Correct output resolution matrix

[p-slash]

This PR redesigns the test for the new resolution matrix output by comparing each row to theoretically expected values.

[weaverba137]

@julienguy, could you please review this? it is possible this fixes #137.

[weaverba137]

I have rebased this branch on main so that test will now run automatically.

[coveralls]

coverage: 74.084%. remained the same
when pulling dc89b5c on correct-output-resolution-matrix
into 51f64f2 on main.

[weaverba137]

@p-slash, apparently we decided to work on this at the exact same time. Please let me know what your next steps are.

[p-slash]

Thanks @weaverba137 . This fix was needed after PR #128 where I changed how the input resolution matrix was downsampled to DESI resolution. It compares the downsampled resolution matrix to theoretical expectations, though a relative tolerance of 0.1% is needed to pass this test.

As far as I know, these two PRs affect only Lya forest P1D, and we have been happy about it.

[weaverba137]

@p-slash, OK, if you have no objections, I'll clean up and merge this PR.

[weaverba137]

I'm also adding a quick fix for #133.

[p-slash]

No objections. Thanks