Skip to content

MuGirl - Add quality cuts on NSW segments

Johannes Junggeburth requested to merge jojungge/athena:ATR-25025 into master

Hi everybody,

I scrutinized few of the spurious events reported in ATR-25025. There is a particular one that is super interesting as this one generates a ton of MuGirl muons that are reconstructed by arbitrary NSW hits.

root [4] BasicTesterTree->Scan("muons_pt:muons_eta:muons_author:muons_truthType:muon_segmentLink:segment_nMM_EtaHits:segment_nMM_StereoHits:segment_nSTGC_EtaHits:segment_nSTGC_PhiHits")
***********************************************************************************************************************************
*    Row   * Instance *  muons_pt * muons_eta * muons_aut * muons_tru * muon_segm * segment_n * segment_n * segment_n * segment_n *
***********************************************************************************************************************************
*        1 *        0 * 8.6900310 * 2.0171515 *         6 *         4 *         0 *         4 *         0 *         7 *         0 *
*        1 *        1 * 12.340089 * -2.508581 *         6 *         0 *         1 *         4 *         4 *         8 *         0 *
*        1 *        2 * 10.537893 * 1.9191445 *         6 *         0 *         2 *         4 *         0 *         7 *         0 *
*        1 *        3 * 7.0811395 * -1.629407 *         6 *         0 *         3 *         4 *         2 *         2 *         0 *
*        1 *        4 * 7.3060441 * -2.133970 *         6 *         0 *         4 *         4 *         2 *         4 *         0 *
*        1 *        5 * 5.3500342 * -2.062107 *         6 *         0 *         5 *         4 *         4 *         4 *         0 *
*        1 *        6 * 5.4994983 * 2.0030026 *         6 *         0 *         6 *         4 *         2 *         7 *         0 *
*        1 *        7 * 4.9618248 * 2.1673636 *         6 *         0 *         7 *         4 *         4 *         8 *         0 *
*        1 *        8 * 5.1116323 * 1.7110877 *         6 *         0 *         8 *         4 *         4 *         8 *         0 *
*        1 *        9 * 4.5482745 * -2.298452 *         6 *         0 *         9 *         4 *         2 *         8 *         0 *
*        1 *       10 * 2.3290331 * -2.433726 *         6 *         0 *        10 *         4 *         2 *         4 *         0 *
*        1 *       11 * 1.5519968 * 2.2765874 *         6 *         0 *        11 *         4 *         0 *         7 *         0 *
*        1 *       12 * 3.8701121 * -2.127248 *         6 *         0 *        12 *         4 *         4 *         7 *         4 *
*        1 *       13 * 1.2498190 * -2.528452 *         6 *         0 *        13 *         4 *         2 *         8 *         0 *
*        1 *       14 * 3.4110505 * -2.585165 *         6 *         0 *        14 *         4 *         2 *         8 *         0 *
*        1 *       15 *  3.605124 * -2.000551 *         6 *         0 *        15 *         4 *         2 *         5 *         0 *
*        1 *       16 * 3.6106908 * -2.065633 *         6 *         0 *        16 *         4 *         0 *         4 *         0 *
*        1 *       17 * 3.5291152 * 2.1401772 *         6 *         0 *        17 *         4 *         4 *         8 *         0 *
*        1 *       18 * 2.5159516 * -1.646621 *         6 *         0 *        18 *         3 *         1 *         4 *         0 *
*        1 *       19 * 3.0232121 * 1.5547832 *         6 *         0 *        19 *         2 *         0 *         7 *         0 *
Type <CR> to continue or q to quit ==>
*        1 *       20 * 3.2614803 * 2.3482048 *         6 *         0 *        20 *         4 *         0 *         8 *         0 *
*        1 *       21 * 2.9061310 * 1.6900132 *         6 *         0 *        21 *         4 *         4 *         8 *         0 *
*        1 *       22 * 1.1399654 * -2.371282 *         6 *         0 *        22 *         4 *         2 *         4 *         0 *
*        1 *       23 * 2.7491135 * 0.8322315 *         6 *         0 *        23 *         0 *         0 *         0 *         0 *
*        1 *       24 * 2.7693615 * -2.073358 *         6 *         0 *        24 *         4 *         2 *         8 *         0 *
*        1 *       25 * 2.8100295 * -2.407633 *         6 *         0 *        25 *         4 *         2 *         8 *         0 *
*        1 *       26 * 2.6617591 * -2.235891 *         6 *         0 *        26 *         0 *         0 *         4 *         4 *
*        1 *       27 * 2.6348168 * 1.6742838 *         6 *         0 *        27 *         4 *         4 *         8 *         0 *
*        1 *       28 * 2.6246531 * 2.1547827 *         6 *         0 *        28 *         4 *         0 *         6 *         0 *
*        1 *       29 * 2.5497019 * 2.2802910 *         6 *         0 *        29 *         4 *         4 *         8 *         0 *
*        1 *       30 * 2.4655008 * 2.3677763 *         6 *         0 *        30 *         3 *         1 *         4 *         0 *
*        1 *       31 * 2.4270520 * 1.3973145 *         5 *         0 *        31 *         0 *         0 *         0 *         0 *
***********************************************************************************************************************************
(long long) 33

