[PWGHF] KFParticle reconstruction of 3-prong decay

[lubynets]

1. In this PR the usage of Kalman Filter Particle algorithm is implemented for 3-body decay reconstruction (by way of analogy with 2-body decay).
2. A table with KF-specific variables (mainly topological) is added.
3. Changes related to item 1 are introduced into selection and treeCreator tasks.

[fgrosa]

Hi @fgrosa, thanks for your feedback and proposed changes, I implemented them. What about stability of the DCAFitter part of the code and memory usage - I checked both. I have run reconstruction of Lc using DCAfitter both in my version of the code, proposed in this PR, and in master version. The results are identical, see distributions of invariant mass and azimuthal angle - they are on top of each other (master and KF). Also I looked through other distributions - their number of entries, means and std devs agree in all digits that allows me to conclude that the results are identical. What about memory usage I also built both for master and KF versions the proportionalSetSize graphs (if I decoded PSS in your comment correctly) - they also look similar in my opinion. Are other values in performance graphs also worth to be compared? I placed generated AnalysisResults.root files and performanceMetrics.json files by this link https://cernbox.cern.ch/s/XUukbYeVeJtb8Ja

Thanks! It looks good to me. Approved now, even if we cannot merge yet because an issue in the CI with macOS-arm (will do as soon as solved)

[vkucera]

@lubynets Please fix the MegaLinter errors.

[lubynets]

@lubynets Please fix the MegaLinter errors.

Done

[vkucera]

@lubynets Thanks for addressing my comments.
Not strictly required for this PR, but there is a lot of duplication between fillTablesData and fillTablesMc. It would be worth exploring the possibility to reduce it with refactoring that code.

[fgrosa]

@lubynets Thanks for addressing my comments. Not strictly required for this PR, but there is a lot of duplication between fillTablesData and fillTablesMc. It would be worth exploring the possibility to reduce it with refactoring that code.

I agree on this, but I also agree that it can be done in a second PR (@lubynets please consider it) in order to move on with this one. I cannot merge however because the utils header was included in Tools, for which I have no approval rights (maybe @ktf can merge? Thanks!)