Xcoll release 0.8.0

What's Changed

• Improvements to the Geant4 interface:
  • Ensuring compatibility with the latest BDSIM version (in the develop branch.
  • More robust implementation of Geant4 environment compilation, compatible with different BDSIM versions.
  • Neutral and unstable particles (when using the latest BDSIM version) can be handled.
  • Large speed-up in tracking (Xcoll postprocessing on returned particles was very inefficient).
  • New collimator type: Geant4CollimatorTip which has a thin layer of different material on the jaw (code by @gbrogginess).
  • Automatically adapt the mass of the reference particle to the mass used by Geant4, if known.
  • One can select which particle types to return from Geant4.
• Improvements to the LossMap:
  • Importing from a json failed due to a check for Geant4-collimators in a non-existing line. Fixed.
  • Plotting now can make the distinction between warm and cold losses, and can optionally provide a zoom plot. These can be provided as arguments to LossMap.plot() or can be set on the line in general using line.metadata['collimation'] = {'cold_losses': [[s1_start, s1_end], [s2_start, s2_end], ...], 'warm_losses': ..., 's_range': {betatron : [s_start, s_stop], ...}}. Note that either colde_losses or warm_losses has to be provided, and that the s_ranges are named ranges to able to do e.g. plot(zoom='betatron').
  • Some fixes in the aperture losses and in the different properties.
  • Aperture interpolation failed when time-dependent variables are present and activated (e.g. during an RFSweep). These are now temporarily deactivated during interpolation at LossMap instantiation / particle adding.
• Fix in FLUKA name for MoGR64.
• The CollimatorDatabase now accepts the s-position as arguments for each collimator (to specify the position when no Marker or Drift is present in the line for that collimator).

Develop Updates

• Bugfix: old and/or malformed config files were blocking.
• Geant4: after tracking, the C++ code auto-sizes the return array (the capacity check is done in Xcoll).
• Changed API of beam element transformations active by @szymonlopaciuk in #164.
• Update monitor to handle zero emittance case by @JPTS2 in #170.
• Emit moni zero emit by @JPTS2 in #171.
• The built-in material database is built dynamically at import, but this was making the package import very slow (due to cascading look-ups). This is now fixed.
• When installing a collimator, the check if the element name already exists is sped up.
• The Constants class now allows to define extra metadata data to its elements. Used e.g. for particle masses where besides the name and the mass, also the pdg_id is provided.

New Contributors

• @JPTS2 made their first contribution in #170

Full Changelog: v0.7.2...v0.8.0