Installation

Using PIP

For normal use, it is enough to install fpipy using pip from git:

pip install git+https://github.com/silmae/fpipy.git

Optionally, you may also specify dask as an extra dependency if you wish to enable parallel computation. Currently this is just for convenience (you will get the same result if you install dask separately):

pip install git+https://github.com/silmae/fpipy.git[dask]

If you need to access ENVI files (such as the included example hyperspectral datasets) you will need to install rasterio. Specifying ENVI as an extra dependency will install it if you have the required system libraries for GDAL:

pip install git+https://github.com/silmae/fpipy.git[ENVI]

However, due to the difficulty of installing the necessary libraries (especially on Windows) it is recommended you use Conda if you wish to access ENVI files.

Using Conda

Using conda, you can create an environment with rasterio and xarray, then install fpipy using pip:

conda create -n <env_name> rasterio xarray
pip install git+https://github.com/silmae/fpipy.git

There are also multiple ready-made environments under the fpipy/envs directory which may be used to create suitable conda environments using:

conda env create -n <env_name> --file fpipy/envs/<env_name>.yml

See Contributing for more info on using development environments.