Add a calculator to integrate range from energy loss

[amandalund]

As we refactor the physics table construction we'll probably move toward replacing the Process::step_limits with macro_xs and energy_loss (initially the process-integrated energy loss, but eventually letting each process provide its own and summing them) and then integrating the range from the energy loss ourselves. This adds a calculator for integrating the range from a dedx grid and uses it in the physics grid construction.

I'm leaving this as a draft for now because it changes a lot of test results. I think this could be because of differences in the spline interpolation, but it would be good to understand and discuss what we want to do before I go too far with it. I added a failing test comparing range values calculated from the imported dedx to the imported range to get a rough idea of the magnitude of the differences.

[github-actions]

Test summary

 4 573 files   6 999 suites   13m 21s ⏱️
 1 715 tests  1 709 ✅  6 💤 0 ❌
24 556 runs  24 472 ✅ 84 💤 0 ❌

Results for commit 249bd09.

♻️ This comment has been updated with latest results.

[amandalund]

I've updated this to use spline interpolation, and the calculated range values match the imported ones. The only tests where the results changed are the mock tests (since we're calculating the range in the physics params instead of constructing it in the mock process) and a few float tests.

One thing I noticed is spline interpolation does not work well if there are a small number of grid points and the scale of the x grid is large. Hopefully Geant4 always uses a dense enough grid, and at least this should give us a little more control of the calculation.