Extend nightly tests of Moore's "generic" Hlt2 lines to include signal samples rather than just min bias

This is prompted by a suggestion from @sstahl in the 15/3/2022 RTA-WP3 meeting, in the context of a discussion about the recent drop in muon ID efficiency.

Currently there is test_hlt2_reco_plus_thor_selections_fastest.qmt, which uses min bias input.

This makes use of the generic lines, which are intended to be stable/static for use in testing, unlike the real WG lines that are regularly tweaked/tuned.

We should make sure that there are tests that cover some key combinations of signal topologies and selections of them. E.g. charged_hadron+charged_hadron, something with muon, something with electron, something with V0, something with photon etc...

What do we currently have in generic_lines?

• Something with muons: Hlt2BsToJpsiPhi_JPsi2MuMu_PhiToKK_Line
• Something with electrons: Hlt2BsToJpsiPhi_JPsi2ee_PhiToKK_Line
• Something with DD V0: Hlt2XiczToLPimPip_DDLine
• Something with LL V0: Hlt2OcToXimKmPipPip_LLLLine
• Combination with photon/pi0. (we only have an inclusive photon selection)

As a side note I think these lines should be renamed to start with Hlt2_generic_... to distinguish them from the real WG lines for production.

@rmatev @sstahl @poluekt @nskidmor @decianm @cmarinbe @mramospe @agilman @rjhunter @graven @gligorov