install-manual-detailed.md

!! This guide is not finished, and might contain innacuracies. Please report any mistakes/bugs by creating an issue. This is a WIP (Work In Progress) !!

Detailed Installation Manual

Installing Python packages with compiled extensions can be a challenge especially on Windows systems. However, when using Miniconda things can be simplified to a great extent as this manual hopefully will show you.

The this approach will require you to use the command line, but it is as easy as copy-pasting them from this page straight into your command prompt.

Installation instructions follow in alphabetical orderby platorm.

Linux

• Basic dependencies:

python (3.5 recommended) git gcc gcc-fortran (gfortran)

• Use either your system package manager, pip + virtualenv, or Anaconda to install the following python dependencies:

numpy, cython, scipy, pandas, matplotlib, pyscaffold, future, nose, sphinx, xlrd, (py)tables, h5py, pytest, pytest-cov, setuptools_scm, setuptools

Note that often the pytables packages is called python-tables instead of python-pytables.

Dependencies on Mac

People who now how to handle Python on the Mac side are kindly requested to complete this guide :-)

Dependencies on Windows

A Python installation with compilers is required. If you already have this, or know how set up such an environment, you skip to here.

Microsft Visual Studio 2010 Compiler

wetb contains extensions that need to be compiled. On Windows things are complicated because you need to use the same compiler as the one used for Python. This means that for compiling extensions on:

• Python 2.7 you need Microsoft Visual C++ Compiler for Python 2.7, or the direct link.
• Python 3.4 you need MS Visual Studio 2010
• Python 3.5 (and later) you need MS Visual Studio 2015
• You can install Microsoft Visual C++ Compiler for Python 2.7 alongside MS Visual Studio 2010, but you can not install Visual Studio 2010 and 2015 in parallel.

You can find more background information and installation instructions here, here, or here.

Command line

This guide will use the command line (aka command prompt) frequently. You can launch a Windows terminal as follows: press Start> and type "cmd" + . A link to the command prompt should be visible now.

In case you want an alternative, more capable windows terminal, you could consider using ConEmu (this is optional).

ConEmu-Maximus5 is a Windows console emulator with tabs, which presents multiple consoles and simple GUI applications as one customizable GUI window with various features.

Git

• Download and install Git version control system for Windows 64-bit here. Only select the Windows Portable option if you know what you are doing or if you do not have administrative rights on your computer.

• Git comes with a simple GUI, but there are more and different options available if you are not happy with it, see here.

• If you would like to use a GUI for git, we recommend you to use tortoisegit

Recommended python distribution: Anaconda

Installing Anaconda, activate root environment

• Anaconda is a professional grade, full blown scientific Python distribution.

• Download and install Anaconda from https://www.continuum.io/downloads.

Note: the choice of Anaconda for Python 2.7 or Python 3.5 only affects the root environment. You can always create additional environments using other Python versions, see below.

• Update the root Anaconda environment (type in a terminal):

conda update --all

• Activate the Anaconda root environment in a terminal as follows:

activate

and your terminal will do something like:

C:\Users\> activate
[Anaconda3] C:\Users\>

note that the name of the environment is now a prefix before the current path.

use deactivate to deactivate the environment.

Optionally, create other independent Anaconda environments

• By using environments you can manage different Python installations with different versions on your system. Creating environments is as easy as:

conda create -n py27 python=2.7
conda create -n py34 python=3.4
conda create -n py35 python=3.5

• These environments can be activated as follows:

activate py27
activate py34
activate py35

use deactivate to deactivate the environment.

Install dependencies with conda and pip

• Install the necessary Python dependencies using the conda package manager:

conda install setuptools_scm future h5py pytables pytest nose sphinx
conda install scipy pandas matplotlib cython xlrd coverage xlwt openpyxl

• Not all packages are available in the conda repositories, but they can be easily installed with pip:

pip install pyscaffold pytest-cov --no-deps

And Finally: install wetb

git clone https://gitlab.windenergy.dtu.dk/toolbox/WindEnergyToolbox.git
cd WindEnergyToolbox
pip install -e . --no-deps

Note that pip install -e . --no-deps will install wetb in the source directory. This works best if you are also developing and regularly updating this package.

You can run the tests cases from the source root directory:

python setup.py test