[PWGHF] Bugfixes of correlated background processing in 3-prong decays

[lubynets]

1. In evaluation of the swapping flag the case of $\Xi^+_c\rightarrow pK^-\pi^+$ decay was missing. This PR fixes it.
2. Removed Only from names of keep variables for signal, background and correlated background, since it is wrong (keeps are connected with logical OR).
3. Reflected $\Lambda^+_c\rightarrow pK^-\pi^+$ added into a definition of correlated background.

[github-actions]

O2 linter results: ❌ 0 errors, ⚠️ 0 warnings, 🔕 8 disabled

[vkucera]

@lubynets @mfaggin I think point 2 was not a good move because it fixes the "intent vs behaviour" mismatch at the wrong place.
AFAIK, the purpose of the keepOnly... configurables was indeed to store only a given subset of candidates. If the code behaviour doesn't match the intent, it is the code behaviour that needs to be fixed, not the names that communicate the intent.

[lubynets]

Hi @vkucera, I think after adding correlated BG by Mattia it is fully possible situation (from my experience) when one needs to save 2 out of 3 species - e.g. signal for efficiency determination and corrBG for its impact study. When there were only two "keeps", signal and BG, of course it did not make sense to enable both, but now it makes.