ranking_HM.Rank#
Back to Module page
Summary#
No class docstring summary available.
Method |
Summary |
|---|---|
No docstring summary available. |
|
Deprecated on 6/2024 Saves the class to a pickle |
|
Saves the class to a hdf5 file. Ensure that you ran check_and_fix_bg_fg_lists if you added extra injections (scoring does it automatically) Note that saving to a new path doesn’t replace what is in the instance unless path is None, if you want to use the new hdf5 file created, load it using from_hdf5 |
|
Load the class from a hdf5 file |
|
Load attributes from a hdf5 file into an instance of the class |
|
Removed from init to make saved rank objects independent of the cogwheel version |
|
After maximizing over banks, there are fewer entries in cands_preveto_max, which are no longer in order of subbank_id. Apart from the return, this function 1. Recomputes the median normfacs for the subbanks 2. Reevaluates the sensitivity penalties for all candidates 3. Reorders the entries in cands_preveto_max |
|
It can be slow to loop over the entries in scores_(non)vetoed_max, read them into memory, and compute something, in bulk. It is faster to dereference the hdf5 arrays in bg_fg_by_subbank once, compute what we need and read the values relevant to scores_(non)vetoed_max |
|
Compute quantities needed to assign scores to the triggers Responsible for populating: 1. coherent scores 2. rank functions (CDF of P(SNR^2\|H0) in each detector) If we are rerunning with a different subset of vetoes, update the desired set in self.mask_veto_criteria and, if apply_veto_before_scoring is True, set redo_bg = True |
|
TODO: Speed this function up like rank_scores_calc_MZ Calculating ranking scores for all the bg, fg, lsc, inj triggers. We first group the templates based on their “glitchiness” and then make separate rank functions for different groups. |
|
Compute FARs using ranking scores as test statistics |
|
No docstring summary available. |
|
Run after scoring the bank bg |
|
Calculating ranking scores for all the bg, fg, lsc, inj triggers. The rank functions are constructed in a sub-function make_score_funcs_MZ(). We first group the templates based on their “glitchiness” and then make separate rank functions for different groups |
Class docstring#
No docstring text available.