Projection of 2 point functions to k_par, k_per basis.

[rainwoodman]

[nickhand]

This is relatively easy to do, but not sure the best way to incorporate it into the existing FFTPower. Maybe a new algorithm?

[cbyrohl]

Would it be to messy just to introduce a distinction of cases to project_to_basis() called in FFTPower?

[rainwoodman]

perhaps subclassing the result class like WedgeBinnedStatistics could work? but it means the result object needs to carry the full 3d ampl per mode. several existing projection ops needs to be refactored there too.

FFT power can still return the tradittional projected result when mode=? is set. but an empty mode will do the full result which can be projected anyway we like?

no i think this is messier than adding new cases in project to bases.

Yu

On Tue, May 15, 2018, 9:59 AM cbyrohl [email protected] wrote:

Would it be to messy just to introduce a distinction of cases to project_to_basis() called in FFTPower?

— You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub https://github.com/bccp/nbodykit/issues/340#issuecomment-389239461, or mute the thread https://github.com/notifications/unsubscribe-auth/AAIbTOZw_7EZHQhNzUUDaa_B98Eqzq-oks5tywlWgaJpZM4NsLJW .