Setup with the existing python#
We will present how to compile the code, install and run the various scripts with setuptools.
Requirements#
Because the programs are written in Python, and some parts of the library in C++, you must have Python 3, at least Python version 3.10, a C++ compiler and cmake installed on your system to build the library.
Note
The C++ compiler must support the ISO C++ 2014 standard
The compiling C++ requires the following development library:
Eigen3 can utilize optimized libraries such as Intel Math Kernel Library (MKL), OpenBlas, or Accelerate on MacOSX to enhance computation speed. These libraries replace the default BLAS implementation provided by Eigen3.
You can install these packages on Ubuntu by typing the following command:
sudo apt-add-repository multiverse && sudo apt-get update # MKL is in the multiverse repo.
sudo apt-get install g++ cmake libeigen3-dev libmkl-dev libboost-dev
You need, also, to install Python libraries before configuring and installing this software:
You can install these packages on Ubuntu by typing the following command:
sudo apt-get install python3-numpy python3-netcdf4
Build#
Once you have satisfied the requirements detailed above, to build the library,
type the command python3 setup.py build
at the root of the project.
You can specify, among other things, the following options to build-ext command to customize the build:
Option |
Description |
---|---|
|
Additional arguments for CMake |
|
Preferred C++ compiler |
|
Selected CMake generator |
|
Use MKL as the BLAS library. The MKL library is searched in the
Python prefix path. Alternatively, you can set the environment variable
|
|
Forces CMake to reconfigure this project |
Run the python setup.py build-ext --help
command to view all the options
available for building the library.
Test#
Requirements#
Running tests require the following Python libraries:
Running tests#
The distribution contains a set of test cases that can be processed with the standard Python test framework. To run the full test suite, use the following at the root of the project:
pytest -v
Automatic Documentation#
The source code of this documentation is managed by sphinx. It is possible to generate it in order to produce a local mini WEB site to read and navigate it. To do this, type the following command:
sphinx-build -b html docs/source docs/build
Install#
To install just type the command pip install . at the root of the project.