Revise sweepformula select operation #2109

MichaelHuth · 2024-05-13T23:07:38Z

close #2012

…m a graph + sweepNo - Added SFH_GetLabNoteBooksForSweep to retrieve LNB only - Added SFH_GetLabNoteBooksAndDFForSweep to retrieve LNB and SweepDFR These functions abstract the differences between a graph that is a DB and a SB. For an SB graph a valid mapIndex must be given as argument. preparation commit for the following changes

t-b · 2024-10-14T19:59:35Z

I've adapted PSX to now correctly use the LBNs with mapIndex. For that I also pulled out the earlier adapations to now have them in only one commit.

Range-diff: range-diff.txt

Will merge once CI passes.

problem described in #2151 SF requires to retrieve the correct LNBs and sweepDFR for sweeps. Through the analysisbrowser sweeps from multiple devices and/or experiments can be loaded. This must be handled correctly or wrong data is retrieved/analysed. Thus, SF functions that retrieve sweep data are adapted to handle that correctly. For selections a forth column was added to store the mapIndex of the sweepMap that is unique for a sweepNo, experiment, device combination. For graphs from SB the mapIndex is used to retrieve the correct LNB, SweepDF. ** THIS BUGFIX IS LIMITED TO SWEEP HANDLING IN SWEEPFORMULA ** Other parts of MIES may be still unaware of multiple experiment/devices in the SB.

- allows to select device and experiment - only one of these opeations are allowed for select and they must each yield a single result, but allow wildcards - the remaining behavior is like selrange where the information is taken from the outermost select only. - Add sweepMapIndex to trace user data in CreateTiledChannelGraph - Rewrite of selection gathering and adaptations to sweep getting to be aware of different experiment that require to work with sweepmap indices instead of sweep numbers. - adaptations of all SF functions that retrieve LNB and sweepDFR to be aware of multiple experiments present in a sweeepbrowser that could be selected by select

- remove now unused and obsolete BSP_GetSweepDF function

selexpandsci: Extends the selected sweeps with all sweeps from the same stimset cycle id selexpandrac: Extends the selected sweeps with all sweeps from the same repeated acquisition cycle The operations take no arguments. The feature is enabled if they are added to select(....) if both are used then first SCI and then RAC is applied.

- both operations take exactly one numerical argument that sepcifies a specific setCycleCount and/or setSweepCount - if not used in select then the filter is ignored

both operation take exactly one numeric argument specifying the index selRACIndex selects all selections that have a specific index within a repeated acquisition cycle selSCIIndex selects all selections that have a specific index within a stimset cycle

For some cases select was returning a zero sized wave for no results. This was unified to be a null wave and the select argument parsing was adapted to distinguish between regular filters from sel* operations and select arguments

- selsweeps can return a null wave that was used as default value for the select filter. This clash was resolved by adding a seperate flag to indicate if select got any sweep numbers as argument. The checks for the default were changed to check against the new flag instead. The flag approach also keeps the check against specifying selsweeps multiple time intact.

- for seldev, selexp, selsetcyclecount, selsetsweepcount, selsciindex, selracindex, selexpandsci, selexpandrac

… usage In principle the sweep browser supports multiple devices/experiments. However not all sweep browser functions imlement a suitable API for that. SB_GetSweepIndexFromMap is such function. When running tests it asserts now if sweep numbers appear more than once, which can happen when multiple experiments are loaded and this function is called.

- selsetcyclecount, selsetsweepcount, selsciindex, selracindex - selexpandsci, selexpandrac

- sometimes PS_PB6 acquires sweep 2 as full sweep if the computer is slow. However the test checked if there are no results. If there was a full sweep acquired it tests now against existing results.

- LNB is handed down from SF_GetSelectData - For the displayed case the LNBs are gathered per trace with TUD data from CreateTiledChannelGraph

- LNB handed down from SF_GetSelectData - Uses correct LNB now for SB case

…tilities_Logbook.ipf - use GetHeadStageForChannel to retrieve headstage - Add function documentation - adapt call sites

This avoids having invalid data in there. Missed in the initial implementation in abe808f (SF: Add select operation to select sweeps for data, adapted data operation, 2022-02-04).

In the same spirit as d5b560f (Bugfix: SF use proper handling of sweepMap for SB, 2024-10-04) but for PSX. We also need to include now the experiment name in the combo key to make it unique. And for the stats equivalence data we also include the sweepmapindex next to the sweepNo.

MichaelHuth added the SweepFormula label May 13, 2024

MichaelHuth self-assigned this May 13, 2024

MichaelHuth requested review from timjarsky and t-b as code owners May 13, 2024 23:07

MichaelHuth force-pushed the feature/2109-revise_sf_select branch from 040b6ef to 2102576 Compare May 15, 2024 13:14

MichaelHuth mentioned this pull request May 21, 2024

Extend GetSetIntersection to return indices of intersecting elements #2121

Merged

MichaelHuth force-pushed the feature/2109-revise_sf_select branch 2 times, most recently from ce5b821 to d8ccbec Compare May 23, 2024 16:26