Skip to content

3.0.0

Compare
Choose a tag to compare
@hernando hernando released this 20 Mar 10:41
· 78 commits to master since this release
  • 250:
    Added new overloads of brain::Simulation::getGIDs to get random GID sets.
  • 249:
    • brion::CompartmentReport::loadFrame changed to return brion::Frame
    • Implemented missing brain::CompartmentReportView::getReport
  • 239:
    • Added two query options to configure chunking and chunk cache size when
      opening/creating SONATA compartment reports.
    • Added a new overload of brion::CompartmentReport::write to allow
      optimizations for full frame writes.
  • 229:
    • Added a method to brain::Circuit to retrive custom node attributes. This is
      only supported in C++ for SONATA circuits.
    • Implemented morphology recentering for SONATA circuits.
  • 227:
    Read/write support for SONATA compartment reports. This file format replaces
    the old file format when writing reports.
  • 225:
    Addition of some new methods and classes:
    • Added the class Simulation for accesing the contents of simulation
      configurations
    • Circuit::getSource
    • Circuit::getMorphologyNames
    • Made CompartmentReport and SpikeReporReader movable
    • Soma::getMaxRadius
    • neuron::Section::getNumSamples
    • neuron::Section::operator[]
    • Support morphologies with single point somas
    • Added default Synapses::iterator constructor
    • Changed Synapses and Synapse memory magement implementation to allow
      use in Python wrappings of client code.
  • 222:
    Implemented basic support for reading circuit in SONATA file format
    • The implementation is limited to node, no edges are read
    • Only one population and one node group are read, the rest are ignored.
  • 204:
    Fixed brain::Circuit methods to not crash with empty gids lists.
  • 191:
    • Optimizations in brain::CompartmentReport opening, only the metadata is
      read.
    • Removed compartmentCount and cellCount from brain::CompartmentReportMetadata
      because they can't be figured out efficiently in all report types.
    • New function getFrameSize added to CompartmentReportMapping to replace
      compartmentCount.
    • New function getCellCount added to CompartmentReport to replace cellCount.
  • 180:
    Bugfixes in compartment reports:
    • The report converter was messing up the order of the data when writing to
      .h5 from an out-of-order binary.
    • Timestamps reported were wrong for reports not starting at 0.
    • Read time and data unit attributes from H5 reports is they exist.
    • Do not assume the first section is always present post-processing the
      mapping in H5 reports.
    • Avoid crashing when calling getNumCompartments from Python with and out of
      bounds index.
    • Corrections in compartmentReport comparison code.
  • 176:
    Removed the dependency on HDF5++ and use HighFive instead.
  • 172:
    • zeroeq::Server/Client based morphology loader.
    • New brion::Morphology read API and removed write API (see doc/feature/morphologyRead.md)
    • Removed apical point API
    • morphologyConverter can only write H5 V1.1 morphologies
  • 171:
    Ban obsolete morphology repair stage from API: The stage parameter from
    certain Morphology methods has been removed. In the case of h5v2 morphologies,
    the implementation will pick the most complete repair stage.
  • Performance optimizations in compartment reports and synapses
  • Python-wheel generation for Linux and Mac OS/X (versions 3.5, 3.6 and 3.7): scripts and CMake fixes
  • Use gitsubmodule for Pydoxine
  • Make ZeroEQ optional and off by default