PWGHF: Add ThnSparse of the ML scores vs Vars to the Lc task

[zhangbiao-phy]

No description provided.

[fgrosa]

Hi @zhangbiao-phy thanks a lot! I have few comments (the most important about the maximum number of axes), see below

[zhangbiao-phy]

Hi @zhangbiao-phy thanks a lot! I have few comments (the most important about the maximum number of axes), see below

Hi @fgrosa, Thanks for the review! Very helpful! I commit the changes based on your suggestion and comments.

[fgrosa]

Hi @zhangbiao-phy thanks for implementing my comments! I have a couple of additional suggestions, then it's good for me.

[alibuild]

Error while checking build/O2Physics/o2 for f938929 at 2024-03-08 12:40:

## sw/BUILD/O2Physics-latest/log
/sw/SOURCES/O2Physics/5040-slc7_x86-64/0/PWGHF/D2H/Tasks/taskLc.cxx:52:92: error: redeclaration of 'o2::framework::ConfigurableAxis HfTaskLc::thnConfigAxisBdtScoreLc'
/sw/SOURCES/O2Physics/5040-slc7_x86-64/0/PWGHF/D2H/Tasks/taskLc.cxx:278:43: error: 'thnConfigAxisBdtScoreBkg' was not declared in this scope; did you mean 'thnConfigAxisBdtScoreLc'?
/sw/SOURCES/O2Physics/5040-slc7_x86-64/0/PWGHF/D2H/Tasks/taskLc.cxx:278:89: error: no matching function for call to 'o2::framework::AxisSpec::AxisSpec(<brace-enclosed initializer list>)'
/sw/SOURCES/O2Physics/5040-slc7_x86-64/0/PWGHF/D2H/Tasks/taskLc.cxx:279:46: error: 'thnConfigAxisBdtScoreSignal' was not declared in this scope; did you mean 'thnConfigAxisBdtScoreLc'?
/sw/SOURCES/O2Physics/5040-slc7_x86-64/0/PWGHF/D2H/Tasks/taskLc.cxx:279:98: error: no matching function for call to 'o2::framework::AxisSpec::AxisSpec(<brace-enclosed initializer list>)'
/sw/SOURCES/O2Physics/5040-slc7_x86-64/0/PWGHF/D2H/Tasks/taskLc.cxx:280:105: error: no matching function for call to 'o2::framework::AxisSpec::AxisSpec(<brace-enclosed initializer list>)'
ninja: build stopped: subcommand failed.

Full log here.

[vkucera]

@zhangbiao-phy , your changes replaced the filter with partitions which broke the candidate grouping by collision. As a result, all candidates are now processed multiple times, once for each collision. To fix the grouping, please add slicing per collision index.
Also, the added configurable applyMl is misleading, as it only activates histograms filling and does not apply any ML selection, which is only done in the selector.
Please, always test your changes before making a PR.

[vkucera]

Tagging @DelloStritto to keep him in the loop.

[zhangbiao-phy]

@zhangbiao-phy , your changes replaced the filter with partitions which broke the candidate grouping by collision. As a result, all candidates are now processed multiple times, once for each collision. To fix the grouping, please add slicing per collision index. Also, the added configurable applyMl is misleading, as it only activates histograms filling and does not apply any ML selection, which is only done in the selector. Please, always test your changes before making a PR.

Hi @vkucera, Thanks for spotting this! It's a mistake. actually, I have tested it on my laptop. I will make a PR soon.