Skip to content

Tags

Tags give the ability to mark specific points in history as being important
  • v0.7.0

    eccac433 · hotfix: update version ·
    v0.7.0
    
    - feature: add `mc_parameters` (v2) field to determine wich normalized parameters will be used to produce Monte-Carlo samples. It works with names of parameters from the storage, not their covariance uncertainties aliases.
  • v0.6.0

    v0.6.0
    
    - feature: new versions of model where added: `v1a_neutrino_rate` and `v2`. Calculation is based on total neutrino rate of each reactor.
    - fix: `full.covmat.chi2cnp` was removed to avoid confusions. `full.covmat.chi2cnp_alt` was renamed to `full.covmat.chi2cnp`, it is defined as (18)-(19) equations from the [arXiv: 1903.07185](https://arxiv.org/pdf/1903.07185). Only `v1a*` and `v2` were updated.
  • v0.5.2

    Version v0.5.2
    
    - hotfix: add `absolute_efficiency` gruop for covariance matrix and pull terms. Works with model `v1a`, and `v1a_distorted`.
  • v0.5.1

    Version: 0.5.1
    
    - hotfix: add `absolute_efficiency` gruop for covariance matrix and pull terms. Works with model `v1a`, and `v1a_distorted`.
    - hotfix: update `reactor_refine_data` function for model `v1a_distorted`.
  • v0.5.0

    Version 0.5.0
    
    - feature: add `v1a_distorted` to study extreme spectral distortions.
    - feature: add `covariance_groups` parameter to control passed nuisance parameters to covarince matrix, works only with `strict=False`. Supported for `v1a` and `v1a_distorted`.
    - feature: add `pull_groups` parameter to control passed nuisance parameters to `nuisance.pull_extra`. Supported for `v1a` and `v1a_distorted`.
    - feature: add `is_absolute_efficiency_fixed` parameter to switch between fixed/variable absolute correlated detector efficiency. Supported for `v1a` and `v1a_distorted`.
    - feature: prepare code for reading variable periods of reactor data.
    - tests: optimize creation of models, ordering of tests.
  • v0.4.2

    - chore: disable `numba` caching as it may cause problems for parallel
    execution. Configurable.
    - chore: step `dgm-reactor-neutrino` dependence version to 0.2.2.
  • v0.4.1

    97d30207 · Merge branch 'release' ·
    - feature: validate the version from `data_information.yaml` and use it
    to determine the
      `source_type` (data format).
    - feature: check overridden indices exist (model v1a).
    - fix: apply `Abs` transformation to scaled fission fractions. Fit
    becomes much stable.
    - fix: update keys related to graph visualization for
    [scripts/scripts/run_dayabay.py](extras/scripts/run_dayabay.py).
  • v0.4.0

    - feature: configure the final binning (v1a) via configuration file of
    an argument.
  • v0.3.0

    fc8fa066 · CI: add python build ·
    v0.3.0
    
    First delivering via CI
  • v0.2.3

    - feature: Add `switch_data` metoh from version v1a. It can switch model
    output between real data and Asimov.
    - feature: Support all the inputs via `uproot` (`ROOT` is supported just
    as before).
    - feature: v1a, read antineutrino spectra from tsv file (not python).
  • v0.2.2

    - [feature] Add key `leading_mass_splitting_3l_name` to switch between
    |Δm²₃₂| and |Δm²₃₁| to v1a version. In Day Bay official analysis |Δm²₃₂|
    as leading mass splitting.
    - [feature] Add key `override_cfg_files` that allows to override paths
    to configuration files. Key is supported for models v1+.
    - [feature] Remove `source_type` parameter from version v1+. Now
    `source_type` is determined automatically from `path_data`. To override
    default value, just path `path_data` to your model.
  • v0.2

    first PYPI version
  • v0.1

    first public (pre)release
  • archive/fork-feature-fits

    archive stale branch
  • backup/covmatrix0

    Contains some code to compute covariance matrices