jWave is a library for performing pseudospectral simulations of acoustic signals. Is heavily inspired by k-Wave (in its essence, is a port of k-Wave in JAX), and its intented to be used as a collection of modular blocks that can be easily included into any machine learning pipeline.

Following the phylosophy of JAX, jWave is developed with the following principles in mind

1. Differntiable
2. Fast via jit compilation
3. Easy to run on GPUs

Before installing jwave, make sure that you have installed JAX. Follow the instruction to install JAX with NVidia GPU support if you want to use jwave on the GPUs.

Install jwave by cd in the repo folder an run

pip install -r requirements.txt
pip install -e .

If you want to run the notebooks, you should also install the following packages

pip install jupyter, tqdm

A lot of functionality is still missing compared to kWave, expecially in terms of functions constructing objects. If you want to use some kWave methods from python (for example the kArray class implementing off-grid sources), you should install the MATLAB engine for python. For example:

conda activate myenv
module load Matlab/2019a
cd /apps/software/Matlab/R2019a/extern/engines/python/
python setup.py build --build-base=$(mktemp -d) install

To access the documentation, you need to first forward the port 8000 of the machine you are connecting to. In your local bash, type

ssh -L 8000:localhost:8000 username@remote.address

This is not needed if you are working on your local machine.

After login, activate the environment. The first time you also need to install some extra packages packages

conda activate myenv
pip install -r dev_requirements.txt

To build and serve the documentation, type

mkdocs serve

from the root folder of the repository and then navigate to http://localhost:8000/ in a browser.

If the above command fails, you may need to install some extra packages via pip before being able to build the documentation, such as mkdocs, markdown-katex and mkdocstrings: they will be listed at the end of the output returned by mkdocs serve.

Note: Those links work only if mkdocs serve is running

1. Getting started
2. Simulation examples
3. Optimization examples
4. API Reference

1. ADSeismic.jl: a finite difference acoustic simulator with support for AD and JIT compilation in Julia.

1. Adler et al. 2021: Deep Learning for Seismic Inverse Problems