Erwan Pannier

@dcmvdbekerom I updated the test conditions in the main PR tests ; so far I see mostly an improvement over a regular grid ; and no penalty in the worse...

Decided on 16/08/24 with @minouHub : merge and make it a default option as it; if and only if errors are acceptable (compared to other interpolation errors)

I'm curious about how it is done and implemented in other packages. Do you know of any gpu-accelerated Python package that doesn't require the C++ compiler & CUDA Toolkit ?...

Btw the `deploy` step of Radis (activated on pushes to the master branch) already builds a wheel https://github.com/radis/radis/blob/develop/.travis.yml#L84

Said today : we'll release next version 0.15 with `cibuildwheels` and see if it works ! I'm taking this.

@dcmvdbekerom is this still relevant, ie do we still use Cython code?

@dcmvdbekerom aren't the test failing just because you aren't running on a Travis-GPU enabled instance ? It can be activated here: https://docs.travis-ci.com/user/reference/overview/#gpu-vm-instance-size

(side note, if building opacity databases have a look at the SpecDatabase class which is basically a folder of precomputed `.spec` files. For instance if changing path lengths you don't...

@menasrac please add a runnable test code in the PR main text so we can test it

@EkulRF have a look and confirm this is what you need; then we'll merge