test

test

…DR5_tenToA0Tuned-Q_injection_boris.ipynb

removing a loop for y0 in completeness computation with scattering. updating averaged noise when using averaged Q.

average Q is computed from input files from source injection method, this is binned lkl update

…g completeness

update to make it clear what option is chosen for completeness and which Q data is sourced, T_CMB in CCL is set to use the default value to match NEMO, starting point of fine binning at low redshift in binned lkl is corrected, now Ntot from binned and unbinned are in good agreement (there is still some issues with Q but this is applied for both), option to cut some low redshift bins for binned lkl is added, cash lkl is tidied up, column order when reading catalogue in possion lkl is corrected, both code works when including intrinsic scatter but slow at the moment

 into dev-clusters-both

first attempt of trying pytest for both binned and unbinned cluster likelihood

speeding up the unbinned lkl without scatter (no repeated reading in for injection Q) and tidying up (removing obsolete funcs and renaming some params and funcs for clarity). rms_bin_index is not needed anymore for prediction, but still needed for rate densities.

for some reason, vectorisation for Pfun_per didn't speed up the code at all on my side. went back to the older version and read the entire catalogue at once and pass the tile index in computation of y0. starting preparation for inclusion of scatter.

synchronizing intrinsic scatter computation between binned and unbinned lkl (log10(y0) is switched to ln(y0) in unbinned)

switching back to error function for completeness

comoving volume element added for the rate density calculation, an interpolator for the comoving volume element added, source injection method for the unbinned implemented, low redshift cut for the unbinned implemented, application of debias factor checked

minor change with cash statistics, binned intrinsic scatter part

… thanks to @inigozubeldia

matches class_sz