Reduce SMOG2 sprucing bandwidth

This MR significantly reduces the bandwidth used by the SMOG2 sprucing lines.

The main changes are the following:

• The SMOG2 sprucing lines that are currently merged all have persistreco = True. To reduce bandwidth, many of the sprucing lines are now changed to the selective persistency model with persistreco = False.
• Lines that are kept persistreco = True are prescaled, with the exception of the dimuon lines which have extremely low rates and therefore very small bandwidth.
• Some sprucing lines are removed, as the corresponding HLT2 lines are being moved from FULL to TURBO stream (which is why this MR depends on !3266 (merged))

Current bandwidth with merged sprucing lines: 0.9 GB/s

New bandwidth with this MR: 0.13 GB/s (rate 3.3 kHz). Rates for individual lines are heresmog2_sprucing_rates.txt

I am aware this bandwidth is still slightly large. However, IFT is unique in that we do not have TurCal space, therefore we have some calibration lines included here with persistreco = True. Also these rates were computed with pAr + pp MC, for smaller gases such as hydrogen or helium the corresponding rates will be much smaller.

Finally, this MR does not modify any HLT2 lines. The changes in this MR that show changes in HLT2 are all due to !3266 (merged), which will be merged imminently. Once it is merged, the remaining changes will only be in sprucing lines.

@samarian @thboettc @oboenteg