v2.3.1

This minor release adds the recent LHCb measurements of the B_s→φμμ angular observables FL, S3, S4, and S7 from arXiv:2107.13428.

v2.3.0

This release adds many new measurements and fixes several bugs.

New and updated measurements

• The new LHCb measurement of BR(Bs->mumu), LHCb Bs->mumu 2021 and a combination with ATLAS and CMS data, AS combination Bs->mumu 2021
• The Bs->mumu effective lifetime measured by LHCb in 2017 and 2021 and by CMS in 2019
• The new LHCb measurement of RK, LHCb RK 2021
• The 2021 muon g-2 measurement by the "Muon g-2" experiment
• The 2021 BR(Bs->phimumu) measurement by LHCb
• Small typos have been fixed in LHCb B+->Kmumu BR 2014 and CDF B+>Kmumu 2012 (thanks @jafuen for finding them)
• LHCb KS->mumu 2019 has been updated to the published version (thanks to @hoodyn)
• LHCb B->ee 2020 has been added (thanks to @hoodyn)

New observables

• The B0->K*mumu - B0->K*ee LFU difference D_AFB

Bug fixes

• A wrong sign in the B->D(*) form factor correlation matrix that led to too large uncertainties has been fixed (thanks to @Jorge-Alda)
• The subleading contributions to B->Vll at high q^2 had not been added correctly (thanks to @langenbruch or fixing this)
• A wrong factor of 1/2 has been removed from the scalar Wilson coefficients used in semi-leptonic tree-level decays (thanks to @whereforebound)

v2.2.0

New features

• The new flavio.citations module provides a bibliography of theory papers (thanks to @MJKirk)
  • At the end of a flavio computation, the theory papers on which the computation is based can be obtained from flavio.citations.set in terms of a set of INSPIRE citation keys. The bibliography can be reset by calling flavio.citations.reset().
  • The new theory_citations() method of the Observable class returns the theory papers on which the prediction of a given observable is based.

New observable

• The ratio of the branching ratio of B⁰→K*γ and the time-integrated branching ratio of B_s →φγ, BR(B0->K*gamma)/BR(Bs->phigamma) (thanks to @MartinoBorsato for the suggestion)

New measurements

• The B⁺→K*μμ angular observables by LHCb
• The combination of B_s→μμ by ATLAS, CMS, and LHCb
• Some LEP measurement (thanks to @MJKirk)
• Some LHCb and ATLAS measurements (thanks to @olcyr)

Bug fixes

• Some bugs in the implementation of mu-e conversion have been fixed (thanks to @xmarcano for reporting the issue)
• A problem with the PyYAML loader SafeIncludeLoader on some Windows systems has been fixed

Other improvements

• flavio now supports Python 3.9
• The PyYAML loader SafeIncludeLoader now supports absolute file paths in the !include and !include_merge_list constructors
• The config files have been updated to avoid matplotlib deprecation warnings (thanks to @MJKirk)
• The interface to iminuit has been upgraded to use iminuit v2.x

End of support for Python 3.5

This release requires at least Python 3.6. Python 3.5 has reached end-of-life on 30 September 2020. If you are on an outdated system and cannot update Python, have a look at conda. It allows you to install the newest version of Python in a virtual environment without root privileges. Conda is also the best choice if you are on Windows.

v2.1.0

New features

• The new class FlavioParticle based on the particle package (https://github.com/scikit-hep/particle) replaces several functions that were previously used for reading PDG data.
• New versions of flavio.plots.contour and flavio.plots.likelihood_contour_data allow specifying the exact minimum (e.g. obtained by numerical minimization) to be used in contour plots.
• The ParameterConstraints class has the new method read_default, which allows resetting parameters to their default values.

