/ptsa_new

PTSA - Python Time Series Analysis

Primary LanguagePythonGNU General Public License v3.0GPL-3.0

PTSA

https://travis-ci.org/pennmem/ptsa_new.svg?branch=master

For documentation and tutorials, please see https://pennmem.github.io/ptsa_new/

Install via conda

Available on Linux, Mac, and Windows 64 bit:

conda install -c pennmem ptsa

Report bug or feature request

To report a bug or a feature request please use https://github.com/pennmem/ptsa_new/issues.

Build from source

Install dependencies:

conda install -y numpy scipy xarray swig traits

You will also need to install FFTW. To install FFTW with conda on Linux or Mac:

conda install -y -c conda-forge fftw

Alternatively, it can be installed with the system package manager on Linux (Debian-based command shown below):

sudo apt-get install libfftw3-dev

or on Mac using homebrew:

brew install fftw

To read EDF files, you will also need to install pybind11:

conda install -y -c conda-forge pybind11

Install PTSA:

python setup.py install

If you encounter problems installing, some environment variables may need to be set, particularly if you installed FFTW with conda. If your anaconda distribution is installed in $HOME/anaconda3 and the environment name is ptsa, set the CPATH and the LD_LIBRARY_PATH as follows:

export CPATH=$HOME/anaconda3/envs/ptsa/include
export LD_LIBRARY_PATH=$HOME/anaconda3/envs/ptsa/lib

Running Tests

To run the PTSA test suite locally, first set up a testing environment:

conda env create -f environment.yml
source activate ptsa

and then build build the extension modules and run the test suite:

python setup.py develop
pytest tests/

The shell script run_tests will also run the test suite, assuming the environment is configured.

To skip tests that depend on Rhino the NO_RHINO environment variable must be set:

export NO_RHINO=TRUE

When running tests which require rhino access, the path to the root rhino directory is guessed based on common mount points.

Building conda packages

If you don't already have it installed, you'll need to install the conda build tool:

conda install conda-build

Update the version number in conda.recipe/meta.yaml.

Build packages with:

conda build conda.recipe

To allow uploads you need to install anaconda-client:

conda install anaconda-client

After that installing anaconda-client you need to to provide your anaconda.io login credentials:

anaconda login

At this point you will be ready to upload newly built conda PTSA packages. After the build is successfully completed you go to the directory where package tarballs have been generated and type:

anaconda upload --user pennmem ./ptsa-*.tar.bz2

Hint: conda packages will be most likely generated in <conda installation dir>/conda-bld/<architecture_folder> where <architecture folder> denotes name of the arget architecture for which conda package was build. e.g. on 64-bit Windows the architecture folder will be called win-64 (hence conda packages will be generated in <conda installation dir>/conda-bld/win-64

Hints for Windows

You'll want to install Microsoft Visual Studio 2015 (or newer) community edition in order to compile extensions. For environment variables to be setup correctly, use the "Developer Command Prompt for VS20xy" which can be found for example in Start->All Programs->Visual Studio 2015->Visual Studio Tools.

License

PTSA is licensed under the GNU GPL version 3.

This repository also includes: