[DPG] Refactor tpcSkimsTableCreator

[lubynets]

Reduced code repetition:

• merged process functions "with" and "without" dEdx correction;
• merged process functions "WithdEdxTrQA" and "WithTrQA";
• iteration by V0s (tracks) in V0 (TOF) tree creator replaced with a for-loop over a dedicated struct for mentioned objects;
• removed a redundant if-esle construction with identical consequent and alternative in struct TreeWriterTPCTOF.

Readability and code cosmetics:

• added const and const& where possible and necessary, removed const from function arguments passed by value;
• introduced constexpr shortcuts for particles' masses and ids;
• removed sqrtsNN from member functions' arguments since sqrtsNN is a struct's configurable;
• renamed TreeWriterTPCTOF::processStandard2() function into TreeWriterTPCTOF::processStandardWithCorrecteddEdx() for keeping the same naming style in the entire source file.

Minor bugfixes:

• use downsamplingTsalisTritons and downsamplingTsalisDeuterons variables for tritons and deuterons respectively (instead of downsamplingTsalisProtons);
• use Preclise perCollisionTracksWithCorrecteddEdx instead of perCollisionTracks in TreeWriterTPCTOF::processWithTrQAWithCorrecteddEdx().

Addressed linter error messages:

• introduced enums instead of hardcoded constants for track- and event-selection;
• removed underscore _ symbol from configurable names, configurable string names are made consistent with their variable names;
  full commit of linter error fixes.

FYI @amaringarcia

[github-actions]

O2 linter results: ❌ 1 errors, ⚠️ 0 warnings, 🔕 0 disabled

[vkucera]

@lubynets Nice work!
Did you use clang-tidy?
Why did you remove const from function arguments passed by value? Having const prevents them from being modified inside the function which is a valuable protection and indication of intent.

[vkucera]

Suggested change

	constexpr static double MassElectorn{o2::track::pid_constants::sMasses[PidElectron]};
	constexpr static double MassElectron{o2::track::pid_constants::sMasses[PidElectron]};

[lubynets]

Done

[vkucera]

Suggested change

	template <bool doUseCorrecteddEdx = false, typename T, typename C, typename V0Casc>
	void fillSkimmedV0Table(V0Casc const& v0casc, T const& track, C const& collision, float nSigmaTPC, float nSigmaTOF, float dEdxExp, o2::track::PID::ID id, int runnumber, double dwnSmplFactor, float hadronicRate)
	template <bool DoUseCorrectedDeDx = false, typename T, typename C, typename V0Casc>
	void fillSkimmedV0Table(V0Casc const& v0casc, T const& track, C const& collision, float nSigmaTPC, float nSigmaTOF, float dEdxExp, o2::track::PID::ID id, int runnumber, double dwnSmplFactor, float hadronicRate)

[lubynets]

Done

[vkucera]

Suggested change

	template <bool IsCorrecteddEdx, bool IsWithdEdx, typename TrksType, typename BCType>
	void runWithTrQAGeneric(Colls const& collisions, TrksType const& myTracks, V0sWithID const& myV0s, CascsWithID const& myCascs, aod::TracksQAVersion const& tracksQA, Preslice<TrksType> const& perCollisionTracksType)
	template <bool IsCorrectedDeDx, bool IsWithDeDx, typename TrksType, typename BCType>
	void runWithTrQAGeneric(Colls const& collisions, TrksType const& myTracks, V0sWithID const& myV0s, CascsWithID const& myCascs, aod::TracksQAVersion const& tracksQA, Preslice<TrksType> const& perCollisionTracksType)

[lubynets]

Done

[vkucera]

Why are these needed? The PDG masses are in the common constant header.

[lubynets]

Done (used common constant header).

[vkucera]

Suggested change

	template <bool doCorrectdEdx, bool isWithdEdx, typename T, typename TQA, typename C>
	void fillSkimmedTPCTOFTableWithTrkQAGeneric(T const& track, TQA const& trackQA, bool existTrkQA, C const& collision, float nSigmaTPC, float nSigmaTOF, float nSigmaITS, float dEdxExp, o2::track::PID::ID id, int runnumber, double dwnSmplFactor, double hadronicRate, int bcGlobalIndex, int bcTimeFrameId, int bcBcInTimeFrame)
	template <bool DoCorrectDeDx, bool IsWithDeDx, typename T, typename TQA, typename C>
	void fillSkimmedTPCTOFTableWithTrkQAGeneric(T const& track, TQA const& trackQA, bool existTrkQA, C const& collision, float nSigmaTPC, float nSigmaTOF, float nSigmaITS, float dEdxExp, o2::track::PID::ID id, int runnumber, double dwnSmplFactor, double hadronicRate, int bcGlobalIndex, int bcTimeFrameId, int bcBcInTimeFrame)

[lubynets]

Done

[vkucera]

Suggested change

	const auto trackocc = collision.trackOccupancyInTimeRange();
	const auto ft0occ = collision.ft0cOccupancyInTimeRange();
	const auto trackOcc = collision.trackOccupancyInTimeRange();
	const auto ft0Occ = collision.ft0cOccupancyInTimeRange();

[lubynets]

Done

[vkucera]

Suggested change

	double nSigmaTofTpctof;
	double nSigmaTpcTpctof;
	double nSigmaTofTpcTof;
	double nSigmaTpcTpcTof;

[lubynets]

Since initially it was nSigmaTPC_TPCTOF, I think the first part TPC must be separated from the second part TPCTOF. That's why I merged Tpctof in a single substring.

[lubynets]

@lubynets Nice work! Did you use clang-tidy? Why did you remove const from function arguments passed by value? Having const prevents them from being modified inside the function which is a valuable protection and indication of intent.

@vkucera, thanks!
No, I did not use clang-tidy.
About const arguments by value, honestly I had a feeling that the const should not be remove exactly because of the reasons which you mentioned. But I have read that it's a common practice not to pass arguments by value with const - isn't it the case for O2Physics? Or maybe I am missing something, and there is no such a "rule" in general?

[vkucera]

@lubynets Nice work! Did you use clang-tidy? Why did you remove const from function arguments passed by value? Having const prevents them from being modified inside the function which is a valuable protection and indication of intent.

@vkucera, thanks! No, I did not use clang-tidy. About const arguments by value, honestly I had a feeling that the const should not be remove exactly because of the reasons which you mentioned. But I have read that it's a common practice not to pass arguments by value with const - isn't it the case for O2Physics? Or maybe I am missing something, and there is no such a "rule" in general?

I don't think there is a rule for it in O2Physics. The C++ Core Guideline Con.1 considers it to be "pedantic" though. :-)

Btw, the misc-const-correctness check in clang-tidy can fix the missing const automatically.

[lubynets]

@lubynets Nice work! Did you use clang-tidy? Why did you remove const from function arguments passed by value? Having const prevents them from being modified inside the function which is a valuable protection and indication of intent.

@vkucera, thanks! No, I did not use clang-tidy. About const arguments by value, honestly I had a feeling that the const should not be remove exactly because of the reasons which you mentioned. But I have read that it's a common practice not to pass arguments by value with const - isn't it the case for O2Physics? Or maybe I am missing something, and there is no such a "rule" in general?

I don't think there is a rule for it in O2Physics. The C++ Core Guideline Con.1 considers it to be "pedantic" though. :-)

Btw, the misc-const-correctness check in clang-tidy can fix the missing const automatically.

Ok, I mixed it with another practice - avoiding const in function declaration when it is separated from the definition (link).

[lubynets]

Dear @alcaliva, thanks for enabling auto-merge! Could you please merge the PR?