Tracking/TrackHitContributions/hit_matching: 2DStrip-compatibility.

[ybedfer]

i) Association efficiencies of 'p' and 'n' strips are histo'd independently, as well as simultaneous 'p&n'.
ii) r vs. z: 'p' and 'n' RecHits are combined to obtain precise 3D info. For the BarrelOuterTracker, this is achieved by retrieving WorldToLocal transforms from a TGeometry file.
Also: some cosmetics changes, e.g.: input filename, output to ROOT file.

Briefly, what does this PR introduce? Please link to any relevant presentations or discussions.

What is the urgency of this PR?

• [ *] Medium

What kind of change does this PR introduce?

• [* ] Bug fix (issue = processing of 2DStrip RecHits)

[ybedfer]

Hello,

This PR requires PR #1077 of `epic (eic/epic#1077).

This PR improves the processing of 2DStrip hits in the hit_matching script, which otherwise yields a bizarre parabolic shape for the r vs. z distribution of the BarrelOuterTracker:

It also introduces a new feature, viz. the independent histogramming of the association efficiency of 'p' and 'n' strips:

The processing of pixel (as opposed to 2DStrip) data keeps being unchanged:

In order to bring into play the special processing of 2Dstrip data, one needs to symlink a TGeometry file to hit_matching.geometry.root in the $PWD.

Note that concerning the association efficiency histo:

• The so many hits obtained for MPGDBarrelRecHits and OuterMPGDBarrelRecHits are due to the factor 2 in # of entries from 'p' and 'n' RecHits × the factor 5 of the Multiple Sensitive SubVolume implementation of 2DStrip.
• There is also here and there a larger fraction of associated outliers:
  • in the 2DStrip MPGDs: I guess it can be easily explained by the more difficult track reconstruction task when strips instead of pixels.
  • in the SiEndcapTracker: there, I don't know...
• In the plots above, only the Barrel (as opposed to Endcaps) MPGDs are 2DStrip.

[ybedfer]

Note: I haven't followed the indentation style of the original file. For various reasons:

• it's inconvenient, given the amount of nested loops,
• it does not correspond to what's used in the rest of ePIC software.

[bschmookler]

Hi Yann. Thanks a lot for updating this.

Should we update the README (https://github.com/eic/snippets/blob/main/Tracking/TrackHitContributions/README.md) to include the instructions for including the TGeometry file, and how to toggle the verbose flag and max number of events?

[ybedfer]

Hello,
I committed a new README.md.

While doing this I tried to execute the recommended (in README) procedure, viz.:

root -l -b -q hit_matching.C

to check that it still works now that I added arguments to the macro (since those have default values, there is a priori no reason that I shouldn't work). But then I got the following error:

Processing /local/home/bedfer/eicA/snjppets/Tracking/TrackHitContributions/hit_matching.C...
libedm4hepDict dictionary payload:741:7: error: field has incomplete type '::edm4hep::Vector3f'
      referencePoint{}; ///< Reference point of the track parameters, e.g. the origin at the IP, or the position  of the
      ^
libedm4hepDict dictionary forward declarations' payload:8:26: note: forward declaration of 'edm4hep::Vector3f'
namespace edm4hep{class  Vector3f;}
                         ^
_etc_..._etc_...

This, whether I execute my new version or the old one.

[bschmookler]

Hi Yann. I see this issue with the script in the main branch also. It seems to be a CLING issue. I was able to resolve it by compiling the script:

root -l -b -q hit_matching.C+

I'll check if it still works for the version here.

[bschmookler]

Hi Yann. The code seems to run file if I compile it (with the + option). I added this information to the README.

There were two warnings when compiling the code.

You had this:

double lpos[3], Ur, Vr;

Ur and Vr are only assigned inside if (geometry->WorldToLocal(...)). So, I initialized them to zero

You also had this:

double X, Y;

X is only assigned when pn == 0 and Y only when pn != 0. So, I again initialized them to zero.

If you think that is reasonable, then we can merge this into the main branch now.

[ybedfer]

OK