Local primordial non-Gaussianities in initial conditions

[jordanflitter]

This PR includes the ability to have local PNGs in the initial conditions. At this moment, it only works with POWER_SPECTRUM!=CLASS, which is why it's in a draft state. The CLASS transfer function, being defined differently than the other transfer functions, requires a different implementation, so this PR will be in a draft state until there's also a support with having POWER_SPECTRUM=CLASS and PRIMORDIAL_NON_GAUSSIANITIES=True.

Some notes:

• According to this paper, https://arxiv.org/pdf/1304.8049, it is not enough to simply change the initial conditions, but it's also necessary to modify EPS (eq. 5). This requires computing analytically the 3-point function in linear theory. This is much harder and will be done in a future PR.
• I added an output called hires_Phi, just to be able to see directly how the primordial gravitational field is modified due to PNGs. It is only allocated when PRIMORDIAL_NON_GAUSSIANITIES=True, and it's a single (high-resolution) box in the initial conditions (unlike output boxes that we make at each node_redshift), so I'm not sure if users should be concerned about this extra memory allocation.
• I added some functions in InitialConditions.c like add_pngs_to_delta_box, add_pngs_to_Phi_box, multiply_by_factor_in_Fourier_space and Poisson_factor. I am not entirely sure that the last two functions should indeed be placed in InitialConditions.c. There is a lot of resemblance between filter_box and multiply_by_factor_in_Fourier_space, and I actually think that we could use the latter to perform the former (since filtering is actually multiplying the box in Fourier space by a k-dependent factor, this is exactly what multiply_by_factor_in_Fourier_space does). I am not sure this PR is the best place to do this, but I do think it worths of making an issue.

[codecov]

Codecov Report

❌ Patch coverage is 60.00000% with 2 lines in your changes missing coverage. Please review.
✅ Project coverage is 88.23%. Comparing base (02ed909) to head (009268d).
⚠️ Report is 17 commits behind head on release-v4.2.
✅ All tests successful. No failed tests found.

Files with missing lines	Patch %	Lines
src/py21cmfast/wrapper/outputs.py	33.33%	1 Missing and 1 partial ⚠️

Additional details and impacted files

@@               Coverage Diff                @@
##           release-v4.2     #631      +/-   ##
================================================
+ Coverage         88.01%   88.23%   +0.22%     
================================================
  Files                32       32              
  Lines              4755     4760       +5     
  Branches            803      804       +1     
================================================
+ Hits               4185     4200      +15     
+ Misses              410      399      -11     
- Partials            160      161       +1     

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

[steven-murray]

This looks cool @jordanflitter!

I will refrain from a full review since it's in draft state, but I will note that I think hires_Phi is a bad name :--) Phi is a symbol name, not a physically meaningful name.

Otherwise, I agree in spirit with all your notes!