EFT implementation

[chrgeorgiou]

This PR implements the calculation of integrals in the EFT model of Large-Scale Structure, with particular emphasis on the galaxy intrinsic alignments.

The equations that are implemented are summarised in 2303.15565. The (22)-integrals are in Eq. (2.39) with the associated kernels in Eq. (A.1) and the (13)-integrals are in Eq. (2.40) with kernels in Eq. (A.3).

The FAST-PT implementation here has been benchmarked against an external code that calculates these integrals using FFTLog in mathematica. The agreement is below 1% for the (22)-integrals and below 3% for the (13)-integrals.

Happy to receive comments on the implementation and change things around if necessary.

[chrgeorgiou]

Last few things to do (after discussion with @jablazek):

• Migrate to v4.
• Add option to not subtract the low-k limit.
• Compute the low-k limit properly (now it is just the first P(k_0) value).
• Calculate the optimal/allowed \nu values (optional).

[chrgeorgiou]

Me and Yijie had a chat to port all my work to v4. So far, it only works by utilising the to_do list, so one has to set to_do=['EFT'] to have FAST-PT initialise the integrals necessary for the EFT calculations. Yijie will look at updating this to the v4 infrastructure, where quantities are computed as needed without explicitly requiring the to_do list.