Skip to content

Getting started

Jump to bottom
Santiago Peñate Vera edited this page Jun 26, 2016 · 27 revisions

Installation

You'll need a version of Python 3 installed on your system with the following packages:

  • Numpy
  • Scipy (version 0.17 at least. The simplex solver linprog is provided since 2015)
  • Scikit-learn
  • NetworkX
  • Pandas (make sure you have the latest version)
  • xlwt
  • xlrd
  • xlsxwriter
  • Matplotlib
  • Numba
  • PyQt4

Windows:

The standard Anaconda distribution works out of the box. Otherwise install the listed packages in your Python distribution using pip.

Stand-alone Python + Gridcal installation

  • Download the anaconda installer from here. Pick the python 3.x version in 64 bits.
  • Install python in a separated folder. Do not mark the installation for all the users nor to add it to the PATH.
  • Copy Gridcal to the same folder where the Anaconda Python was installed.
  • Once installed, create a .bat file containing .\Anaconda\python .\GridCal\main_gui.py
  • Gridcal is executable by double-clicking the .bat file. Should work out of the box.

Linux:

I have experience in Ubuntu only, but you should be available to install everything from the package manager.

  • The stand-alone instructions are valid for Linux only that instead of a .bat file it must be a .sh file

Updating Numpy and Scipy

You will be asked for the BLAS/LAPACK libraries and for a FORTRAN compiler, for that do:

sudo apt-get install gfortran libblas-dev liblapack-dev

Then try upgrading the Python3 Numpy and Scipy packages

sudo pip3 install --upgrade numpy

sudo pip3 install --upgrade scipy

Somehow the Pip3 version installed from the Ubuntu repositories might not be able handle Numpy/Scipy updates. In that case we need to update the Pip3 command first: sudo easy_install3 -U pip3

Getting to install Numba

This instructions are to install the numba package in linux systems using the python 3 packages provided by the distribution. If you don't manage to install it, you can always install the Anaconda python distribution on your linux system and it'll be all there without having to install any new package.

Following this link the commands to install Numba on linux systems are:

Ubuntu 14.04

sudo apt-get install zlib1g zlib1g-dev libedit2 libedit-dev llvm-3.6 llvm-3.6-dev llvm-dev

pip3 install enum34 funcsigs

LLVM_CONFIG=/usr/bin/llvm-config-3.6 pip3 install llvmlite

LLVM_CONFIG=/usr/bin/llvm-config-3.6 pip3 install numba

Ubuntu 14.10

sudo apt-get install zlib1g zlib1g-dev libedit libedit-dev llvm-3.5 llvm-3.5-dev llvm-dev

pip3 install enum34 funcsigs

LLVM_CONFIG=/usr/bin/llvm-config-3.5 pip3 install llvmlite

LLVM_CONFIG=/usr/bin/llvm-config-3.5 pip3 install numba

Ubuntu 15.04

sudo apt-get install zlib1g zlib1g-dev libedit2 libedit-dev llvm-3.6 llvm-3.6-dev llvm-dev

pip3 install enum34 funcsigs

LLVM_CONFIG=/usr/bin/llvm-config-3.6 pip3 install llvmlite

LLVM_CONFIG=/usr/bin/llvm-config-3.6 pip3 install numba

Ubuntu 15.10 and 16.04 (LTS)

sudo apt-get install zlib1g zlib1g-dev libedit2 libedit-dev llvm-3.7 llvm-3.7-dev llvm-3.7-runtime llvm-3.7-tools sudo pip3 install enum34 funcsigs

Then we need to install llvmlite from the source :(

git clone https://github.com/numba/llvmlite

cd llvmlite/

sudo ch -c "LLVM_CONFIG=/usr/bin/config-3.7 python3 setup.py build"

sudo sh -c "LLVM_CONFIG=/usr/bin/llvm-config-3.7 python3 setup.py install"

Finally

sudo pip3 install numba

OSX:

Since I don't have an apple system I have no clue, but if you manage to install the Anaconda python distribution you should be good.

  • The stand-alone instructions should be valid for OSX, only that instead of a .bat file it must be a .sh file.
Clone this wiki locally