skeleton for O2 analysis with multiparticle correlations

[abilandz]

No description provided.

[victor-gonzalez]

There is not problem on creating the MultiparticleCorrelations as the PAG working area but I will recommend the following structure

PWGCF/MultiparticleCorrelations/Core
PWGCF/MultiparticleCorrelations/DataModel
PWGCF/MultiparticleCorrelations/TableProducer
PWGCF/MultiparticleCorrelations/Tasks

In Core you can put your header files although in the way you use them they apply more to the Tasks directory where perhaps you can create a Headers sub-directory. Be aware that if you define variables in header files they have to be handled with care if they can be 'included' by different entities. But it seems it is not your case.
Core would be more applicable to internal structures or classes for your functional area which will end up in a library

In DataModel you would put the tables definition that apply to your functional area. None for the time being.

In TableProducer you would put the tasks that populate internal tables which you will use in your functional area. None for the time being.

In Tasks you put your functional area tasks ;-)

Each directory should have its own CMakeFiles.txt file which addresses how the directory is built (the empty directories don't need to be created or having that txt file)
At the PWGCF/MultiparticleCorrelations level should be another CMakeFiles.txt which will incorporate the different directories under PWGCF/MultiparticleCorrelations (only the ones that will have CMakeFiles.txt file).

Finally, the CMakeFiles.txt at PWGCF level should incorporate the MultiparticleCorrelations directory.

[abilandz]

Dear Victor,

Thanks a lot for your detailed feedback - I have followed your suggestions in the new commit!

For symmetry, I have supplied also CMakeLists.txt files in empty subdirectories MultiparticleCorrelations/TableProducer and MultiparticleCorrelations/DataModel - at the moment those CMake files have only copyright info.

Please let us know if any further action from our side would be needed!

Cheers,
Ante

[victor-gonzalez]

The structure is ready to be approved

Have a look at the detailed comments I left regarding the use of the HistogramRegistry and the access to alien files (they also apply to the files in their new places) and let me know your opinion

[abilandz]

Further feedback on this exchange: I just did the first test run in Hyperloop for my first O2 executable ( https://alimonitor.cern.ch/hyperloop/wagon-test/28805 ) , and everything went perfectly fine! I could access the custom file on Grid in init(...), initialize the data members of my task with its content, and then use that content via data members in process(...) without any performance or I/O penalty.

Thanks a lot to the O2 community for making this possible!

[abilandz]

Final feedback from my side: Also the centrally submitted hyperloop went through perfectly fine: https://alimonitor.cern.ch/hyperloop/train-run/28808