Releases: madminer-tool/madminer
Releases · madminer-tool/madminer
v0.9.6
Bug fixes
- Fix typo in
_calculate_partition_bounds
method calls (#524).
Internal changes:
- Drop support for Python 3.7 (#527) (support ends on June 2023).
- Lint all codebase with Black 22.12.0 (#525).
- Lint all codebase with Isort 5.11.3 (#526).
- Migrate default branch from
master
tomain
.
Full Changelog: v0.9.5...v0.9.6
v0.9.5
Bug fixes
- Fix calculation of Fisher information when there are no nuisance parameters (#519).
- Fix missing
sympy
dependency on Condaenvironment.yml
file (#522).
Internal changes:
- Change HDF5 settings load messages from
ERROR
toINFO
level (#518).
Full Changelog: v0.9.4...v0.9.5
v0.9.4
New features:
Internal changes:
- Removed the default values of the hdf5.py
save_events
public function (#504). - Added Python 3.10 to the list of CI/CD tested Python versions (#510).
Full Changelog: v0.9.3...v0.9.4
v0.9.3
Bug fixes
- Fix HDF5 serialization for
Observable
,NuisanceParameter
andSystematic
models. #503
Internal changes:
- Make save_events
sampling_benchmarks
argument mandatory. #503
Full Changelog: v0.9.2...v0.9.3
v0.9.2
v0.9.1
v0.9.0
New features:
- New CI workflow step,
lint
, to ensure notebooks style consistency 2047dc7.
Internal changes:
- Drop support for Python 3.6 #484 (support ends on December 2021).
- Added support for Python 3.9 #484.
- Upgrade notebook examples to Python3 syntax #480.
- Upgrade
numpy
minimum version to version1.20.0
#486. - Complete rewrite of the HDF5 interface functions #483.
- Definition of dataclass models to ease attribute access #485:
Benchmark
.FiniteDiffBenchmark
.AnalysisParameter
.NuisanceParameter
.Cut
,Efficiency
andObservable
.Systematic
(and theSystematicType
enum 6183d55).
Breaking changes
There are multiple backward compatibility breaking changes to consider
- The HDF5 file keys to store madminer settings and / or serialized data have changed. We recommend re-creating them.
- The methods LHEReader.add_cut() and DelphesReader.add_cut() got their
pass_if_not_parsed
arg. renamed torequired
. - The method LHEReader.add_efficiency() got its
value_if_not_parsed
arg. renamed todefault
. - The function combine_and_shuffle got its
overwrite_existing_file
unused argument removed.
v0.8.3
Bug fixes:
- Fixed particles
phi
property access after scikit-hep/vector migration (not callable anymore). - Fixed missing "modified_score" mode within
Ensemble.calculate_fisher_information()
accepted ones.
Internal changes:
- Improved
math
functions mappings within the utilities (6a7ead3). - Replaced most of the hardcoded particle IDs by the scikit-hep/particle package.
- Removed legacy
check_required_data
utilities function (13bf559). - Removed legacy
UploadCommand
setup.py class (now using GHA workflows) (243e1e8).
Documentation:
- Defined new
CITATION.cff
file to ease package citation.
v0.8.2
Bug fixes:
- Fixed MadminerParticle initialization by using @ classmethod helpers. Thanks @lyazj.
- Replaced
MadminerParticle
inherited class (VectorObject4D
->MomentumObject4D
). Thanks @lyazj. - Bumped
vector
package to version0.8.4
, first version to properly allow subclassing.
Internal changes:
- Replaced Python2 compatible super-class calls by Python3 style.
Documentation:
- Fixed README relative links (unavailable on Madminer's PyPi information page).
v0.8.1
Bug fixes:
- Fixed a Python2 -> Python3 migration issue when dealing with dictionary values (thanks @sbrass).
Internal changes:
- Replaced legacy
scikit-hep>=0.5.0, <0.6.0
requirement (underused), by thevector
package. - Replaced legacy rootproject/root-ubuntu Docker base image by rootproject/root equivalent.
Documentation:
- Fixed readthedocs automatic documentation building (broken since July 1st 2020).
- Define explicit
.readthedocs.yml
file with the desired configuration. - Define GitHub
docs.yml
workflow, to validate documentation building uponmaster
branch pushes / PRs.