Installation instructions¶
Installation of RegPy¶
We provide currently a setup of RegPy on your machine by:
- Building/Installing using pip 
- Building/Installing from sources 
- Running in a docker image 
In the following we discuss the installation steps for each method in more detail. If you observe any problems during the installation, please create an issue on github with detailed information so that we may reproduce the problem github issue tracker.
Regpy in its cor version depends only on numpy and scipy. However, we provide a woking interface for ngsolve. This can be installed as an optional dependency and is by default included in the docker image.
Installation using pip¶
We publish RegPy on the PyPi, which can then be installed running the command:
pip3 install regpy
Note that you can add standard options to pip3 influencing the installation, as well as specify a specific version. If you wish to install with the optional dependency for ngsolve use can use
pip3 install regpy[ngsolve]
Installation from source¶
You can directly install (through pip) the latest version of RegPy from github using
pip3 install git+https://github.com/regpy/regpy.git@master
or clone and install it as an editable library if you wish to make modifications
git clone https://github.com/regpy/regpy.git@master
cd regpy
pip install --editable .
Again you may add the optional dependency of ngsolve
pip3 install git+https://github.com/regpy/regpy.git@master[ngsolve]
or
git clone https://github.com/regpy/regpy.git@master
cd regpy
pip install --editable .[ngsolve]
Using Docker image¶
For convenience we provide a docker image of RegPy. How to install and use docker daemon can be found in the dockerdocs.
Assuming you have a working docker daemon you can simply run the latest RegPy docker with
docker run -i -t regpy/regpy:latest /bin/bash
The image is also stuffed with a jupyter server. Thus running:
docker run --name regpy-jupyter -p 8000:8000 regpy/regpy:latest
This will lunch a docker container with a jupyter server having RegPy already installed. To access the jupyter server just open the link that is displayed when running the above command.