Provide SEG-Y I/O functions about deepwave

[erellaz]

Thanks again for sharing Deepwave with the community. It is a fantastic package. Thanks also for your support.

Here is a contribution: a SEGY wrapper for Deepwave (or any other python forward modeling package). It includes a full example going from a classic benchmark SEGY velocity model (the BP tooth model) to SEGY shots generated by Deepwave and usable by any seismic package.
The example is 2D, but the package will work in 3D (with a few judicious reshape).
It also uses Obspy (built on SEGYIO) which is more convenient (or higher level) than SEGYIO alone to write actual SEGYs from scratch.

https://github.com/erellaz/SEGY_Wrapper/blob/main/README.md

[ar4]

Nice work! I have always found writing data to SEG-Y from Python to be troublesome, so this will be useful. Thank you. I am also very pleased to hear that you have found Deepwave to be useful. The dispersion that you saw is a bit concerning. I suspect that it is due to using quite a low order finite difference approximation. I used a low one to make the process faster, especially for FWI, but it's probably too low for some situations. I would like to add a reference to your repository to the Deepwave README. How does this sound: @erellaz has created [a nice tool](https://github.com/erellaz/SEGY_Wrapper) to make loading and saving SEG-Y files easier, and a demonstration of using Deepwave for forward modeling with a realistic 2D model.

[erellaz]

Awesome, thanks. Feel free to use as you see fit.
I will keep working on the wrapper in the next few days to populate key headers of the Segy from the Deepwave geometry (at least shot number, receiver number, shot and receiver X and Y and offsets), and add a few details (custom EBCDIC from files).