From 9f322f7aac5ca6b5025d076a4d775cea658a9e26 Mon Sep 17 00:00:00 2001
From: Farid Mohammadi <farid.mohammadi@iws.uni-stuttgart.de>
Date: Wed, 16 Feb 2022 11:51:58 +0100
Subject: [PATCH] [bayes_inference] update paramter names from ExpDesign
object.
---
.../bayes_inference/bayes_inference.py | 4 ++--
src/bayesvalidrox/bayes_inference/mcmc.py | 20 +++++++++----------
2 files changed, 12 insertions(+), 12 deletions(-)
diff --git a/src/bayesvalidrox/bayes_inference/bayes_inference.py b/src/bayesvalidrox/bayes_inference/bayes_inference.py
index 85f68c4aa..3ba266cb9 100644
--- a/src/bayesvalidrox/bayes_inference/bayes_inference.py
+++ b/src/bayesvalidrox/bayes_inference/bayes_inference.py
@@ -1034,10 +1034,10 @@ class BayesInference:
axes = np.array(figPosterior.axes).reshape((len(par_names), len(par_names)))
for yi in range(len(par_names)):
ax = axes[yi, yi]
- ax.set_xlim(PCEModel.BoundTuples[yi])
+ ax.set_xlim(PCEModel.bound_tuples[yi])
for xi in range(yi):
ax = axes[yi, xi]
- ax.set_xlim(PCEModel.BoundTuples[xi])
+ ax.set_xlim(PCEModel.bound_tuples[xi])
# Extract the axes
# axes = np.array(figPosterior.axes).reshape((NofPa, NofPa))
diff --git a/src/bayesvalidrox/bayes_inference/mcmc.py b/src/bayesvalidrox/bayes_inference/mcmc.py
index 1c918be1a..4edf8be72 100755
--- a/src/bayesvalidrox/bayes_inference/mcmc.py
+++ b/src/bayesvalidrox/bayes_inference/mcmc.py
@@ -95,16 +95,16 @@ class MCMC():
# Pick samples based on a uniform dist between min and max of
# each dim
initsamples = np.zeros((self.nwalkers, ndim))
- BoundTuples = []
+ bound_tuples = []
for idxDim in range(ndim):
lower = np.min(self.initsamples[:, idxDim])
upper = np.max(self.initsamples[:, idxDim])
- BoundTuples.append((lower, upper))
+ bound_tuples.append((lower, upper))
dist = st.uniform(loc=lower, scale=upper-lower)
initsamples[:, idxDim] = dist.rvs(size=self.nwalkers)
# Update lower and upper
- PCEModel.ExpDesign.BoundTuples = BoundTuples
+ PCEModel.ExpDesign.bound_tuples = bound_tuples
# Check if sigma^2 needs to be inferred
if Discrepancy.optSigma != 'B':
@@ -116,8 +116,8 @@ class MCMC():
# Update ndim
ndim = initsamples.shape[1]
- # Update BoundTuples
- PCEModel.ExpDesign.BoundTuples += Discrepancy.ExpDesign.BoundTuples
+ # Update bound_tuples
+ PCEModel.ExpDesign.bound_tuples += Discrepancy.ExpDesign.bound_tuples
print("\n>>>> Bayesian inference with MCMC for "
f"{self.BayesOpts.Name} started. <<<<<<")
@@ -278,7 +278,7 @@ class MCMC():
for i in range(len(Discrepancy.InputDisc.Marginals)):
par_names.append(Discrepancy.InputDisc.Marginals[i].Name)
- params_range = PCEModel.ExpDesign.BoundTuples
+ params_range = PCEModel.ExpDesign.bound_tuples
# Plot traces
if self.verbose and self.nsteps < 10000:
@@ -358,9 +358,9 @@ class MCMC():
PCEModel = self.BayesOpts.PCEModel
Discrepancy = self.BayesOpts.Discrepancy
- nSigma2 = len(Discrepancy.ExpDesign.BoundTuples) if Discrepancy.optSigma != 'B' else -len(theta)
+ nSigma2 = len(Discrepancy.ExpDesign.bound_tuples) if Discrepancy.optSigma != 'B' else -len(theta)
priorDist = PCEModel.ExpDesign.priorSpace
- params_range = PCEModel.ExpDesign.BoundTuples
+ params_range = PCEModel.ExpDesign.bound_tuples
ndimTheta = theta.ndim
theta = theta if ndimTheta != 1 else theta.reshape((1,-1))
nsamples = theta.shape[0]
@@ -379,7 +379,7 @@ class MCMC():
# Check if bias term needs to be inferred
if Discrepancy.optSigma != 'B':
if self.check_ranges(theta[i,-nSigma2:],
- Discrepancy.ExpDesign.BoundTuples):
+ Discrepancy.ExpDesign.bound_tuples):
if all('unif' in Discrepancy.ExpDesign.InputObj.Marginals[i].DistType for i in \
range(Discrepancy.ExpDesign.ndim)):
logprior[i] = 0.0
@@ -397,7 +397,7 @@ class MCMC():
PCEModel = BayesOpts.PCEModel
Discrepancy = self.BayesOpts.Discrepancy
if Discrepancy.optSigma != 'B':
- nSigma2 = len(Discrepancy.ExpDesign.BoundTuples)
+ nSigma2 = len(Discrepancy.ExpDesign.bound_tuples)
else:
nSigma2 = -len(theta)
# Check if bias term needs to be inferred
--
GitLab