Installation

xsar relies on gdal and other shared libs, that are not provided by pip. So insallation in a conda environement is recommended.

conda install

  1. Install mamba

For a faster installation and less conflicts between packages, it is better to install xsar with mamba

conda create -n xsar
conda activate xsar
conda install -c conda-forge mamba
  1. Install xsar (this won’t include the readers)

mamba install -c conda-forge xsar
  1. Add optional dependencies (readers)

  • Add use of Radarsat2 :

mamba install -c conda-forge xradarsat2
  • Add use of Sentinel-1

mamba install -c conda-forge xarray-safe-s1
  • Add use of RCM

pip install xarray-safe-rcm

pip install

  1. Install xsar (this won’t include the readers)

conda create -n xsar
conda activate xsar
pip install xsar
  1. install xsar with optional dependencies (to use Sentinel-1, Radarsat2, RCM…)

  • Install xsar including Sentinel-1 :

pip install xsar[S1]
  • Install xsar including Radarsat2 :

pip install xsar[RS2]
  • Install xsar including RCM :

pip install git+https://github.com/umr-lops/xarray-safe-rcm.git
pip install xsar
  • Install xsar including multiple readers/dependencies (here Radarsat2 and RCM):

pip install xsar[RS2,RCM]
  • Install xsar including Radarsat2, Sentinel-1 and RCM:

pip install xsar[RS2,S1, RCM]

Update xsar to the latest version

xsar conda package can be quite old:

https://anaconda.org/conda-forge/xsar/badges/latest_release_relative_date.svg

To be up to date with the developpement team, it’s recommended to update the installation using pip:

pip install git+https://github.com/umr-lops/xsar.git

Developement installation

git clone https://github.com/umr-lops/xsar
cd xsar
# this is needed to register git filters
git config --local include.path ../.gitconfig
pip install -e .
pip install -r requirements.txt