Introduce handling of new Run3 layouts to master

Hi,

this MR is a manual sweep of all developments in the 21.3 branch which were made in the past to allow to use the new Run3 layouts ATLAS-R3-2021-01-00-00 (asymmetric layout with NSW only on Aside) and ATLAS-R3S-2021-01-00-00 (symmetric layout with NSW on both sides) in simulation/digitisation/reconstruction.

The MRs into 21.3 which have been swept here manually are: !26471 (merged), !26102 (merged), !26028 (merged), !25963 (merged), !25736 (merged), !25732 (merged), !25410 (merged), !25167 (merged), !24562 (merged), !24467 (merged), !24013 (merged) and !23363 (merged).

Notice that this MR will break both the overlay test (RunTier0Tests.py -o -n) and the q221 and q431 tests (RunTier0Tests.py -n) because of the following reasons:

• The sweep of !24013 (merged) fixes a bug in the MuonSegmentRegionRecoveryTool which was excluding the EI MDT hits also when running with the Run2 layout. Correctly using those hits in the recovery changes the reconstruction output.
• The sweep of !24467 (merged) changes the output of the q221 test since the data_SampleMap of the CSC chambers in CscDigitizationTool::FillCollectionWithNewDigitEDM was previously sorted by 'geometrical hashes' (not the position of the CSC identifier in the vector of identifiers), which results in a different order of reading the entries from data_SampleMap and thus in a different order of calling CLHEP::RandGaussZiggurat when calculating the noise values. This slightly changes the digitisation output although there is no real change of physics.
• The overlay test will fail due to a difference in the CSCs. This is because previously, the overlay test for the CSCs was based on the 'geometrical hashes' (not the position of the CSC identifier in the vector of identifiers). In my opinion this has some disadvantages, e.g. the hashes are not continuous and their range is not starting from 0 and going up to the maximum number of hashes-1. Furthermore, in case of an asymmetric detector with only half of the CSC chambers there is a huge gap of hashes which needs to be accounted for correctly. This is why I am strongly in favour of using real hashes (the position of the CSC identifier in the vector of identifiers) for the CSCs as it is done for all other muon subdetectors. However, this requires that we create new input files for the overlay test which already contain the correct information.

Important: ATLAS-R3-2021-00-00-00 is not supported anymore after this MR is merged (since this layout contains both NSWs and CSCs on both sides which is not possible in reality)!

Adding @jchapman for simulation

Adding @rosati, @wleight and @cchau for muon software

Adding @alaurier who might want to follow here because of the overlay

Adding @pboescho who is in charge of the migration of the MuonIdHelperTool and might have some overlap with this MR

Adding @sroe and @emoyse for having a look

Best, Nico