Projected correlation function support

[chrgeorgiou]

This PR adds support for computing projected correlation functions with CCL. These have the form:

$w_{ab}(r_p)=\int dz \ W(z)\int\frac{dk \ k}{2\pi^2} \ P(k,a) \ J_n(k \ r_p)$

with $W(z)$ a window function that depends on the redshift distribution, $P(k,a)$ the relevant power spectrum and $J_n$ the bessel function. This function is often used in direct measurements of galaxy clustering and intrinsic alignments.

This PR uses the FFTLog wrapper in CCL (similarly to what is used in the halo model) to speed-up the integration.

[coveralls]

Pull Request Test Coverage Report for Build 14642990103

Warning: This coverage report may be inaccurate.

This pull request's base commit is no longer the HEAD commit of its target branch. This means it includes changes from outside the original pull request, including, potentially, unrelated coverage changes.

• For more information on this, see Tracking coverage changes with pull request builds.
• To avoid this issue with future PRs, see these Recommended CI Configurations.
• For a quick fix, rebase this PR at GitHub. Your next report should be accurate.

Details

• 35 of 51 (68.63%) changed or added relevant lines in 1 file are covered.
• 41 unchanged lines in 7 files lost coverage.
• Overall coverage decreased (-0.2%) to 97.241%

---

Changes Missing Coverage	Covered Lines	Changed/Added Lines	%
pyccl/correlations.py	35	51	68.63%

Files with Coverage Reduction	New Missed Lines	%
pyccl/baryons/baccoemu_baryons.py	1	98.72%
pyccl/cells.py	1	97.92%
pyccl/nl_pt/ept.py	1	99.63%
pyccl/pk2d.py	3	98.71%
pyccl/_nonlimber_FKEM.py	5	96.12%
pyccl/tracers.py	14	95.1%
pyccl/halos/pk_4pt.py	16	96.25%

Totals
Change from base Build 11816689318:	-0.2%
Covered Lines:	6591
Relevant Lines:	6778

---

💛 - Coveralls

[elisachisari]

Some top level feedback without going into the details:

• could you add a benchmark?
• could you add an example in CCLX?
• could you edit the docs files so this new function is included there?

[damonge]

@chrgeorgiou can I check what the status of this PR is?

[chrgeorgiou]

Hi @damonge, I just received a benchmark code from Elisa, I will tick the points raised above soon and update the PR.

[chrgeorgiou]

@elisachisari Here is an update on this. I have run the benchmark code, saved the resulting file in the benchmark data and created a benchmark test. I also created an example in the CellsCorrelations.ipynb in CCLX. In terms of documentation, I take it you are referring to the CCL/doc, I can add this there too.

I have to note that the code developed here is only able to compute the projected correlations over infinite baseline (so $\Pi_\mathrm{max}=\infty$). Having an arbitrarily defined $\Pi_\mathrm{max}$ will need more development if the FFTLog method is to be used. I am happy to continue with this PR as is and add this feature at a later time if requested.

[damonge]

@chrgeorgiou @elisachisari can I check what the status of this is? Is it meant to be ready for review?

[chrgeorgiou]

Hi David, I think the development is enough to have this merged from my side, but I would like to hear what Elisa thinks as well.

[nikosarcevic]

Hi David, I think the development is enough to have this merged from my side, but I would like to hear what Elisa thinks as well.

what is the status of this? woudl be great if this was in ccl now.

[chrgeorgiou]

Hi Niko, I think the work is implemented and ready but there are some drawbacks to the implementation:

1. The current implementation re-codes the fftlog code in the correlations.py file, which is okay but not ideal.
2. The correlations assume infinite baseline (Pi_max to infinity) which is also not ideal, but speeds up the computation significantly.

It is likely still a useful implementation to be in CCL and I am happy to move forward with this but I'm curious if @elisachisari has a comment on this.

[elisachisari]

I think it's useful to have for now. I think we could add later a case with Pimax and also multipoles.