v1.4.0

This release changes the pseudo-data definition slightly, now including t-channel single top. This was necessitated by a previous bug, where this sample was not getting picked up correctly in the statistical model. Reference histograms have been updated accordingly.

In addition to that, the update includes a fix for the custom ServiceX dataset grouping when names get so long that they become mangled. It also adds support for specifying URI prefixes for ATLAS xcache use cases.

Full list of changes:

• fix: handle long file names in ServiceX dataset grouping (AGC v1) (#196)
• fix: pick up correct single top sample in statistical model (AGC v1) (#195)
• feat: backport client update for EAF to v1 branch (#193)
• feat: support URI prefixes for xcache in AGC v1 (#190)

Full Changelog: v1.3.0...v1.4.0

v1.3.0

This release adds support for reading input data from public EOS via xrootd via a new INPUT_FROM_EOS setting in the config.yaml. It also improves the value used for a bin edge cut (no impact on behavior, but now the cut is put at the actual bin edge).

Full list of changes:

• fix: sync changes from #185 to script (#186)
• feat: support input files on public EOS (#185)
• fix: set bin edge cuts to 110 GeV instead of 120 GeV (#181)

Full Changelog: v1.2.0...v1.3.0

v1.2.0

This updates the reference implementation of the AGC analysis task version 1. The b-tagging cut is fixed to use > instead of >= to follow the task specification. Bin yields in histograms no longer contain an offset of 1e-6 that is no longer relevant and was used as a workaround in the past. This update also includes an improved histogram validation utility.

Full list of changes:

• feat: fixes towards v1.2 (#179)
• feat: do not require pseudodata hists in validate_histograms.py (#178)
• feat: updated validation script to check partial bin sums (#175)

Full Changelog: v1.1.0...v1.2.0

v1.1.0

This updates the reference implementation of the AGC analysis task version 1. The propagation of pT systematic variations is fixed to correctly affect observables in the 4j2b region. Histograms are now saved in their original binning they were produced in, the histograms are subsequently re-binned only for statistical inference. A validation utility is added to compare histograms produced with any AGC implementation to references provided in this repository.

Full list of changes:

• feat: add validation script and histogram references (#170)
• feat: save histograms without re-binning (#165)
• fix: pt variation change carried through signal region calculations (#163)

Full Changelog: v1.0.0...v1.1.0

v1.0.0

The ttbar CMS Open Data analysis now uses an input dataset in NanoAOD format. These new files were created from the same MiniAOD files that were previously used to create the custom ntuple format used in earlier versions. Functionality is otherwise equivalent to v0.2.0, the only aspects that changed were those relating to the NanoAOD files (such as list of columns for I/O testing).

The AGC now also has a new documentation webpage on readthedocs: https://agc.readthedocs.io/.

Full list of changes:

• fix: update I/O branches (#126)
• fix: GiB vs GB for file sizes (#121)
• fix: minor changes to docs (#125)
• changed jupytext synced file for plotEvents to py (#124)
• fix: added pandoc for building docs in CI (#123)
• docs: Read the docs website for AGC (#114)
• feat: switch to NanoAOD input files for AGC CMS ttbar notebook (#102)
• fix: add axis labels and binning to cabinetry config file (#117)

Full Changelog: v0.2.0...v1.0.0

v0.2.0

Improved ServiceX pipeline, otherwise corresponding to the ACAT 2022 setup from v0.1.0. The ServiceX executor in coffea is no longer used. Transforms are instead all launched at once and subsequently processed in coffea once they are all done. Downloads of the transform are skipped, the results are directly streamed into coffea.

Full list of changes:

• build: update versions of dependencies from CMS ttbar setup (#112)
• feat: combine ServiceX transforms for CMS ttbar notebook (#107)
• fix: update func-adl query for H>ZZ notebook (#106)
• feat: ATLAS H>ZZ ServiceX transform names (#104)
• feat: serve ROOT file URIs with ServiceX for CMS ttbar notebook (#103)
• feat: serving ROOT file URIs with ServiceX in HZZ notebook (AGC demo day) (#96)
• feat: use jupytext for demonstrator notebooks (#92)
• [pip] Update AGC related packages to latest versions (#89)
• docs: add Zenodo DOI badge (#88)

v0.1.0

Setup used for benchmarking with the CMS ttbar analysis implementation, as shown at ACAT 2022: https://indico.cern.ch/event/1106990/contributions/4998188/.