Damian's notes – Open Sourcing of Noiz - new Ambient Seismic Noise processing software

Damian Kula

Open Sourcing of Noiz - new Ambient Seismic Noise processing software

Posted on 2023.05.03

TLDR: We have released to the public Noiz , a new Ambient Seismic Noise processing software.

During my time at Unistra in Strasbourg, I was working on a PhD that focused on Ambient Seismic Noise. In the end, I did not finish the degree, but there is something else left over from those days. Software!

I think I never really fit into academia. I prefer to write code that is of better quality than just a quick and dirty sketch to do some quick analysis with. And then reuse it for the next few years or, if you are particularly unlucky, for most of your career. My aim in writing code was to at least use decent industry standards, tools and techniques and play with best practice, which is not necessarily compatible with the pressures that academia puts on people.

During my time in Strasbourg, I wrote Noiz. Noiz is a web application based software focused on ambient seismic noise processing. It took a slightly different direction than MSNoise or Yam. I thought it would be more cloud-oriented, with the target deployment being a Kubernetes cluster rather than a local machine or a Slurm cluster. Pretty non-standard for scientific software.

That's why it comes out of the box with a Docker-compose suggested deployment (I know, I mentioned k8s before. A Helm chart is coming. Sometime. Promise.) in a separate repo noiz_deployment. It also comes with a minimal dataset against which I have written system tests that can also be used for testing purposes noiz-minimal-dataset. Noiz itself is in a third repository noiz repository.

Unfortunately, the whole open-sourcing process took quite a long time due to bureaucracy, and Noiz has got a bit dusty since then. I still have some obvious technical debts to pay, such as

  • proper packaging
  • testing on a matrix of Python versions (and libs too?)
  • proper CI releases

I hope to be able to solve these issues relatively soon, so that I can happily release code to PyPi, even if it will be under more development in the future.

Merci Beaucoup to Alex for all the help with Noiz as well as his dedication to making open sourcing a reality!