Allow for correlations between redshift bins in P1D

[andreufont]

Irsic et al. (2017) provides a covariance for different redshift bins, and Karacayli will do as well.

Owr p1d_data_base assumes that we have a list of covariances for each redsfhit, we might need to update that at some point.

For now we are ignoring correlations between redshifts.

[andreufont]

This is relevant both for high-resolution P1D from Karacayli et al. (2022), and from DESI mock tests

[andreufont]

Even more relevant now that we have QMLE P1D from DESI EDR in Karacayli et al. (2023), we should add these!

https://arxiv.org/abs/2306.06316