[DISCUSSION] Needed modeling in CCL and cosmosis

[jablazek]

Needed modeling

What are the modeling capabilities that are most useful for community analysis of new and existing measurements?

Contacts: Jonathan Blazek, Elisa Chisari
Participants:
When will this discussion take place?
Wednesday TBD

Discussion Items

• Current CCL and cosmosis capabilities
• Near-term planned capabilities
• Parameterizations, etc.
• Missing documentation?
• What else is needed?
  -- analysis of both projected correlation functions and angular statistics
  -- additional effects that must be included (ggl, lens magnification, etc)

Detailed description

[add additional details about the topic]

[danielgruen]

Is a 'forward modeling' capability for indirect measurements (e.g. given a joint distribution of z,L,sSFR or whatever, of a sample) in scope for this discussion?

[harrysjohnston]

On Hankel transformations to get projected alignment statistics (w_g+, w_gx, w_xx) from P(k)'s:
C_ell’s and xi_+/- are straightforward, but the Hankel transforms for w_g+ and w_++ from IA P(k)’s are not explicitly implemented in many places. The hacks which have been mentioned are exploiting the similar form of angular C_ell → angular CF integrations. These are in CosmoSIS @ joezuntz / cosmosis / wiki / default modules / cl to xi nicaea 1.0 — Bitbucket and in CCL @ CCLX/CellsCorrelations.ipynb at master · LSSTDESC/CCLX · GitHub (“correlation functions”, at the bottom)