A new and (hopefully) correct implementation of TwinPeaks using the SwitchSampling idea.

(c) 2018 Brendon J. Brewer.

LICENSE: GNU General Public License version 3. See the LICENSE file for details.

For the C++ to compile, you'll need:

• yaml-cpp and its dependency boost

You can probably get these from your operating system's package manager. You'll also need the C++ header files, which are sometimes put into a separate package with the suffix -dev or -devel. For example, on Ubuntu, this will do the trick:

sudo apt-get install libboost-all-dev libyaml-cpp-dev

On the Python side, you'll need:

• Python 3 and some packages (numpy, matploblib, and pandas). Anaconda is a good Python distribution that includes these. Otherwise they might come with (or be installable through) Python 3 as provided by your OS.
• LaTeX and dvipng so that LaTeX can be used with matplotlib.

First, compile the C++:

make

Then install the python package:

cd python
python3 setup.py install

To run the demo example:

./main

Then wait a long time. At any time during the run (or after it's completed), you can execute

python3 showresults.py

to see the results.

The authors of this paper.