marginal samples or (TBD) means and Gaussian errors.
         if isinstance(self._D, PseudoMarginalData):
             for i, ph in enumerate(self.phi):
-                index = np.nonzero(self.label == i)
+                index = np.nonzero(self.label == i)[0]
                 data = self._D[index]  # a PseudoMarginalData instance
                 # calculate weights for selecting a representative sample
                 ps = self.prior.like1(self.manip(data.data), ph) / data.interim_prior
                 ps /= np.sum(ps, axis=1)[:, np.newaxis]  # think this line can go.
                 for j, p in enumerate(ps):
-                    self.D[index[0][j]] = data.data[j, pick_discrete(p)]
+                    self.D[index[j]] = data.data[j, pick_discrete(p)]
             # Need to update the r_i probabilities too since self.D changed.
             # self.r_i = self.alpha * self.prior.pred(self.mD)
             self.p[:, 0] = self.alpha * self.prior.pred(self.mD)

dpmm/shear.py

@@ -126,6 +126,8 @@ def __init__(self, g):
         # g should be a length-2 array representing the real and imag parts of the complex
         # representation of the reduced shear.
         self.g = g
+        self.Nproposals = 0
+        self.Nacceptances = 0
 
     def init(self, D):
         """A quick and dirty estimate of g is just the average over D."""
@@ -150,7 +152,7 @@ def update(self, D, phi, label, prior):
         # Whoops!  The weak shear limit proposal doesn't lead to *any* acceptances when ngal is
         # large.  Need to try something more clever.
         # prop_g = draw_g_2d_weak_shear(D, phi, label)
-        prop_g = np.random.multivariate_normal(mean=self.g, cov=np.eye(2)*0.0005**2)
+        prop_g = np.random.multivariate_normal(mean=self.g, cov=np.eye(2)*0.003**2)
 
         current_e_int = unshear(D, self.g)
         prop_e_int = unshear(D, prop_g)
@@ -162,7 +164,10 @@ def update(self, D, phi, label, prior):
             prop_lnlike += prior.lnlikelihood(prop_e_int[index], ph)
         if prop_lnlike > current_lnlike:
             self.g = prop_g
+            self.Nacceptances += 1
         else:
             u = np.random.uniform()
             if u < np.exp(prop_lnlike - current_lnlike):
                 self.g = prop_g
+                self.Nacceptances += 1
+        self.Nproposals += 1