TrackState Covariance matrix

[andresailer]

Replacing: key4hep/k4LCIOReader#8

At the moment the LCIO covariance matrix is lower triangle, while the one in EDM4hep is the upper triangle.

http://lcio.desy.de/v02-09/doc/doxygen_api/html/classEVENT_1_1TrackState.html#a667b8774ce1b0f42dbae11a59dc7e09b

EDM4hep/edm4hep.yaml

Line 63 in e17d47a

- std::array<float, 15> covMatrix // upper triangular covariance matrix of the track parameters. the order of parameters is d0, phi, omega, z0, tan(lambda).

It was decided to change the edm4hep one to lower triangle as well.

Is there any sane way to prevent existing places where this conversion was done after reading the documentation to not silently break?

[andresailer]

For example
https://github.com/key4hep/k4SimDelphes/blob/8561d3fc39b2accc99c62a561090421fad3f1e8e/converter/src/DelphesEDM4HepConverter.cc#L488-L493

[clementhelsens]

Hi @andresailer , thanks for deciding :)
As far as I know, this is the only place where the covariance matrix is filled.

[andresailer]

I didn't decide this: https://indico.cern.ch/event/1000558/

[clementhelsens]

sorry, I misread you comment and missed the "t was" :)
Anyway, this part of the code

https://github.com/key4hep/k4SimDelphes/blob/8561d3fc39b2accc99c62a561090421fad3f1e8e/converter/src/DelphesEDM4HepConverter.cc#L488-L493

will have to be changed once we have properly patched Delphes to get the full matrix.

[andresailer]

I should have linked the minutes in the ticket from the start.

The delphes part can already be changed to assume lower triangle though. Still basically makes a mess that there will be files with one and some with a different interpretation.

[clementhelsens]

For FCC-ee, I do not think this is problematic. I will anyway remove all the files we have been producing so far for a first clean production where I want to have the full matrix stored, so no show stopper from my side a least!