ShapePipe was not designed to be a stand-alone Python library. Instead users are expected to install the full ShapePipe environment on the system(s) where data should be processed.

Standard Installation


Check out Miniconda for a light weight and easy installation of Conda.

The standard installation of ShapePipe manages dependencies and scripts using a Conda environment. Therefore, to follow the standard installation Conda must be available on the system.

The ShapePipe package should first be cloned (or downloaded) from the GitHub repository.


Developers should simply clone the repository as usual.

git clone -b <VERSION> --depth 1
cd shapepipe

where <VERSION> is a tagged release of ShapePipe (e.g. v1.0.1). It is recommend to use the latest release unless you want to reproduce an older set of results.

Then, the entire ShapePipe environment, including dependencies, can be built using the install_shapepipe script as follows.


The install_shapepipe script will create the recommended Conda environment along with all of the required core and module dependencies. The script also provides a checklist for the success or failure of installing each of the dependencies.

Once the installation is complete the shapepipe environment needs to be activated.

conda activate shapepipe

A list installation options can be seen using the --help option.

./install_shapepipe --help

In particular, the --develop flag can be used to install additional tools for developers.

MPI Installation

The standard installation of ShapePipe will install and enable MPI on a given node (i.e. a given machine). However, mpi4py needs to be built from source in order to take advantage of a preinstalled MPI distribution on the compute nodes of a cluster.

This can be done as follows

./install_shapepipe --mpi-root=<PATH TO MPI>

where <PATH TO MPI> is the full path to MPI root directory (i.e. where the bib, include and lib directories can be found). If the installation is successful, ShapePipe will be able to submit jobs to all of the nodes in the cluster.

Uninstalling ShapePipe

The --uninstall flag can be passed to install_shapepipe to remove the entire ShapePipe environment.

./install_shapepipe --uninstall

Installing the ShapePipe Library Only

The ShapePipe library, i.e. the core package not including module dependencies, can be installed in the following ways.

After cloning the repository.


Note, this method will not include any system executables or examples.

pip install .

Without cloning the repository.

pip install git+


If you encounter any problems installing ShapePipe please open an issue and we will do our best to help you.