Daniil Ponomarenko and Yury Smirnov TRD test beam analysis meeting November 6… · 2019. 12. 1. · Daniil Ponomarenko and Yury Smirnov TRD test beam analysis meeting November 6,

Online monitoring

Daniil Ponomarenko and Yury Smirnov

TRD test beam analysis meeting

November 6, 2019

• To create a GUI that would take the raw-data files, run the entire reconstruction, analysis, and show the final

plots;

• every next step should be able to take the results of the previous one;

• everything should run non-interactively (exceptions are possible but discouraged – see next slides);

• the code of these four packages is taken from live copies (“mirrors”) of the respective repositories, treated

as git submodules;

• when a developer of any of these packages pushes something into his repository, this change will

propagate to the mirror, and we may choose to accept it on a machine wherever it is supposed to run (but

nothing gets updated if no actions are taken);

• it’s still up to the original developers to maintain their code to comply with this workflow!2/14

The idea

Yury Smirnov November 6th 2019 TRD test beam analysis meeting

Data unpacking

Event building

Pixel-energy calibration and clusterization

TTreeanalysis

histograms

(“unified format”)

3/14

How it looks now

https://gitlab.cern.ch/trtTestBeam-experts/TRStand_pixel/monitoring_gui/tree/dev_Yury





4/14

Compile-time issues


Data unpacking

Event building


TTreeanalysis

histograms

• These three require different environments to compile;

• the exact requirements are present in the README of the 3rd repository only;

• would be nice to know the requirements for the first two.

• I figured how to compile all three, but only on lxplus;

• not sure if the same method will work on a computer where it’s supposed to run – probably not (see next

slides).

5/14

Run-time issues


Data unpacking

Event building


TTreeanalysis

histograms

• These two have the paths and names of input files hardcoded – this has to be changed.

6/14

Run-time issues


Data unpacking

Event building


TTreeanalysis

histograms

• This one is configured via a config file, i.e. interactively;

• we can ask a user to input the distance between the detector and radiator and whether he/she wants clustering via edges

only or via edges+corners, rewrite the config file on-the-fly and run;

• but these are only two parameters, there are more:

• calibration constants;

• whether we want to apply the time-walk correction;

• RoI definition (we usually throw away a few columns and rows of pixels too close to the detector edge);

• etc.

• Wouldn’t it make more sense to have the

in the config file by default instead of

?

good_event_flags = [2,4]

good_event_flags = [0,2]

7/14

Run-time issues


Data unpacking

Event building


TTreeanalysis

histograms

• Processed several runs with electron and muon/pion triggers;

• electron runs are okay, but muon/pion runs aren’t;

• no matter what I have in the config.toml, trigger 4 keeps getting ignored;

• ergo, no muons and no pions.

• What’s wrong? How do we usually run it?

o Config-file content - ?

o Other command-line parameters - ?

o Input files - ?

o Should it even be the unified-format code?

8/14

Run-time issues


Data unpacking

Event building


TTreeanalysis

histograms

• Electron/pion/muon triggers issue:

• up to now, we relied on the run number to figure out the desired trigger selection: 𝑒 / 𝜇 / 𝜋;

• we can introduce something like the following, but this will pop up for every run we process:

• Maybe makes more sense to introduce it in the previous step and propagate user’s choice to the

analysis.

9/14

Environment issues

Data unpacking

Event building


TTreeanalysis

histograms

• All four compile and run on lxplus (the first two definitely run from outside of the GUI tool);

• The computer they are supposed to run eventually has Windows installed;

• Planning to use Docker to run everything in;

• Probably only one ROOT version will be available.


10/14

How it works so far


11/14

How it works so far

• Choosing an input file for step 3 (since the first two aren’t supported yet);

• Let it be electrons, 2m, 90 Mylar foils, 50/3.


12/14

How it works so far


13/14

three minutes later…


14/14

How it works so far


THANKS!

Daniil Ponomarenko and Yury Smirnov TRD test beam analysis meeting November 6… · 2019. 12. 1. · Daniil Ponomarenko and Yury Smirnov TRD test beam analysis meeting November 6,

Documents