Wiki

Clone wiki

vide_public / Installation

Prerequisites

VIDE requires the following libraries installed prior to installation:

  • A C/C++ compiler
  • CMake
  • Python 2.7
  • NumPy (>= 1.6.1)
  • Matplotlib (>= 1.1.1rc)
  • PyFITS (>=3.3.0)

Compiling

To install, in the main VIDE directory execute

ccmake ./
make

All necessary additional libraries will be downloaded and installed, unless the user sets the relevant INTERNAL_NNN to OFF. These libraries include:

  • Python:

    • netcdf4
    • cython
    • healpy
    • setuptools
    • scipy
    • kdtree_scipy
  • C/C++:

    • GSL
    • boost
    • netcdf
    • HDF5
    • gengetopt
    • qhull

If there is a problem downloading a package (this often occurs if libcurl is not installed with SSL support), then you will have to download the indicated problem packages separately and point CMake to the location. In the CMake console:

t to toggle advanced mode
search for missing library
edit path to point to local download
configure, generate, and quit

If you already have any of these libraries installed and included in your PATH or PYTHONPATH, then in the CMake console search for INTERNAL_* and choose the libraries for which you want to disable internal loading.

Run make to finish the compilation.

If you are installing on a Mac, good luck!

Updating your path

While optional, adding the location of path/to/vide_public/ext_build/python/lib/python2.7/site-packages/ to your PYTHONPATH (e.g., in your .bash_profile) makes execution of the pipeline easier, and makes the VIDE analysis tools available to you outside of the VIDE directory.

If you do not add this directory to your PYTHONPATH, then for all the scripts you need to run with the helper ./run_python.sh.

This wiki will assume that you have updated your PYTHONPATH.

Updated