Skip to content

Latest commit

 

History

History
207 lines (158 loc) · 6.58 KB

CHANGELOG.md

File metadata and controls

207 lines (158 loc) · 6.58 KB

Changelog

v0.1.26:

Added:

  • Simulation config: synapse_replay input files must be .h5 (#351)
  • SimulationConfig: Adds "represents_physical_electrode" field in "inputs" section (#354)

Fixed:

  • Update HighFive SHA to 2.9.0. (#343)
  • Make documentation available for Python API (#345)
  • Build static macos and linux wheels (#355)

v0.1.25:

Added:

  • Implement fall-back for make_collective_reader. (#326)
  • Inject dataset reading via Hdf5Reader. (#307)
  • Use using Range = std::array<Value, 2>. (#319)

Fixed:

  • SimulationConfig: spike_threshold is float (#340)
  • Check more variants of CMake invocations. (#333)
  • Remove CMAKE_SKIP_INSTALL_ALL_DEPENDENCY. (#336)
  • add SimulationConfig to __all__ (#339)
  • Make sure defaults for condition blocks and others are applied (#338)
  • Allow sonata::sonata_{shared,static}. (#335)
  • Update pybind11==2.11.1. (#332)
  • Refactor edge_index::resolve. (#314)
  • Refactor _readSelection. (#315)
  • Prevent pytest from picking up pybind11 tests. (#320)
  • update to require using cmake 3.16 (#321)
  • test for completely empty datasets (#281)
  • Update HighFive to v2.8.0. (#318)
  • be neighborly, do not use using in headers (#316)
  • work around brew erroring in github actions (#317)
  • Move Selection to its own header. (#306)
  • Simplify HighFive usage. (#304)
  • Simplify single group check. (#303)
  • Optimize loading ?fferent_edges. (#298)
  • Allow disabling coverage an 'Debug'. (#299)
  • Simplify Selection::Selection. (#300)
  • Use RawIndex = std::vector<std::array<.., 2>>. (#294)

v0.1.24:

  • small cosmetic fixes/py 3.12 wheels (#290)
  • Fix toJSON incorrect empty, node_id and compound problems (#289)
  • NodeSet update: merge NodeSet objects together (#283)

v0.1.23:

  • Create non-tuple Spike Report access (#285)
  • Add 'lfp' report type and 'electrodes_file' field (#271)
  • Add missing reversal property of shotnoise family (#284)
  • Move std::vector methods above scalar methods in bindings (#280)

v0.1.22:

  • bump HighFive to v2.7.1 (#278)
  • Extend metadata type to be int, float, bool and string (#277)
  • Prevent node_set materialization for @libary with integers (#276)

v0.1.21:

  • SimulationConfig: set the default value of random_seed in inputs to None (#270)
  • Allow empty arrays in nodesets (#272)
  • implement spec for population properties for vasculature, astrocytes and endfeet (#275)

v0.1.19:

  • switch to having connection_overrides being a list (#263)
  • expose circuit config status (#264)
  • Don't read the full node sets file before parsing (#261)

v0.1.18:

  • speed up nested compound nodesets

v0.1.17:

Added:

  • Allow for partial circuit configs (#241)
  • Added spatial index accessors for segment and synapse indices (#250)

Fixed:

  • Add include(Catch) to CMakeLists.txt. (#248)
  • Enable setting the C++ standard. (#247)
  • SONATA sim config updates (#245)
  • make chemical the default connection type (#242)
  • add check for duplicate populations (#239)
  • refactor python tests so they are not in a monolithic file (#238)
  • Fix get_population_names in README.rst (#235)
  • update github actions

v0.1.16:

Fixed:

  • Fix #226 for empty selections (#232)

v0.1.15:

Added:

  • Parse metadata and beta_features sections in SimulationConfig (#224)
  • SimulationConfig parser: add modifications properties in conditions section (#228)
  • make nodes/edges paths files available through PopulationProperties (#230)

Fixed:

  • Updates on the SimulationConfig parser (#227)
  • don't make many small reads via HighFive/hdf5 (#226)
  • Bump pybind11 from v2.5.0 to v2.10.0; supports python 3.11 (#223)
  • fix report files not respecting the output path (#222)

v0.1.14:

Fixed:

  • Parse synapse mod variables in conditions/mechanisms section (#217)

v0.1.13:

Added:

  • Parse SimulationConfig #197, #203, #207, #208, #210, others

Fixed:

  • throw if biophysical_neuron_models_dir is missing for biophysical population (#206)
  • for biophysical node populations, don't raise error if alternate_morphologies exists (#209)

v0.1.12:

Added:

  • Add a check when feeding a directory path to readFile (#188)
  • Add support for IO-blocks (#183)

Fixed:

  • Bump HighFive to v2.4.0 (#190); fixes errors on archs: aarch64, s390x and ppc64le
  • Update hdf5 to 1.12.1 in linux wheel building (#191)
  • Add -Wp,-D_GLIBCXX_ASSERTIONS to more likely find out of bounds memory access (#182)
  • Improve memory operations + NodeID-ElementID mapping bindings (#179)

v0.1.11:

Fixed:

  • Prefer building Release mode for python wheels, can control with SONATA_BUILD_TYPE (#176)
  • Only do a single read when getting a report frame (#174)

v0.1.10:

Added:

  • Have the shim module document functionality (#162)
  • added simulation config implementation (#158)
  • reporting getNodeIdElementIdMapping function(#168)
  • getTimes inside SpikeReader::Population (#169)

Fixed:

  • move to C++14 (#159)
  • Exclude musllinux wheels building
  • Update HighFive submodule to latest 2.3.1 (#156)
  • update to new pybind11::mkdocs method (#173)
  • switch away from nose for test discovery (#172)

v0.1.9: Thu 01 Jul 2021

Added:

  • circuit_config.json implementation (#142)
  • allow for simple expressions (#147) in node_sets as proposed to the offical SONATA spec in AllenInstitute/sonata#129

Fixed:

  • upgrade Catch2 to 2.13.6; autodetect test names (#148)
  • move to fmt 7.1.2 (#145)

v0.1.8: Wed 10 Feb 2021

Fixed:

  • moved extlib/nlohmann/json.hpp one level deaper so that building with non-versioned nlohmann json is easier

v0.1.7: Thu Jan 21 2021

Added:

  • Implemented node_sets functionality (#107)
  • Handle pathlib.Path for paths (#113)
  • Add repr to python Selection binding, aswell as the (Node|Edge)Population (#114)

Fixed:

  • Fix python bindings for Selection with negative values (#112)
  • Fix regression from move to GitHub actions

v0.1.4 Wed Jul 22 2020

  • Use common docstrings between C++ and Python
  • Better documentation
  • Report readers documentation
  • Bump HighFive and pybind11
  • Numpy arrays for the report outputs

Fixed:

  • Fixed python sdist distribution by disabling test requirements
  • Fix report_reader to have several time the same index

v0.1.3: Fri Mar 13 2020

Added:

  • convenience method to create a Selection of all members of a Population (#60)
  • make bbp::sonata::Selection comparable (#62)
  • comparison for node attributes to values (#64)
  • intersection and union for selections (#67)
  • output reader for spikes, somas and elements (#71)

Fixed:

  • Mac python wheels (#77)