This library is meant to provide a user-friendly implementation of the Consistent Bayesian framework for solving Stochastic Inverse Problems.
Easiest way: execute ./fresh_install.sh
This will create a new python kernel for you and link it to jupyter notebook.
It will be called test_py_env. This file is a wrapper around the instructions that follow. You can activate this and have access to all the cbayes packages and dependencies with source activate test_py_env
If you create a fresh Anaconda environment with
conda create -n py3b python=3.6 -y; source activate YourEnvironmentName
or simply using your existing python distribution (run which python), you can
execute python setup.py install in the main ConsistentBayes directory to install all dependencies.
You will have to link your kernel to jupyter if you want to use the notebooks.
Additionally, if you want to use widgets (interactive workbooks with sliders), you can check the series of commands listed in the file fresh_install.sh to ensure you have done so correctly.
You can verify that everything is working as expected by executing the unit tests by running nosetests from the parent directory.
Please make sure to run nosetests before making any commits if you plan to contribute pull requests.
It is suggested that you start with the examples/CBayes.ipynb file.
Then examples/CBPaper_Examples walks through Example 6.1 and 6.2 from the seminal paper on this work, cited below.
These notebook walks you through the method and several example files equipped with rich interactive multi-dimensional visualizations.
A non-interactive python script that carries out the same computations can be found in examples/consistentbayes_example.py
Some notebooks that have PDE solvers use the python package progressbar, which you can install with pip (it is not a main dependency so it is not installed by setup.py, or simply remove the use of bar in the for-loop of the PDE solver code in the notebook.
See the examples/ directory for both script and jupyter-notebook files demonstrating how to solve stochastic inverse problems in the Consistent Bayesian framework, which was developed in tandem by:
- Dr. Troy Butler, CU Denver Dept. of Mathematics & Statistical Sciences
- Dr. Timothy Wildey, Sandia National Laboratories
- Dr. Scott Walsh, and
- Michael Pilosov, MS, CU Denver Dept. of Mathematics & Statistical Sciences
You can find general overview in the explanations presented in the jupyter notebooks.
For more detail, please see this draft on arxiv pending publication
Author: Michael Pilosov
Updated: 2/27/2018
ConsistentBayes.com
This software was released under an MIT open-source license. It is provided AS IS and with NO WARRANTY OR GAURANTEE.
Please see LICENSE for more information.