The Python-based command-line interface (CLI) and the Python API that we call renku
is available for use via RenkuLab sessions and you can also install it on your local machine. Most available CLI commands are documented in our cheatsheet <../renku-python/docs/_static/cheatsheet/cheatsheet.pdf>
.
The renku
command-line interface is already installed if you start a Session from RenkuLab with a project you created on RenkuLab or initialized via renku init
in a local repository.
See renku_cli_upgrade
for upgrading to the latest version of the CLI for Sessions.
Install and execute renku in an isolated environment using pipx
. It will guarantee that there are no version conflicts with dependencies you are using for your work and research.
Note
This is the method of installation in the Renku docker images, i.e. the default environment you use when you launch a JupyterLab session via the Renku browser interface.
Install pipx and make sure that the $PATH
is correctly configured.
$ python3 -m pip install --user pipx
$ pipx ensurepath
Once pipx
is installed use following command to install renku
.
$ pipx install renku
$ which renku
~/.local/bin/renku
To upgrade renku to the latest stable version:
$ pipx upgrade renku
To install renku at a specific version:
$ pipx install --force renku==0.10.3
To upgrade to the latest development version:
$ pipx upgrade --pip-args=--pre renku
If you run into dependency problems during the CLI installation, have a look at the cli-troubleshooting
.
For development or if the above methods do not work for you, Renku can also be installed with pip
or pipenv
:
pip install --pre renku
If you want to have Renku as a project dependency, you may install it using pipenv
in your project's directory:
pipenv install --pre renku
pipenv shell
Using conda
, we can very easily install renku from source into an isolated environment.
To install the latest bleeding-edge version of renku and keep it from polluting your application environment, an easy solution is to place it inside an isolated conda
environment. If you don't have conda
already, you should install miniconda. Once you have it installed, you can run
conda create -y -n renku python=3.6
$(conda env list | grep '^renku\s' | awk '{print $2}')/bin/pip install -e git+https://github.com/SwissDataScienceCenter/renku-python.git#egg=renku
mkdir -p ~/.renku/bin
ln -s "$(conda env list | grep '^renku\s' | awk '{print $2}')/bin/renku" ~/.renku/bin/renku
This will create an isolated environment for renku and link the binary to .renku/bin
in your home directory. If you want to use it, you should add this to your PATH
:
export PATH=~/.renku/bin:$PATH
If you want it to be done automatically for your shell (bash), add it to .bashrc
:
echo "export PATH=~/.renku/bin:$PATH" >> $HOME/.bashrc
source $HOME/.bashrc
When you want to update the installed version again, simply do
$(conda env list | grep '^renku\s' | awk '{print $2}')/bin/pip install -e git+https://github.com/SwissDataScienceCenter/renku-python.git#egg=renku
To install a specific version of renku the procedure is nearly identical the above, but instead of installing from source you install a version with pip
. For example, after creating the conda
environment as described in the previous section, you can install renku v0.3.0 with
$(conda env list | grep '^renku\s' | awk '{print $2}')/bin/pip install renku==0.3.0
Note
You may get a ValueError: unknown locale: UTF-8
- see here for instructions on how to fix it.