Introduction of SegMIP BDT + other improvement on ECAL veto

[tvami]

I am updating ldmx-sw, here are the details.

What are the issues that this addresses?

Resolves #1287
Resolves #1305

We have shown the results of the software changes already at a SWAN meeting. The latest BDT that's now trained on the full E14 trigger skimmed data is included now too. This is to be shown at a SWAN meeting.

Future improvements in another PR will include factorizing the MIP tracking out of this class.

Check List

• I successfully compiled ldmx-sw with my developments
• I ran my developments and the following shows that they are successful.

[tvami]

We expect the tests to fail, both logs and histogram comparisons.

[tvami]

Error in TFile::TFile: file hist.root does not exist

this should be not related

[tvami]

Nvm, the issue is in the prev line

ERROR: ONNX model file '/home/runner/work/ldmx-sw/ldmx-sw/install/data/Ecal/segmip.onnx' does not exist.

Let me fix this typo

[tvami]

Lots of diffs in the validation, many actually come from a simple change that now the electron beam energy in the ecal veto is configurable, now the plots are correct (meaning consistent) for the signal. For the ecal pn, we need to change the config to be at 8 GeV. I also think we should run tracking in the validation configs. I can do that in another PR.

I propose to do that next and have new golds / subversion change after that so then we can launch a BDT skim.

[tomeichlersmith]

As usual, I mainly have some C++ technical comments. I think this implementation looks good and resolves two ( 2️⃣ ❗ ) issues so I'm happy its going to be merged.

I will continue to grumble about #1306 but leave that for future work.