Flexible modeling and simulation software for free-surface flows.
Zoomy's main objective is to provide a convenient modeling interface for complex free-surface flow models. Zoomy transitions from a symbolic modeling layer to numerical layer, compatible with a multitude of numerical solvers, e.g. Numpy, Jax, Firedrake, FenicsX, OpenFOAM and AMReX. Additionally, we support the PreCICE coupling framework in many of our numerical implementations, to allow for a convenient integration of our solver with your existing code.
See our Documentation for details on
- how to get started
- tutorials
- examples
- API
- ...
The Zoomy code is free open-source software, licensed under version 3 or later of the GNU General Public License. See the file LICENSE for full copying permissions.
T.b.d.
ZoomyLab consists of a
- Base repository (Zoomy): Symbolic Layer (Models), Pre/Postprocessing, NumPy solver
- Solver backends, each with it's own (sub-)repository
- JAX (zoomy_jax)
- Firedrake (zoomy_firedrake)
- FenicsX (zoomy_fenicsx)
- AMReX (zoomy_amrex)
- Utility backend
- meshes (currently GMSH mesh definition files)
- data (data repository for large scale test cases)
A good way to have an interactive first impression of Zoomy is to use one of the following options
WIP
You can either clone the repo with all related subrepositories via
git clone --recurse-submodules https://github.com/ZoomyLab/Zoomy.git
or you can start by cloning the main repository (Zoomy) and selected subrepositories, e.g.
git clone https://github.com/ZoomyLab/Zoomy.git
cd Zoomy
git submodule update --init meshes
git submodule update --init library/zoomy_core
git submodule update --init library/zoomy_jax
The different subrepositories are listed at ZoomyLab
Clone the repository (see above) and open the repository in your IDE / VS-Code. A pop-up will appear if you want to open the workspace in a container. You can choose between:
- Zoomy + Jax
- Zoomy + Firedrake
Requirements:
- Docker installation
- 'Dev Container' extension in VS-Code
Restrictions:
- On Windows, you need to use Linux Containers for Docker.
Standalone versions: These versions do not allow for changes of the zoomy library.
- Zoomy + JAX:
docker pull docker push ghcr.io/zoomylab/zoomy_jax_standalone:latest - Zoomy + Firedrake:
docker pull docker push ghcr.io/zoomylab/zoomy_firedrake_standalone:latest
Development versions: These versions do allow for changes of the zoomy library. They require to 'pip install' the local zoomy libraries or the use of the .devcontainers.
We offer Docker containers for
- Zoomy + JAX:
docker pull docker push ghcr.io/zoomylab/zoomy_jax:latest - Zoomy + Firedrake:
docker pull docker push ghcr.io/zoomylab/zoomy_firedrake:latest
Restrictions:
- On Windows, you need to use Linux Containers for Docker.
- Clone the repository (see above)
- Install Zoomy and the respective solver backends, e.g.
- Zoomy (Numpy solver backend + GMSH support):
conda env create -f install/Zoomy.yml
conda activate zoomy
conda env update -f install/meshes.yml
python3 -m pip install library/zoomy_core
- Zoomy + JAX (Numpy + JAX solver backend + GMSH support)
conda env create -f install/Zoomy.yml
conda activate zoomy
conda env update -f install/zoomy_jax.yml
conda env update -f install/meshes.yml
python3 -m pip install library/zoomy_core
python3 -m pip install library/zoomy_jax
Note that the AMReX installation is completely indepdenent and the AMReX solver does not depend on the Conda/Mamba environment. Follow the instructions on the AMReX Webpage
WIP
WIP
WIP
See the install/*.yml for a complete list of requirements.
TODO: complete list of dependencies
The following environment variables can be set
ZOOMY_DIR=/path/to/Zoomy
JAX_ENABLE_X64=True
PETSC_DIR=/path/to/petsc/installation
PETSC_ARCH=architecture used for compiling petsc
WIP
WIP
WIP