Skip to content

Latest commit

 

History

History
418 lines (379 loc) · 19.3 KB

Changelog.md

File metadata and controls

418 lines (379 loc) · 19.3 KB
Raw

Changelog {#Changelog}

Release 3.3.8

  • 344 3.3.8 Release preparation
  • 343 Update dependencies to latest version and fixes legacy HDF5 compartment reader

Release 3.3.7

  • 342 3.3.7 Release preparation
  • 341 Updates GLM submodules from 0.9.9.3 to 0.9.9.8

Release 3.3.6

  • 339 Set numpy minimun version to 1.19.5 and release preparation

Release 3.3.5

  • 332 Improves submodule handling. Fix morphio target aliasing

Release 3.3.4

  • 327 3.3.4 Release preparation
  • 326 Add extra checks to choose the appropiate synapse handling class

Release 3.3.3

  • 325 3.3.3 Release preparation
  • 324 Sort returned SONATA report values by node ID
  • 323 Adds warning message about using Brion to read SONATA format
  • 322 Update copyright and funding notice

Release 3.3.2

  • 321 3.3.2 Release preparation
  • 320 Allow tabs '\t' as key - value separator in BlueConfig
  • 319 Fix wrong frame index calculation from report timestamp, drop support for python < 3.6
  • 318 General submodule update
  • 317 Custom BlueConfig file support for multipopulation
  • 316 Fix BlueConfig class to follow documentation about morphology type

Release 3.3.1

  • 314 3.3.1 Release preparation
  • 313 Improve error message for GIDs out of range in report
  • 312 Handling MorphologyType from BlueConfig
  • 311 Handle 0 based GIDs on sonata files
  • 310 Miscelaneous fixes and updates
  • 309 Use env vars for setting PYTHON3_ROOT_DIR (wheel distribution)
  • 307 Fixes version number typo
  • 306 Fix for wrongly load check in synapse positions

Release 3.3.0

  • 305 Preparation for 3.3.0 release.
  • 304 Rename brain python package to brion.
  • 303 Improve SONATA syanapse support.
  • 302 Fix glm targets not installed properly.
  • 300 Initial commit for python wheel distribution package support.
  • 299 Removing unused functions from old SONATA circuit implementation.
  • 298 Support for relative paths in BlueConfig/CircuitConfig.
  • 297 SONATA circuit support.
  • 295 Misc. fixes and updates.
  • 293 Fix mvdtool bug when mecombo file is empty.
  • 290 Replacing morphology readers and tests with MorphIO.
  • 289 Update README.md.
  • 288 Update MVDTool submodule.

Release 3.2.0 (10-06-2020)

  • 287 Fix wrong file access mode when opening spike report for reading.
  • 286 Fix cmake build with newer boost, python and cmake version. Fix tests.
  • 285 Attempt to fix intel compile errors.
  • 284 Fix installation structure from CMake.
  • 283 Preparation for 3.2.0 release.
  • 282 Few fixes related to glm.
  • 281 Strips Pydoxygen, Servus and CMake/common.
  • 280 Bumps libsonata submodule.
  • 279: Removed lunchbox library from Brion.
  • 276: Removed vmmlib library from Brion. Replaced with GLM.
  • 275: Removed Keyv library from Brion. This change also means the end of support for BlueBrain's key value map based reports.
  • 273: Fixes to improve compilation against libsonata.

Release 3.1.0 (07-01-2020)

  • 269: Introduce SONATA support
  • 263: Detect start.target in CircuitPath
  • 255: Adding Python 3.7 and removing 3.4 from wheels

Release 3.0.0 (20-03-2019)

  • 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.

Release 2.0.0 (30-05-2017)

  • 159: Changed mapping index according to Pandas requirements for indexing.
  • 157:
    • Add Synapse.gid() python wrapping
    • Improve synapse attribute reading performance
  • 154:
    • Improved help and command line inverface for spikeConverter and compartmentConverter.
    • Timestamps in brion::CompartmentReport changed from float to double.
    • New implementation of binary compartment report reader using POSIX AIO.
  • 153: Updated Python wrapping of SpikeReportWriter to handle nympy arrays of spikes
  • 152: Added the function brain.neuron.compute_morphological_samples to help getting sample positions from morphologies
  • 151: brain::Circuit accepts full paths to mvd files in CircuitPath
  • 150:
    • Added a new method in brain::Circuit to obtain synapses for afferent projections from outside the circuit (e.g. thalamocortical projections).
    • Support for Projections sections in brion::BlueConfig
  • 141: Fixed documentation to specify that morphology section samples report diameters and not radii.
  • 130,138: Added brain::CompartmentReportReader and its Python wrapping brain::CompartmentReport : python API v2
  • 136: Implement readable null:// compartment report for benchmarking
  • 120, 131:
    • New SpikeReport API.
    • Reimplementation of the high level SpikeReportReader. The new implementation uses numpy arrays to provide the requested spikes.
  • 126: Add erase for map compartment reports
  • 122: Support loading of individual neurons for binary and map compartment reports
  • 121: Performance optimizations for map and binary compartment reports, fix sub target loading of map compartment reports
  • 119: Optimized brain::Circuit::loadMorphologies for slow stat'ing filesystems (GPFS)

Release 1.9.0 (09-12-2016)

