Combine v10.0.2

What's Changed

• Replace old iterators in python script (ROOT 6.32 compatibility), fixing python scripts by @anigamova in #977
• fix seg fault with FitDiagnostics plots by @kcormi in #991
• Fix #987, incorrect dependency in conda_env.yml by @pitkajuh in #988
• Fix bugs in shapeN + CMSHistFunc/Sum by @ajgilbert in #995
• Update cvmfs-ci workflow by @anigamova in #993
• Update Impacts.py (making --setParameters option in Impacts.py compatible with the one in src/Combine.cc ) by @anigamova in #996
• Adding getProcessNorms() to CMSHistErrorPropagator by @nucleosynthesis in #994

Tutorials and documentation updates

• Add tutorial material by @kcormi in #979
• Add material for tutorial on model building by @anigamova in #981
• Updating CH instructions for checkout since not needed as much now (… by @nucleosynthesis in #976
• Add approx. impacts to docs by @nucleosynthesis in #985

New Contributors

• @pitkajuh made their first contribution in #988

Full Changelog: v10.0.1...v10.0.2

Combine patch v10.0.1

This is a patch fixing a bug reported here.

What's Changed

• Fix a bug check_inf_parameters in #974

Full Changelog: v10.0.0...v10.0.1

Combine v10.0.0

This version is targeting CMSSW_14_1_0_pre4 and is compatible with ROOT >= 6.30/07 and el9 architecture.
Installation instructions are available at http://cms-analysis.github.io/HiggsAnalysis-CombinedLimit/v10.0.X/#installation-instructions.

Major updates

• ROOT v6.30/07 support: #969 #971 #972 #947 #951 #955 #950

• Move combineTool.py scripts to Combine in #902 #970

Code and documentation improvements

• Follow up on ordered collections for nuisance parameters in #940
• Force updateGrid in expectedFromGrid HybridNew.cc in #941
• Updates in DataFrameWrapper #963
• Update logic for Feldman Cousins intervals and add example in #966
• Updated CI workflows in #965 #946 #956
• Use override keyword where necessary in #948
• Support compilation with StatAnalysis in #967

Full Changelog: v9.2.1...v10.0.0

Combine v9.2.1

Main update

• Moved to ordered collections to fix the order of nuisance parameters for the workspace build with text2workspace.py in #936

Documentation and tutorial

• Changes to mkdocs and CI for using mike to version the docs in #918 #926
• Installation instructions updates in #924 #928
• Tutorial updates in #930 #931 #932
• Physics models and general documentation updates #921 #914 #922

Full Changelog: v9.2.0...v9.2.1

Combine v9.2.0

Main updates

• Add flags for SinglePdfGenInfo to avoid using gen spec when changing nuisances per toy. in #836
• Add regex for --floatNuisances in #851
• Updates to ChannelCompatibilityCheck in #760 #805
• New logger in #868
• Updates for ROOT versions > 6.22 in #873 #874 #915 #890 #905
• Spellcheck #894 #907
• Random start points option for profiled parameters of interest in #713

Physics models

• Adding physics model for (STXSto)SMEFT fits in #855
• Accelerated interference models in #842

Documentation/tutorial updates

• New Documentation Section focused on underlying stats without code in #839
• Parametric exercise updates in #913
• Unfolding tutorial in #862
• Add Contributing Document in #838
• Update impact docs to explain multiple POI usage in #852
• Documentation language edit /spell check in #871
• Include docker installation in #864
• Adding folder CAT23001 with datacards used in the paper in #865
• Improve structure of tutorial sections in #909

Full Changelog: v9.1.0...v9.2.0

Combine v9.1.0

What's Changed

• Documentation and tutorial updates across several PRs: #812 #813 #814 #815 #816 #824
  #826
• Replace CLs and CLsplusb with Pb and Pmu in the documentation and code: #832
• Separated --dataset option in combine from text2workspace to avoid overwriting workspace location used in datacard parser when an alternative dataset is provided to combine in #832
• Added an option to assign RooUniform pdf for flatParam nuisances and restructured the loop over the systematics in #834

