Installing solaris

There are several methods available for installing solaris:

Prerequisites

Regardless of installation method, you’ll need Python version 3.6 or greater. More details on installing Python can be found here. Additionally, if you plan to use the SpaceNet dataset with solaris (it features prominently in many of the tutorials), you’ll need a free Amazon Web Services account and the AWS CLI installed and configured. If you’re just going to work with your own data, you can skip these steps.

Installing from GitHub using a conda environment and pip

If you wish to install a bleeding-edge version of solaris that isn’t available on conda-forge yet, you can install from GitHub. You’ll need anaconda for this installation as well.

From a terminal, run:

git clone https://github.com/cosmiq/solaris.git
cd solaris
git checkout [branch_name]  # for example, git checkout dev for bleeding-edge

If you have access to a GPU where you’re installing solaris, use the following:

conda env create -f environment-gpu.yml

If you don’t have access to a GPU:

conda env create -f environment.yml

Finally, run the last two lines (for installs both with or without GPU):

conda activate solaris
pip install .

The above installation will create a new conda environment called solaris containing your desired version of solaris and all of its dependencies.

Installing with only pip

Use this method at your own risk!

If you have already installed the dependencies with underlying binaries that don’t install well with pip (i.e. GDAL and rtree), you can easily pip install the rest:

pip install solaris

Note that this will raise an error if you don’t already have GDAL installed.