PyDakota

Python wrapper of Dakota surrogate and linear algebra utilities.

The swig_examples directory contains a number of simple examples of how to use
swig

//----------------//
// SYSTEM INSTALL //
//----------------//
The following assumes C++ libraries have already been built
To Install run to site-packages
python setup.py install

To uninstall from site packages
$ rm -r path/to/site-packges/PyDakota*

To get location of site-packages use
$ python -m site

//----------------//
// LOCAL INSTALL  //
//----------------//
The following assumes C++ libraries have already been built
python setup.py install -prefix=/path/to/local/install
To uninstall
$ rm -r /path/to/local/install/PyDakota*

//--------------//
// Dependencies //
//--------------//
swig, boost, numpy, scipy, matplotlib, cmake, gcc, g++ and gfortran

//-------------------//
// Conda Environment //
//-------------------//
Miniconda or anaconda can be used to install the dependencies
needed to build PyDakota. See https://conda.io/docs/test-drive.html
for a basic tutorial on conda.

First install conda then create a conda envrionment
$ conda create -n env_name
Activate the conda environment
$ source activate env_name
$ conda install numpy scipy matplotlib cmake boost=1.61.0 openblas=0.2.19 gcc swig
on OS X may need to set the following in your .bash_profile
DYLD_FALLBACK_LIBRARY_PATH=~/miniconda2/envs/env_name/lib
to put gfotran library on library path

//---------------------//
// Build C++ libraries //
//---------------------//
build c++ libraries locally
may need to provide more options to cmake
$ cd /path/to/pecos/
$ mkdir build && cd build
$ cmake  -D CMAKE_CXX_FLAGS:STRING="-fPIC" ..
$ make