Measurements

• Updated LHCb B->K*mumu angular observables
• LEP-1 measurements have been updated due to the luminosity update (thanks to Adam Falkowski)

Bug fixes

• an import error present for scipy verision >= 1.5.0 has been fixed (thanks to Matthew Kirk)
• an erroneous comment has been fixed (thanks to Alexander Bednyakov)
• ee->WW observables are now strictly real numbers without zero imaginary part

v2.0: Higgs, beta decays, EDMs, charm, and a new maintainer

This release is a significant milestone for the flavio project. First, it contains a huge number of new observables, justifying the major version bump. Second, the baton of maintaining the project is passed from @DavidMStraub to @peterstangl from this release on.

After extending from flavour physics to electroweak precision tests in v1.0 and subsequently adding EDMs and many other "unflavoured" precision observables, this release brings flavio several huge steps closer to becoming a universal observable calculator for all observables where new physics contributions can be described in the framework of the SMEFT. The list of observables has been updated in the documentation.

Concentrating on flavio's role as an observable calculator and likelihood generator, we also removed some old and poorly maintained submodules, notably statistics.fits.

Here's everything that's new:

Higgs physics

Based on arXiv:1911.07866 by Adam Falkowski and David Straub, this release contains everything that is needed for an up-to-date analysis of Higgs physics at LHC - predictions and measurements (Run 1 & Run 2) of the signal strengths in all channels!

Nuclear and neutron beta decays

Based on a collaboration of Martín Gonzalez-Alonso, Adam Falkowski, and David Straub, this release contains Ft values of doubly allowed nuclear beta decays necessary for EFT analyses of new physics in d→u transitions. In addition, also the neutron lifetime and all relevant correlation coefficients in neutron beta decay are included!

Atomic and molecular EDMs

Electric dipole moments of paramagnetic atoms and molecules have been added, based on arXiv:1809.09114.

Semi-leptonic D decays

The decays D→(π, K)lν including all relevant measurements (even binned differential distributions).

New B→D and B→D* form factors

The form factors have been updated to use the results from arXiv:1908.09398. Unit tests against EOS added with the help of @dvandyk.

Utilities for units and elements

• The new flavio.physics.units module can come in handy for converting between units. Look at the source to see what it provides.
• The new flavio.physics.elements module provides two functions to convert between element symbols (e.g. Ag) and the atomic number Z (e.g. 47): symbol(Z) and Z(symbol).

Other improvements

• flavio now supports Python 3.8
• Some parameters and measurements updated

Breaking changes

The flavio.statistics.fits and flavio.statistics.fitters modules have been removed as we are concentrating on the main features of flavio as observable calculator and likelihood generator. These modules had been poorly maintained in the recent past. Code using the fits module can be ported fairly simply to the more recent (and more powerful) flavio.statistics.likelihood module.

v1.6.0

New observables

• B_s→llγ implemented by Méril Reboud (@mreboud) (#91)

Parameters changed

• Masses and widths updated to PDG 2019 values

New measurements

• New CMS measurement of B_q→μμ and new global combination added by Peter Stangl (@peterstangl) (#93)
• New LHCb measurement of K_S→μμ
• New Belle measurement of R_K

Change to plot functions (upcoming breaking change)

The default colour scheme of flavio.plots predated the release of matplotlib 2.0, which introduced a much nicer default colour scheme than the ugly one of matplotlib 1.x. This flavio release changes the colour scheme to use the matplotlib one by default.

Besides, the old keyword argument col= of many plotfunctions, that is supposed to be an integer specifying the color from the scheme to be used, is now deprecated in favour of the normal matplotlib 2.0+ way of using color=, with either 'C0', 'C1', ... for the default colors or any other color supported by matplotlib. The col= argument might be removed in the future.

Note that this change makes matplotlib 2.0 the minimum required version (the current version is 3.1.1).

Deprecations (upcoming breaking change)

The flavio.statistics.fits module has been marked as deprecated and will be removed in one of the next versions. Most of the functionality of that module has been reimplemented in flavio.statistics.likelihood and the old module has no longer been maintained. Please migrate your code if you still use any of the classes from the fits module. If problems arise, please file an issue.

