Per cross-C_ell transfer function

[damonge]

Currently the pipeline assumes a single transfer function for all power spectra (essentially assuming equivalent filtering across all maps). We should generalise it to ensure a transfer function can be calculated for any power spectrum combination.

[damonge]

@adrien-laposta , @kwolz and I had the following discussion on slack:

so, I'm thinking, there's an easy generalisation of the pipeline structure if we need to worry about multiple transfer functions

basically we associate filterings with tags, but that association can be one-to-many

i.e. one filtering can be associated to many tags (in our case, a single filtering is associated to all tags)

then, we only calculate the transfer functions for tag combinations that have distinct filtering pairs (which, in our case, is a single transfer function)