Matteo Concas

Thanks for the prompt look to this! > 1. the squashed digits should have the same sorting (in column, then in row) as original ones, this is required by the...

Hi @shahor02, thanks, indeed the sorting made tha trick. 

Hi @shahor02 I added the CLI parameter to the ITS reco workflow and tested with one ctf of run LHC22f_522711 Here some comparisons: **Before: no masking, no squashing**  After:...

> Thanks, the ratio of tracks is reasonable but the eta shape is not as flat is it was in Oct.2021 test. For the clusters the effect is not evident...

Hi @shahor02, MC label assgnment should be online. I attach also the result for run 519041, where it is visible the effect of squashing pn last of the four bars....

Hi @shahor02, I changed the logic to enable this, so that we can configure it with keyvals, eg: ```bash --configKeyValues='ITSClustererParam.maxROFSquashingDepth=2;ITSClustererParam.maxRowColDiffToSquash=1;' ``` The first parameter sets the number of ROFs to...

@shahor02 ok, and by setting e.g. a negative values to `maxSOTMUS` we could assume it's a way to disable the squashing. Can it work? Or do you prefer an explicit...

> applying the same maxBCDiffToMaskBias check to make sure we don't squash over the gaps BTW: I am not currently applying any selection on ROFs to be squashed based on...

> @mconcas ready to be merged? In principle yes, I just wanted to x-check two things with you. 1. I did not implement any explicit logic to forbid the usage...

> For the simultaneous squashing / masking request: I would simply throw an exception with corresponding message. Ok > At the moment I would disable the squashing by default, so...