v0.2.0

This is a breaking change release; please read carefully.

New Features:

• add referencedPopulations; given a nodeset name, it will recursively find all the names of the populations referenced.

Breaking changes:

• Node Population and named populations:

  According to https://github.com/AllenInstitute/sonata/blob/master/docs/SONATA_DEVELOPER_GUIDE.md Node populations and their names are implicitly defined in the Node Set namespace, and needn’t be declared explicitly.
  This doesn't say, however, that it is forbidden to use that name. Thus, if another different nodeset is given the population name, it is used instead.

• Normalizing paths:
  Previously, the absolute paths weren't being resolved for SimulationConfig::output when the default values were chosen.
  This means that, for instance, the spike_file may be resolved to where a command was run, rather than the simulation_config.
  Thus, some fixups may be required for simulation_configs.json so they correctly reference the spike_file.
  For examples of how this may impact real applications, see the changes required in:

  • snap: openbraininstitute/snap#26
  • neurodamus: openbraininstitute/neurodamus#580

What's Changed

• CI check: verify macOS build on current master by @cattabiani in #57
• Fixed absolute paths when SimulationConfig::output is defaulted by @mgeplf in #61
• Add forward iterator to Selection by @cattabiani in #60
• Make sure absolute SimulationConfig output paths are correctly based at output_dir by @mgeplf in #62
• Add ElectrodeReader for SONATA electrode weight files by @cattabiani in #56
• remove useless changelog by @cattabiani in #64
• Switch away from flatten/unflatten when updating config to expandVariables by @mgeplf in #66
• Make population name be an implicitly defined nodeset by @mgeplf in #40
• Node set populations referenced by @mgeplf in #67
• Scikit build by @mgeplf in #65

Full Changelog: v0.1.37...v0.2.0