OT-OT Duplicate Cleanup

[GNiendorf]

Small PR that adds hit-matching checks and loosens some duplicate cleaning cuts to reduce the number of OT-OT duplicates (e.g. T5-pT5, T5-T5, etc.) by 85%. Most duplicates are pLS-OT type, but this PR fixes some low-hanging fruit.

Here was the original prompt and a link to the original plan if curious: https://gist.github.com/GNiendorf/ee534ab10da87b37e3b7175bb32fe932

I'm curious if you can go into plan mode, make a bunch of documentation about how the duplicate cleaning works (both within a TC type and cross cleaning over the different TC types) and then using that documentation, our goal is to make a ntuple sample of like say 300 events with the PU200RelVal sample and see which duplicates exist in the TC collection that are not pLS related. So just OT-OT duplicates like pT3-pT5 or pT5-pT5 or T5-T5 or like any variation of those but no pLS-T5 or pLS-pT5 like I don't care about those for now. My thought is I want to know what caused those OT-OT duplicates to slip through the code, like what cuts failed? Were they far in eta/phi space and that caused those duplicate pairs to slip through? I want to know which specific cuts caused it, and how to fix it so that we have 0 OT-OT duplicates in the 300 event sample ideally with no significant decrease in efficiency (both overall efficiency and displaced efficiency very important) from the fix that you figure out. Can you do this? You will have to see after you do the duplicate study of all the duplicate cleaning + cross cleaning kernels that we have enough info in the ntuple to answer these questions, and if not first add that info to the ntuple in the write lst ntuple .cc file or whatever. Then we generate the sample see which cuts failed and figure out how to fix it without decreasing efficiency (total and displaced track).

[GNiendorf]

run-ci: all

[github-actions]

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

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

Here is a timing comparison:

   Evt    Hits       MD       LS      T3       T5       pLS       pT5      pT3      TC       Reset    Event     Short             Rate
   avg     28.0    323.1    242.2    133.0     45.9    702.0     10.8    114.6    116.5    210.5      0.1    1926.7    1196.7+/- 289.1     594.8   explicit[s=4] (target branch)
   avg     27.7    324.5    241.2    131.9     45.4    677.4     10.9    114.0    115.5    194.2      0.1    1882.8    1177.6+/- 284.2     591.0   explicit[s=4] (this PR)

[github-actions]

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

OOTB All Tracks

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