test first pass pLS cleaning with 2 hits

[slava77]

3 hit requirement seems too restrictive; at least as a reference I wanted to check if performance changes for the better or not by reducing the check down to 2 hits

[slava77]

/run standalone

[github-actions]

The PR was built and ran successfully in standalone mode. Here are some of the comparison plots.

The full set of validation and comparison plots can be found here.

[slava77]

The PR was built and ran successfully in standalone mode. Here are some of the comparison plots. ...

The comparison went off scale (someone may want to contribute to improving the comparison script).
Looking at another test, e.g. https://raw.githubusercontent.com/SegmentLinking/TrackLooper-plots-archive/main/PR355_0f6afe7_standalone/dup_eta_comp.png
the baseline duplicates in the peak are just around 0.14; around x2 reduction.

[slava77]

/run cmssw

[github-actions]

The PR was built and ran successfully with CMSSW. Here are some plots.

OOTB All Tracks

The full set of validation and comparison plots can be found here.

[slava77]

it looks like the improvement is visible, although smaller with cmssw as well.

@ariostas
we usually keep "blue" as a reference in MTV plots so that the ratio is new/ref.
It might be good to update (it's just a matter of swapping the order of inputs to the validation script).

[ariostas]

we usually keep "blue" as a reference in MTV plots so that the ratio is new/ref.

Ah okay, I'll fix that.

[VourMa]

Doesn't seem like this PR is doing any good in the Patatrack case (quad seeds from Patatrack + triplet seeds from legacy pixel seed algorithm).

Patatrack in HLT (blue) vs. Patatrack in HLT with the update from this PR (red):

All plots here:
https://uaf-10.t2.ucsd.edu/~evourlio/SDL/PR365/patatrack_LSTSplitpTTCandpLSTCandT5TC_defaultTrkIdBuiltTracks_VS_2HitsInFirstPassDupClean/

[slava77]

(to note from the meeting) perhaps this can be kept in the |eta|>2.5
but then I guess a test with triplets in post-LST pLC-TCs is needed