Closed georgemilosh closed 10 months ago
@georgemilosh could you format the code with black? Then I am happy to include the changes into the repository.
Sorry what do you mean by formatting in black?
black is the formatter we use to make sure code is always formatted with PEP8, you can do this by installing black
pip install black
and then formatting the source code directory: black src/deepymod
.
The thing is that I added a lot of edits to the fork meanwhile, so maybe that's why it failed?
Basically I re-run the notebooks, and added some of my material, like Allen Cahn equation etc. Tried unsuccessfully Kuramoto-Sivashnsky. Plus I found that Keller - Segel equation is no longer identified in DeePyMoD
, while it was almost identified correctly in DeePyMoD_torch
which also takes gradients with respect to L1 loss unlike DeePyMoD
. Also to be able to run Keller - Segel equation I had to apply a small fix to the library.
Hi George! I think it is a version thing with the black package, your edits look like great fixes and additions to me so I will merge them now.
Concerning the examples added, a couple may need data from another equation discovery repo, which could be addressed later
Yes, if you get any further with that feel free to open anothr PR. For now I pulled everything into the master branch!
In the current version of the repo the noise applied during processing is overwritten by the normalization