Fixed many problems with merge statistics
Rebased from branch data_not_found
, please look at !270 (merged) first.
This merge fixes the following important issues:
-
The
Peak3D::_meanBackground
(anIntensity
) was not initialised withPeak3D
, resulting in many garbage/undefined values, which severely messed up the statistics. I now have reasonable quality metrics for the brd4 dataset: -
There was a fundamental misunderstanding between developers and instrument responsibles about the meaning of "merging": it does not involve merging two peak collections, rather it involves merging symmetry-related peaks within a set of predicted peaks as an interal check of self-consistency. This has been fixed in the GUI by defaulting to selecting a single predicted peak collection for merge statistics.
-
Fixed a few bugs in the display of the merged and unmerged views in
SubframeMergedPeaks
.
This breaks compatibility with existing .nsx
files, which will contain garbage _meanBackground
values in many cases.
Resolves #274 (closed) and #280 (closed)