Skip to content

Future seeding fix found ntracks flagging

Renato Quagliani requested to merge future-seeding-fixFoundNtracksFlagging into future

This branch should Fix the hit flagging in track recovering routine to get the same efficiencies and ghost rate of master branch. As far as i undestand ( maybe @raaij and @graven can comment if i am wrong in the understanding ) Before: The check for NbUsed hits were done without forwarding the FTHitHandler. The PrHybridSeedTrack contains m_hits as vector , so the hit content would have no info at all about what gets flagged after creating the track object and adding the hit.

--> note that ModPrHit contains a pointer to the actual hit in Tes and the flag information.

The trick used here was to flag ALL PrHit in track ( they were pointers before, so they were really modifying things in TES) and then when counting on another set of tracks the nbHits used we implicitely keeped track of what was done before.

NB: probably the same problem is present in Forward tracking algorithm.

The branch contains also other improvements from Future seeding minor branch ( !472 (merged) )

Checking on master branch and with the modification:

master branch

PrChecker2.TTrack    INFO **** TTrack                           1166 tracks including             83 ghosts [ 7.12 %], Event average  5.24 % ****
PrChecker2.TTrack    INFO                          01_hasT :    1111 from     1376 [ 80.74 %]      0 clones [ 0.00 %], purity: 99.57 %, hitEff: 96.33 %
PrChecker2.TTrack    INFO                          02_long :     772 from      831 [ 92.90 %]      0 clones [ 0.00 %], purity: 99.64 %, hitEff: 96.38 %
PrChecker2.TTrack    INFO                     03_long>5GeV :     566 from      588 [ 96.26 %]      0 clones [ 0.00 %], purity: 99.58 %, hitEff: 96.76 %
PrChecker2.TTrack    INFO                    04_long_fromB :      35 from       37 [ 94.59 %]      0 clones [ 0.00 %], purity: 99.22 %, hitEff: 95.40 %
PrChecker2.TTrack    INFO               05_long_fromB>5GeV :      30 from       32 [ 93.75 %]      0 clones [ 0.00 %], purity: 99.09 %, hitEff: 94.93 %
PrChecker2.TTrack    INFO                  06_UT+T_strange :      53 from       59 [ 89.83 %]      0 clones [ 0.00 %], purity: 99.65 %, hitEff: 97.33 %
PrChecker2.TTrack    INFO             07_UT+T_strange>5GeV :      33 from       34 [ 97.06 %]      0 clones [ 0.00 %], purity: 99.70 %, hitEff: 98.94 %
PrChecker2.TTrack    INFO           08_noVelo+UT+T_strange :      20 from       22 [ 90.91 %]      0 clones [ 0.00 %], purity: 99.50 %, hitEff: 97.71 %
PrChecker2.TTrack    INFO      09_noVelo+UT+T_strange>5GeV :      14 from       15 [ 93.33 %]      0 clones [ 0.00 %], purity: 99.29 %, hitEff: 98.70 %

"this" branch

PrChecker2.TTrack    INFO **** TTrack                           1166 tracks including             83 ghosts [ 7.12 %], Event average  5.24 % ****
PrChecker2.TTrack    INFO                          01_hasT :    1111 from     1376 [ 80.74 %]      0 clones [ 0.00 %], purity: 99.57 %, hitEff: 96.33 %
PrChecker2.TTrack    INFO                          02_long :     772 from      831 [ 92.90 %]      0 clones [ 0.00 %], purity: 99.64 %, hitEff: 96.38 %
PrChecker2.TTrack    INFO                     03_long>5GeV :     566 from      588 [ 96.26 %]      0 clones [ 0.00 %], purity: 99.58 %, hitEff: 96.76 %
PrChecker2.TTrack    INFO                    04_long_fromB :      35 from       37 [ 94.59 %]      0 clones [ 0.00 %], purity: 99.22 %, hitEff: 95.40 %
PrChecker2.TTrack    INFO               05_long_fromB>5GeV :      30 from       32 [ 93.75 %]      0 clones [ 0.00 %], purity: 99.09 %, hitEff: 94.93 %
PrChecker2.TTrack    INFO                  06_UT+T_strange :      53 from       59 [ 89.83 %]      0 clones [ 0.00 %], purity: 99.65 %, hitEff: 97.33 %
PrChecker2.TTrack    INFO             07_UT+T_strange>5GeV :      33 from       34 [ 97.06 %]      0 clones [ 0.00 %], purity: 99.70 %, hitEff: 98.94 %
PrChecker2.TTrack    INFO           08_noVelo+UT+T_strange :      20 from       22 [ 90.91 %]      0 clones [ 0.00 %], purity: 99.50 %, hitEff: 97.71 %
PrChecker2.TTrack    INFO      09_noVelo+UT+T_strange>5GeV :      14 from       15 [ 93.33 %]      0 clones [ 0.00 %], purity: 99.29 %, hitEff: 98.70 %

Merge request reports