As a consequence, also the flavio.statistics.fitters module is now deprecated.

These changes allow to focus the development efforts on flavio as an observable calculator (e.g. in the context of the global SMEFT likelihood).

v1.5.1

This minor release fixes a few issues and updates parameters, but also brings two minor new features.

New features

• All classes wiith named instances, i.e. Measurement, Parameter, Observable, Implementation, AuxiliaryQuantity and a few others have a new find method which takes a regular expression as input and returns all matching instance names. This has powerful applications, e.g. finding all observables in tau decays:

flavio.Observable.find('tau->')

or all observables that are total branching ratios in decays with a tau in the final state:

flavio.Observable.find('^BR.*->.*tau')

Combining this with get_measurements introduced in 1.5.0, this allows finding observables and their measurements without browsing the source files.

• The Parameter class now pretty prints in a Jupyter notebook. Try flavio.Parameter['m_b'].

Bugs fixed

• The shift due to dimension-6 contributions to the Fermi constant in electroweak precision observables contained ann erroneous factor of 2 for one Wilson coefficient (68ec2c4)
• The NP contributions to the W mass were off by a factor of 2

Measurements

• D0 measurement of W->taunu added
• Updated B->D*lnu branching ratios to HFLAV 2019

Parameters

• Used more precise value for f_K/f_π (fed6180)

v1.5.0

New measurements

• New LHCb bound on B→Kμe
• New LHCb measurement of time-dependent CP violation and rate asymmetries in B_s→φγ
• Updated HFLAV constraint on CP violation in D mixing

New observables

• The LFU-violating angular observable differences Dmue_P4p and Dmue_P5p in B→K*ll (and their measurement by Belle)

New features

• New function flavio.functions.get_dependent_wcs returning the EFT, basis, scale, and WCxf sector of the Wilson coefficients a given observable depends on
• New method flavio.Observable.get_measurements to list existing measurements of a given observable
• New method flavio.statistics.likelihood.MeasurementLikelihood.get_number_observations to return the number of observations (individual measurements of all observables)
• Convenience function flavio.statistics.functions.pvalue(chi2, dof) to compute p-values

Fixes

• The NP Wilson coefficients (not the SM contribution) for K and D mixing were erroneously evolved with 5 active flavours also below the b mass

v1.4.0

New observables

• Muon-to-electron conversion implemented by Jonathan Kriewald (thanks!)
• The B_c lifetime tau_Bc constraining the decay rates B_c→lν

New measurements

• Muon-to-electron conversion and B_c lifetime (see above)
• Updated Belle measurement of R_D^(*) with semi-leptonic tagging
• Updated Belle measurement of R_K^*
• Updated LHCb measurement of R_K
• New Belle limits on BR(B0->K*emu) and BR(B0->K*mue)
• New LHCb limits on BR(Bs->mutau,taumu) and BR(B0->mutau,taumu)

Updated parameters

• Λ_b polarization parameter from BESIII (thanks Danny van Dyk)
• Nucleon matrix elements from FLAG 2019

New features

• New plot function flavio.plots.diff_plot_exp for plots of measurements of observables depending on a single parameter (like q²) but where the parameter is not binned but constant
• All the diff_ and bin_ plot functions gained a new argument scale_factor to rescale the observables by a constant factor

Fixes

• Warning caused by the new pyyaml release suppressed (thanks Markus Prim & Kilian Lieret)

v1.3

This release does not bring new observables or major new features, but instead a couple of small improvements and fixes, the most important ones being:

• Several hadronic parameters from lattice QCD have been updated to the new FLAG 2019 average
• flavio.statistics.probability.combine_distributions now supports multivariate PDFs
• The CKM elements used by the wilson package for Wilson coefficient basis translation are now set to the values set in flavio (thanks to @jackypheno for reporting a related bug, #78)
• Likelihoods in the flavio.statistics.likelihood module now cache their observable predictions for repeated calls (thanks @peterstangl!)