as one can see, the segments either lack of sTGC phi hits or of Micromega stereo hits. In order to mitigate the situation the following changes are introduces in this MR:

  • MuonSegmentHitSummary: In the past the hits were categorized into MDT, trigger (i.e., RPC or TGC) and CSCs. At the time when the NSW hits were implemented, the CSChit information was abused to collect all NSW hits. The NSW hits are now split into sTGC and MicroMega hits and these are further subdivided into eta & phi hits in the case of sTGCs and eta and stereco in the other case.

  • Changed the categorization of NSW segments MuonSegmentSelection:

    • NSW segments with less than 8 hits are categorized as 0.
    • NSW segments with no precision hit inone direction are categorized as 1. Dropped the requirement on stationEta
    • NSW segments with less than 12 hits or with less than 4 hits are categorized as 2. Dropper the requirement on stationEta here as well
  • MuonLayerInterSection: Add data member to cache the segment quality avoiding to calculate it again & again during the sorting procedure or the seed segment selection

  • xAOD::Segment conversion: Add detailed information about the number of NSW hits. This will be required for early commissioning.

  • MuTagIMO: Resrict to CSC segment selection, where CSC segments are required. For the moment no selection on NSW hits is applied. This will be implemented in a follow-up MR.

  • Misc: Pipe EventContext through where necessary

After the implementation of these requirements the number of fake MuGirl muons drops to 4 in the same event:

root [1] BasicTesterTree->Scan("muons_pt:muons_eta:muons_author:muons_truthType:muon_segmentLink:muon_segment_nMM_EtaHits:muon_segment_nMM_StereoHits:muon_segment_nSTGC_EtaHits:muon_segment_nSTGC_PhiHits:muon_segment_ChamberTech")
***********************************************************************************************************************************************
*    Row   * Instance *  muons_pt * muons_eta * muons_aut * muons_tru * muon_segm * muon_segm * muon_segm * muon_segm * muon_segm * muon_segm *
***********************************************************************************************************************************************
*        0 *        0 * 14.984746 * 0.1195938 *         1 *         7 *         0 *         0 *         0 *         0 *         0 *         0 *
*        0 *        1 *           *           *           *           *         0 *         0 *         0 *         0 *         0 *         0 *
*        0 *        2 *           *           *           *           *         0 *         0 *         0 *         0 *         0 *         0 *
*        1 *        0 * 13.594388 * 2.3430957 *         8 *         0 *         1 *         0 *         0 *         0 *         0 *         0 *
*        1 *        1 * 3.0362854 * 1.5547844 *         6 *         0 *         2 *         0 *         0 *         0 *         0 *         0 *
*        1 *        2 * 2.7491135 * 0.8322315 *         6 *         0 *         3 *         4 *         4 *         8 *         6 *         4 *
*        1 *        3 * 2.7044379 * 1.6742800 *         6 *         0 *         4 *         0 *         0 *         0 *         0 *         0 *
*        1 *        4 * 2.4750762 *  2.154778 *         6 *         0 *           *           *           *           *           *           *
***********************************************************************************************************************************************
(long long) 8
root [2]

Checks on around 120k particle gun muons are currently processing. I'll attach the results as soon as they arrive.

Tagging: @rosati, @wleight, @pscholer, @iodice, @lucam, @stavrop

Edited by Johannes Junggeburth

Merge request reports

Loading