Full Changelog: v9.0.0...v9.0.1

Combine v9

This version is targeting CMSSW_11_3_4 but is broadly compatible with ROOT 6.22 through ROOT 6.26, and hence also CMSSW 12_X release series. Installation instructions are summarized in https://cms-analysis.github.io/HiggsAnalysis-CombinedLimit/#installation-instructions

New features include:

• Full python3 compatibility (#759)
• Improved ROOT version support (#776)
• 2D RooParametricHist class (#545)
• Shapes from pandas dataframe (#548)
• Added instructions and tools for producing inputs for simplified likelihoods (#800)
• Many bug fixes (see e.g. list of merged PRs to 112x)

From this point on, new developments should always target the main branch, with backports to release series as necessary.

combine v8.1.0 for CMSSW_10_2_X on CC7

This release includes the latest developments made for the Higgs preliminary combination HIG-19-005.

Improvements to the code include:

• Options to speed up discrete minimisation (a6950ba). Enabled by including the option --X-rtd MINIMIZER_freezeDisassociatedParams, can now also include the following

  • --X-rtd MINIMIZER_multiMin_hideConstants: hide the constant terms in the likelihood when recreating the minimizer
  • --X-rtd MINIMIZER_multiMin_maskConstraints: hide the constraint terms during the discrete minimization process
  • --X-rtd MINIMIZER_multiMin_maskChannels=<choice> mask in the NLL the channels that are not needed:
    • <choice> 1: keeps unmasked all channels that are participating in the discrete minimization
    • <choice> 2: keeps unmasked only the channel whose index is being scanned at the moment (works only in multipleMinimize mode 0, makes discrete minimization faster)

• Add option to completely skip discrete minimization (for DEBUG purposes only) --X-rtd MINIMIZER_skipDiscreteIterations (b3802ae)

• Added option in text2workspace.py: --just-check-physics-model to test a physics model quickly without building the full workspace (f162ceb)

• Improved tool test/scanScalefactorsFromWS.py to allow producing scans of scale-factors from PhysicsModels using workspace directly.

• General improvements in memory management in workspace creation.

combine v8.0.1 for CMSSW_10_2_X on CC7

Updates in this tag compared to v8.0.0:

• Fix for TTree branch format in FitDiagnostics
• Documentation updates

For more extensive changes in v8.0.0 with respect to the 8_1_X branch see here.

combine v8.0.0 for CMSSW_10_2_X on CC7

Updates in this tag compared to v7.0.12:

• New documentation pages, using the mkdocs framework. The documentation source is included in the repository as simple markdown files. Users are welcome to make additions and corrections as pull requests to this repo.
• It is now possible to include additional constraint terms for regularisiation when unfolding using combine. Detailed documentation for this is given here.
• The option -S 0 to remove all systematic uncertainties has been removed. Instead, to freeze all constrained nuisance parameters the option --freezeParameters allConstrainedNuisances should be used, which replaces the previous shortcut of --freezeParameters all.
• The possibility to use some old method names has now been fully removed. When setting the -M option, FitDiagnostics, AsymptoticLimits and Significance must be used instead of, respectively, MaxLikelihoodFit, Asymptotic and ProfileLikelihood.
• Nuisance edit selections for bins, processes or systematic names now require a complete string match. For example, nuisance edit add procA binA [...] will no longer match procAB and binAB. Note that regex selections can still be used to match multiple labels, but again are now required to match the full strings.
• Nuisance parameters can now be frozen using attributes that have been assigned to the corresponding RooRealVars. Syntax is --freezeWithAttributes attr1,attr2,...,attrN.
• For Higgs analyses: added YR4 cross sections, branching ratios and partial width uncertainties in data/lhc-hxswg/sm/, as used in HIG-17-031
• [EXPERIMENTAL] For binned analyses using autoMCStats a faster implementation of the vertical template morphing for shape uncertainties can be enabled at runtime with the option --X-rtd FAST_VERTICAL_MORPH. Any results using this flag should be validated carefully against the default.