XENONnT / fuse / 28584799675 / 2
74%
main: 74%

DEFAULT BRANCH: main
Ran
Files 52
Run time 1s
Badge
Embed ▾
README BADGES
x

If you need to use a raster PNG badge, change the '.svg' to '.png' in the link

Markdown

Textile

RDoc

HTML

Rst
coverage: 73.569% (+1.3%) from 72.272%
28584799675.2

push

github

web-flow 
Fix get_parent same-time-tag bug in LineageClustering (#383)

* Fix get_parent same-time-tag bug in LineageClustering

Tie-break by spatial proximity to the daughter's creation position when
multiple parent interactions share the same time tag.

For fast gammas that Compton-scatter at one site and photoabsorb at
another within G4's sub-ns time-tag precision (~0.1 ns / 3 cm), the
original time-cut + array-last logic returned whichever same-time
parent step happened to be last in the array. This mis-attributed the
Compton e- daughter at site A to the gamma's photoabsorbtion vertex at
site B, collapsing the two scatter sites into one cluster.

Net effect on single-gamma decays (e.g. K40, Co60):
~65 % of physically multi-scatter events get reclassified as single-
scatter at the recon stage, biasing the MS templates by ~50 %.
Chain isotopes (Th232, Ra226, etc.) and intrinsic-β isotopes (Pb212
with nucleusLimits excluding daughter chain) are unaffected because
their secondaries do not produce same-time-tag γ Compton+phot pairs.

Validated across 5 isotopes (20 jobs × 100k primaries each):

isotope | vanilla MS-ratio | patched MS-ratio | vs WFsim 2026
K40     | 0.385            | 1.006            | reference 1.000
Co60    | 0.351            | 0.998            | reference 1.000
Th232   | 0.997            | 0.996            | reference 1.000
Ra226   | 1.002            | 1.000            | reference 1.000
Pb212   | 1.000            | 1.000            | bit-identical h5

The fix preserves the existing nearest-in-time fallback when no parent
step satisfies t <= particle.t.

Full validation report and forensic walkthrough:
  /project/lgrandi/mpierre/high-ER-analysis/Inference_highER/
  template_builder/notebooks/WFsim_fuse_comparison_checks/
    Report_get_parent_patch_validation.md
    Investigate_one_K40_event_stage_by_stage.ipynb

* Pin docformatter hook to python3.12 to unblock CI

The docformatter v1.7.7 hook fails on the default Python pre-commit picks
in CI... (continued)

2725 of 3704 relevant lines covered (73.57%)

0.74 hits per line

Source Files on job 28584799675.2