NuMI Flux Sysematics Update

[woodtp]

This PR tied closely with sbndata PR#4 Update NuMI flux file to latest version.

The updates here are necessary to access data correctly from the flux file in the above PR, as the internal structure of the file has undergone several revisions.

Specific Updates:

• The file name referencing the file from sbndata that needs to be accessed has been moved from a local variable of NuMIFluxSyst::Shift to a private member variable of the NuMIFluxSyst class as a whole. My opinion is that it will be easier to locate and update down the line, this way. Although, I'm not a fan of having the filename hard-coded into the class, as it has to be updated separately in the new NuMIPpfxWeight class, but I don't have a better solution at the moment.
• @brucehoward-physics wrote a NuMIPpfxWeight class and corresponding Var to facilitate extracting the NuMI flux weights, so I've incorporated that into this PR. Just as a simple test that it's working as expected, I've pulled a small distribution of flux weights:

Other modifications:

• attenuation has been renamed to att
• beam_divergence --> beam_div
• horn1_position_xy separated into horn1_x and horn1_y
• horn_current --> horn_current_plus (the -2kA sample found to be consistent with statistical fluctuations, so we dropped it)
• horn2_position_xy removed
• beam_spot_size --> beam_spot
• beam_shift_minus_y and beam_shift_plus_y split from beam_shift_xy (x-shift samples dropped; consistent with stat. flucts.)
• B_field removed

Again to check that this update still works as expected, below are a selection of true neutrino energy spectra varied according to the various systematics.

[brucehoward-physics]

Thanks for this PR Tony -- It looks like I'm the only person tagged as a reviewer and since some of it is stuff I helped put together, I guess I'm thinking we should add another reviewer 😅

@PetrilloAtWork is this one for you or... do you have someone else in mind? Maybe @jedori0228?

[jedori0228]

Hi, I merged this PR and checked NuMIFluxSysts and NuMIFlux work as expected. Few validation plots added:

[woodtp]

I've made another commit, which changes the following:

1. New feature to the NuMIFluxSyst class allowing the user to specify a path to the fluxfile. It will default to the one defined by the SBNDATA_DIR envvar.
2. Additional systematic has been added to account for differences between the older beam geometry and the newer 1 MW upgrade to NuMI.
3. The statistical uncertainty on the flux prediction was omitted in previous versions, so I've added a few lines to load it.
4. Some minor changes to which histograms are loaded for beam focusing systematics. I realized that these were reading in the fractional uncertainties instead of the systematic shifts, which would cause an improper accounting for correlations between bins.