olender / spyro-1 Goto Github PK

@Olender: The coordinate system should be set up to the usual coordinate system used in Engineering. If there are interactions with other knowledge fields, such as Geophysics, only the output files should be modified according to the convention adopted for them.
Thus, errors due to the misinterpretations of the coordinate system can be avoided.

Add a Dockerfile and workflow instructions in GitHub Actions to automate the build and delivery of images with spyro.

Add elastic wave propagation forward solver with PML.

@Olender: The coordinate system should be set up to the usual coordinate system used in Engineering. If there are interactions with other knowledge fields, such as Geophysics, only the output files should be modified according to the convention adopted for them.
Thus, errors due to the misinterpretations of the coordinate system can be avoided.

Finish new FWI implementation, add unit tests, at least one tutorial and compare with results from the old spyro paper.

I am going to try and create a minimal reproducible example for every error that shows up and if it is not related to our software submit an issue in the Firedrake repo.

Add linting to CI

It might be slower, but with the Mintrop cluster going temporarily down in two months (it is getting moved to a new location), we need to be able to run problems locally.

As mentioned in PR #49, the gradient test sometimes fails. I have not yet been able to reproduce this error. The gradient test always uses a random direction, so I plan to manually try different directions to find out what is happening, focusing on boundaries and other possible edge cases. I can also save the direction tested every time the test fails.

Error on mask and velocity projection appearing on HABC branch

The NDF lab internet infrastructure has changed and we have temporarily lost access to the runner

TotalEnergies provided a model for the Brazilian presalt and it would be interesting to check how our current acoustic solvers perform on this model. Then, this issue involves the implementation of an example that runs the forward acoustic model for the Brazilian presalt deep water model provided by TotalEnergies. Also, notebook explaining how to run this example will be added.