Implement Cluster reduced shear measurement using pos x shear correlation using TreeCorr

[combet]

This will be compared to the TXPipe "CLMM g_t implementation" developed by @eduardojsbarroso in the cluster_source_selection branch.

Starting point will the using/adapting what is in extensions/cluster_counts/rlens.py and twopoint.py

Questions raised during discussion:

• We do not yet have random catalogs for the cluster (lens)
• Do we need the computation including the correction by randoms or raw <g_t> when comparing to the model?

[marina-ricci]

We are trying to implement the cluster-reduced shear cross correlation. We follow To et al. 2021

This should be equivalent to the computation of NG correlation with a random lens catalog, as specified in TreeCorr doc (https://rmjarvis.github.io/TreeCorr/_build/html/ng.html).

Including random catalog for cluster is said to help accounting for PSF systematics, and diluation effects, that are computed as (To et al 2021, Sheldon et al 2004 ) :

Sources:

• [Sheldon et al 2004] https://iopscience.iop.org/article/10.1086/383293/pdf
• [To et al 2021] https://arxiv.org/abs/2008.10757