Velo-SciFi track reconstruction based on SciFi seeding and Matching algorithms

Goes together with Moore!1733 (merged)

Long track reconstruction based on SciFi seeding and Velo-SciFi matching. Information from the UT is not included for the moment. The algorithms can be found in:

• device/SciFi/hybridseeding for the SciFi seeding. The SciFi standalone reconstruction is based on two main algorithms, the seed_xz and seed_confirmTracks. The HLT2 implementation of the HybridSeeding can be found here: https://gitlab.cern.ch/lhcb/Rec/-/blob/master/Pr/PrAlgorithms/src/PrHybridSeeding.cpp
• device/track_matching for the Velo-SciFi matching. The main reconstruction algorithm is track_matching_veloSciFi and is adapted from HLT2 (https://gitlab.cern.ch/lhcb/Rec/-/blob/master/Pr/PrAlgorithms/src/PrMatchNN.cpp )

The output of the track-matching algorithm is long tracks and has been adapted to have the same format as for the forward algorithm, in order to facilitate integration in the HLT1 sequence. The hlt1 reconstruction with seeding+track-matching can be enabled by using the hlt1_pp_matching or hlt1_pp_matching_validation sequences.

More information on the current strategy and performance of the SciFi seeding + matching algorithms can be found in the 17/01/2022 WP2 meeting: https://indico.cern.ch/event/1116110/

Alternative (original) parallelisation scheme for the SciFi seeding can be found in https://gitlab.cern.ch/lhcb/Allen/-/tree/lohenry-uv-2, with the first complete porting to GPU by @bjashal and @jzhuo in https://gitlab.cern.ch/lhcb/Allen/-/tree/HLT1Seeding_v1 . Both branches are kept to allow development on that parallelisation scheme.

Developments on the UT-seeding and UT-SciFi matching are ongoing in: https://gitlab.cern.ch/lhcb/Allen/-/tree/dev_trackmatching_christina

@lohenry @lcalefic @ahennequ @gligorov