HltEfficiencyChecker: HLT2 reconstruction configuration becomes configurable; set default to fastest; run_reco by default; more accurate HLT1-filtering output rate

Built on top of !122 (merged). Goes with Moore!2419 (merged)

This MR does a number of things:

• Makes running of the HLT2 reconstruction the default,
• Requires now that you specify a hlt2_configuration in the yaml wizard, which can either be fastest or light_without_UT. These correspond to make_fastest_reconstruction and make_light_reco_pr_kf_without_UT respectively.
• Make the former ("fastest") the new default in the options and documentation (closes #35 (closed)).
• Replace --using-hlt1-filtered-MC with --input-rate. The former used to be a flag that would internally set the input rate of a min bias sample to 1 MHz (expected HLT1 rate). Recent discussions in #42 (closed) have pointed out that both the two usual min. bias samples are significantly far away from 1 MHz. So now provide the rate directly, and set it to 1.65 MHz in the case of upgrade_minbias_hlt1_filtered.
• Added new example options files hlt2_rate_example_withGEC.yaml and hlt2_rate_hlt1latestBWdiv.py to showcase how to get reliable HLT2 rates out of the UpgradeHLT1FilteredWithGEC min. bias, and also the latest HLT1 BW division samples.

Showcased in RTA/DPA General on 16/06/23. Slides.