  • 117: Changes and fixes in Brain Python module:
    • Functions that take gids now preserve the iteration order of the input in the output arrays/lists
    • Replaced Python sets with numpy arrays in functions returning GID sets
    • Circuit::getXYZNames functions renamed to Circuit::getXYZTypeNames
  • 113: Support for old circuits containing only synapse center positions
  • 110: Break PersistentMap out into keyv::Map
  • 107: Added Sphinx generated documentation of the brain python module
  • 102: Use PersistentMap for caching synapse positions loaded from brain::Circuit
  • 94: Fixed SWC morphology parser for morphologies with soma contour. The parser was creating invalid soma sections when the first order sections where connected to arbitrary soma sample points.
  • 89: Python wrapping of brain classes
  • 88:
    • Brain namespace enums made strongly typed
    • Fix for brain::Section::getSamples for sections where the first point appears repeated
  • 85: Implement brain synapses specification
  • 84: Brain synapses specification
  • 83: Add brain::Circuit::getRandomGIDs()
  • 82: Improve performance when reading synapse attributes from non-merged files
  • 81: Fix GID out-of-bounds handling for MVD3 in brain::Circuit
  • 79: Use PersistentMap for cache in brain::Circuit::loadMorphologies(); add thread safety w/ synapses cache

Release 1.8.0 (30-06-2016)

  • 77: Add brain::Circuit::getMorphologyTypes() and brain::Circuit::getElectrophysiologyTypes()
  • 75: Implement morphology version 1.1 specification
  • 74: Remove deprecated enums and functions:
    • CompartmentReport( const std::string&, const GIDSet& ) and CompartmentReport( const std::string&, const brion::CompartmentReportFormat, const bool ) constructors; use brion::CompartmentReport::CompartmentReport( const brion::URI&, int, const brion::GIDSet& ) instead
    • enum CompartmentReportFormat; use brion::BlueConfig::getReportSource() instead
  • 71: Transparent caching of Synapse loading. See Lunchbox#263 for configuration.
  • 69: Speedup brain::Circuit::getGIDs(), add brion::BlueConfig::getTargetSources()
  • 63: Moved old BBPSDK/Monsteer spike report to Brain (pending refactoring)

Release 1.7.0 (24-03-2016)

  • 56: Improved target parser
  • 55: Add basic provenance metadata for written HDF5 compartment reports
  • 49: Added the method brain::Circuit::getRotations
  • 46: Fixed a bug in enum to string conversions affecting morphologyConverter.
  • 45: Made targets optional.
    • brain::Circuit::getGIDs() returns all GIDs handled by the circuit.
    • brain::Circuit::getGIDs( target ) returns GIDS for a specified target.
    • brain::Circuit::getNumNeurons() returns the total number of neurons in the circuit.
  • 43: Add MVD3 support to brain::circuit
  • 39: Add compartment report converter tool
  • 30, 35: Added a new library, Brain, to provide higher level functions and classes. The library provides:
    • A Circuit class to get basic information from cells and targets and load morphologies at circuit level.
    • A brain::cell namspace with a Morphology class plus other related classes to access morphological data about neurons.
  • 38: Fix crash while reading more than ulimit -Sn (1024 default) morphologies
  • 37: Added support for synapse nrn_extra.h5 files.
  • 31: Fix crash while reading more than ulimit -Sn (1024 default) NEST gdf files
  • 29: New member functions in brion::BlueConfig to provide a semantic API.
  • 28: SpikeReport continues parsing files that have broken lines

Release 1.6.0 (9-11-2015)

  • 24: Add getURI() method in SpikeReport.
  • 22: spikeConverter can process spikes to and from stream-type SpikeReport plugins.
  • 22: SpikeReport DSO plugins in the LD_LIBRARY_PATH are loaded automatically.
  • 12: Extended brion::Synapse to also support non-merged synapse files
  • 12: Add brion::Target::parse() to resolve a given target name
  • 9: Extend SWC parser to support fork and end points and undefined section points. The Brion::SectionType enum has not been extended to include end and fork points, these types are translated into the most reasonable one based on the point ancestors.

Release 1.5.0 (7-07-2015)

  • Add RESTING_VOLTAGE constant.
  • Add support for binary spike reports.
  • Add support for spike writer plugin.
  • Add async IO support in brion::CompartmentReportMap.
  • brion::CompartmentReportBinary::updateMapping() more robust.
  • brion::SpikeReport API extended to support stream-based reports. A reference implementation of a stream-based spike report plugin is also provided.
  • Ensure thread-safety with non-threadsafe HDF5 library.
  • Morphology class implementation changed to support the Lunchbox plugin mechanism.
  • New morphology plugin to read SWC morphologies.
  • #3: New morphology converter tool.
  • New null compartment report for benchmarks.
  • New spike converter tool.
  • Optimized function Circuit::get for large circuits (cut down the loading time of the 3.1 M neuron circuit by 6.5 seconds in a Core i7-4930K @ 3.40GHz).
  • Replace use of iostreams by lunchbox::MemoryMap.
  • Several bugfixes and cleanups.

Release 1.4.0 (8-10-2014)

  • Old report readers have been adapted to be static plugins using lunchbox::IOPluginFactory.
  • New spike report reader plugin for NEST spike reports. The plugin can parse multiple report file using shell wildcards (*, ?) in the filepath provided.
  • New constructor for brion::CompartmentReport accepting an URI and the desired access mode deprecates the old read and write constructors.
  • New compartment report based on key-value stores supported by Lunchbox::PersistentMap.
  • Spike loading has been optimized (in a 12-core GPU the expected speed-up is above 4x for large files).