Add modelSplitLevel as an option

[Robin Tesse]

To use the new module DataUproot without explicitly importing ROOT, we can add an option modelSplitLevel (same structure as for samplerSplitLevel). The default value is 1 to be compatible with pybdsim (pybdsim.data.Getmodel()).
With this option, we can access the model using the following:

from pybdsim.DataUproot import BDSimOutput
data = BDSimOutput('root_file.root')
data.model.df

‌

[Robin Tesse]

Here is a proposal from Laurie:

We can add the option, and we should also add maybe a 'superior' option called uprootCompatible (for example), and when we turn that option on, we then turn on the others required for the split level.
Also, we can warn the user that using a splitlevel = 1 for the samplers can increase the simulation time, especially if the user has a very large number of samplers (e.g. for optics).

‌

[Laurie Nevay]

Thumbs up from me.

What bit of the data is it that uproot has trouble with? Is it an std::vector in a branch? Or a vector of custom class (not a basic type)?

[Robin Tesse]

I think it has to do with the stored data being in the form of a custom class vector. Therefore, we need to change the split level, to have a vector of known types (float, int, ...).

[Laurie Nevay]

Merged in uproot_feature (pull request #6)

Resolves #349

Approved-by: Laurie Nevay

→ <<cset a540078da837>>