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?