refactor(Examples): clean up magnetic field handling

[msmk0]

Clean up the magnetic field types and options in the examples:

• Magnetic field related headers are now consistently located in the ActsExamples/MagneticField include directory and all code is in the ActsExamples namespace. Before, the code still had the old Examples/Plugins/... layout and inconsistent or missing namespacing inherited from the old framework.
• Move the magnetic field variant type ActsExamples::MagneticField into a separate header.
• Magnetic field options have a well-documented order of preferrence (constant field overrides field map). If no field options are given, a null field is created, e.g. to support telescope-like detectors without a field. Expected units are clearly communicated as part of the parameter name, similar to what is done in other algorithms.
• The field map readers are private implementation details and are not publicly available anymore.
• Detector construction and magnetic field setup are now completely orthogonal: all magnetic fields can be setup for all detectors.

On possible interactions with other open PRs

This is somewhat related to #675 as the magnetic field variant becomes obsolete afterwards. I hope that with the clean up in this PR the move to interface-based magnetic fields will be a bit easier, as the ActsExamples::MagneticField variant type could be replace by something like std::shared_ptr<const Acts::BFieldProvider> without having to rename it. This changeset is part of the ongoing Fatras cleanup.

[codecov]

Codecov Report

Merging #696 (6c77383) into master (bd9115d) will not change coverage.
The diff coverage is n/a.

@@           Coverage Diff           @@
##           master     #696   +/-   ##
=======================================
  Coverage   49.01%   49.01%           
=======================================
  Files         325      325           
  Lines       16574    16574           
  Branches     7744     7744           
=======================================
  Hits         8124     8124           
  Misses       3019     3019           
  Partials     5431     5431           

---

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update bd9115d...73485e1. Read the comment docs.

[XiaocongAi]

Hi @msmk0 , it looks all good to me. So I will approve. Meanwhile, I'm wondering if might make sense to move the Examples/Detectors/MagneticField to Examples/MagneticField?

[XiaocongAi]

It looks like this line can be omitted.

[msmk0]

Yes, it could be omitted. It is here intentionally so both parts of the if/else statement have the same structure and it is easier to understand.

[paulgessinger]

Can we hold this until I have tested how many merge conflicts I get? I'll try to do that tomorrow.

[msmk0]

@XiaocongAi Thanks for the review. I'll update the points you raised.

Can we hold this until I have tested how many merge conflicts I get? I'll try to do that tomorrow.

No worries. As previously discussed, I leave the final merge decision/order up to you.

[paulgessinger]

Ok it's not that terrible. I have a rebased branch ready that I believe still works.
So that'll adapt the variant on top of this PR, while still getting the benefits of this PR's other changes.

I'm fine with merging this then.

[msmk0]

@paulgessinger That is nice to hear.
@asalzburger Do you want to review as well or should we go ahead. @XiaocongAi already approved this PR.

[msmk0]

Meanwhile, I'm wondering if might make sense to move the Examples/Detectors/MagneticField to Examples/MagneticField?

@XiaocongAi I missed that. I would leave if there for now. The magnetic field is detector-related (to some extend) and there is not really that much code to warrant a completely independent module.

[paulgessinger]

The header becomes relatively useless anyway with my upcoming changes so I'll probably remove it entirely.