Repository navigation
Conversation
…3 layers (~40 cm) LArStitchingCosmicRayMerging builds its pointing clusters from a sliding fit with HalfWindowLayers = 30 (~9 cm). For cathode-crossing cosmics nearly parallel to the cathode, the ends of the two halves are fitted over too short a stretch, so their directions are poor and the transverse-displacement cut rejects the pair. Use 133 layers (~40 cm) in both master stitching passes. Measured together with the DENT tune (#933) and the larpandoracontent GetClosestVertices y-z fallback, on the DENT-fixed reco of BNB-light runs 18255/18259: crossings at 1-2 deg to the cathode break 32% of the time instead of 78%, all tracks 3.61% instead of 4.03%, with no good stitches lost. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Member
|
This xml update for gen2 production is included in PR #972 , so this PR can be closed |
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
What
Sets
HalfWindowLayersto 133 (~40 cm at the 0.3 cm wire pitch) in bothLArStitchingCosmicRayMergingtools ofPandoraSettings_Master_SBND.xml. The default is 30 layers (~9 cm).The stitching tool takes each candidate's end directions from a sliding fit with this window. For cathode-crossing cosmics nearly parallel to the cathode, a 9 cm fit gives poor directions, so the transverse-displacement (rT) cut rejects pairs that belong together. Growing the rT limit with the gap between the ends doesn't help; a longer end fit does.
Validation
All numbers are on BNB-light data runs 18255/18259.
Offline replay of both master stitching passes on the cosmic-pass halves of 95 broken near-parallel crossings, 60 broken steep ones and 51 good stitched near-parallel ones. Fake joins come from event mixing (East halves of one event against West halves of others).
Full reco2 on the target events, together with the DENT tune (pandora dent tune xml updates #933) and the GetClosestVertices y-z fallback ([v10_14_02_br] Stitching: y-z fallback in LArStitchingHelper::GetClosestVertices instead of throwing LArSoft/larpandoracontent#106): 36 of 95 parallel and 4 of 60 steep broken crossings are rejoined, and 0 of 60 good stitches are lost.
Full DENT-fixed sample (1,818 reco1 files, 88k events in common), same combination: broken cathode crossings (all tracks) drop from 4.03% to 3.61%. By angle to the cathode: 1–2° from 78% to 32%, 2–3° from 44% to 25%, 3–5° from 19% to 14%. Above 20° nothing changes.
Event displays of 12 random new stitches with a > 10° kink: about 10 are clearly one muon (the kink is a hook or distortion at the cathode), 2 are ambiguous, and none join two unrelated tracks.
Notes
🤖 Generated with Claude Code