TPC PID Skim Tree Creation

[AnnalenaKa]

This pull request will add a check in the TPC PID Response class whether a tpc signal was present and the default values are changed to -999 for the resolution and number of sigma and to 0 for the expected signal.

In addition, a column with the expected signal is added to the TPC pid table, which is produced by the pid-tpc-full task.

Furthermore, the task DPG/Tasks/TPC/tpcSkimsTableCreator.cxx will allow to create a table, which will be used to train the neural network that is used for the multidimensional correction of the TPC response. The task is based on V0 and gamma decays plus TPC and TOF track selections at low momentum.

[AnnalenaKa]

Hi,
@jezwilkinson , @ChSonnabend and @njacazio.
I have created the draft pull request adding the following changes:

1. Some hasTPC() checks in the TPCPIDResponse.h class. Furthermore, the default values were changed to -999 for the sigma and nsigma and to 0 for the expected signal.

2. I had to add the expected signal to the tpc table which is produced by the pid-tpc-full task. It was stored for the TOF but not for the TPC and we need it for our tree creation. Otherwise we will have to copy parts of the tpc full task in order to get the expected signal. If Nicolo, you don't agree with that change I can revert that part and change the way we get the expected signal for the tree creation and get the response from the CCDB or a file in addition to running the pid-tpc-full task.

3. I added the skim tree creation task under DPG/Tasks/TPC. Please let me know if you want me to add some modifications.

[jezwilkinson]

Thanks for the commit! I had a couple of comments on the changes to the TPC response - having the NumberOfSigmas return -999 in case of an invalid signal is good, but i think returning -999 on the resolution could have unintended effects if it's not then checked for by the other functions that use it

[wiechula]

Minor comment: Please carefully check which variables can be declared const.

[njacazio]

Hi @AnnalenaKa this looks super nice indeed! I have just some minor comments/suggestions.
Did you estimate the compression factor of the skimmed data WRT the original AODs?

Also I'm wondering if you are interested into carrying out some down-sampling at some point.

[AnnalenaKa]

Hi,
@jezwilkinson @wiechula and @njacazio
I have included all your comments I received here or in the meeting. Please let me know if you have additional wishes or comments. If everyone is satisfied with the changes I would mark the request as ready for review. As you can see I included also the event selection and track selection in the skimming task.

[njacazio]

Thanks a lot @AnnalenaKa very nice, all good from me

[wiechula]

Some more extremely minor suggestions.
Also I would prefer to explicitly use the std:: math functions:

pow,TMath::Power -> std::pow
abs,TMath::Abs -> std::abs
sqrt -> std::sqrt

etc.

[jezwilkinson]

All looks good to me, thanks for the updates!

[mfaggin]

I see that this PR was hugely discussed already, green light from my side then.

[njacazio]

I would just wait #1025 to be merged

[wiechula]

Thanks a lot @AnnalenaKa looks good!

[AnnalenaKa]

I would just wait #1025 to be merged

Hi @njacazio I saw this pull request was merged. Could we proceed with the review now to merge this PR as well?

[njacazio]

I would just wait #1025 to be merged

Hi @njacazio I saw this pull request was merged. Could we proceed with the review now to merge this PR as well?

Sure, for me it's already fine! @jgrosseo can we have this merged?

[njacazio]

I fixed a conflict brought in by a recent PR merge that prevented the merge

[ddobrigk]

Stalled PR, looks ok and reviewed, approved