Ignoring overlap TCs (pileups) option in MLT

[ArturSztuc]

Added an option to ignore TCs that would overlap with a TD, rather than merging them in. This should help to deal with the beam pileup events in np04. I also took the opportunity to cleanup some of the MLT code, but not by much.

This PR goes together with: https://github.com/DUNE-DAQ/daqconf/pull/483

There's now a new daqconf option, mlt_ignore_overlapping_tcs, set to false by default. Use it like so:

"trigger": {
    "mlt_ignore_overlapping_tcs": true,
    "mlt_merge_overlapping_tcs": false
}

Note that mlt_merge_overlapping_tcs has to be set to false (we cannot merge overlapping TCs and ignore them at the same time...), with an appropriate error printout at conf time if the user does try.

Tests done:

1. 3ru_3df_multirun_test.py passes

2. Using custom trigger candidate maker with options:

use_custom_maker: true, 
ctcm_trigger_types: [7, 8],
ctcm_trigger_intervals: [2000, 1000],
ctcm_timestamp_method: kTimeSync

to make fake HorizontalMuon triggers spaced 2k ticks apart (window of 2k), and MichelElectron triggers spaced 1k ticks apart (window of 2k). Ran it in three mlt settings, 60s each:

1. "mlt_merge_overlapping_tcs":false, "mlt_ignore_overlapping_tcs":false to test no merging no overlap, generated 48Mb file, lots of overlapping trigger records, all with the same window.
2. "mlt_merge_overlapping_tcs":false, "mlt_ignore_overlapping_tcs":true, to test ignoring the overlapping TCs, 15Mb file, no overlapping TC records, all with the same window.
3. "mlt_merge_overlapping_tcs":true, "mlt_ignore_overlapping_tcs":false, to test merging the overlapping TCs, 5.6s, no overlapping TC records (all timing), all merged into kTiming with large window.

Plots: Trigger rate per trigger type. First run no merging no ignoring, every TC made into separate TD, MichelElectron at ~55Hz, HorizontalMuon at ~27Hz. Second run no merging but ignoring overlapping TCs, MichelElectron at 14Hz HorizontalMuon at ~5Hz. Third run, merging TCs, all merged into Timing trigger. In hindsight, should have removed the timing trigger...

Number of TCs & TDs dealt with by the MLT. In second run we see the same number of TCs going in, but drop in number of TDs made. Unfortunately, although MLT tracks number of TCs dropped (not merged, ignored), there is no Grafana display for it.