Revisit lines for strange decays

There are a few follow-up items that were missing after !1605 (merged) and must be understood/implemented for Run 3:

• Understand the low (null) efficiency for modes with electrons, where the invariant mass histograms produced from signal samples of, for example K_S^0\to e^+e^- report zero signal candidates.
• Remove the ghost-probability requirements from the selections and verify that the rates are still low enough
• Possibly add a CI-test in order to test the efficiency of the lines and ensure that changes to the different algorithms do not impact the efficiencies drastically. This is motivated from the fact that people tend to apply requirements at the basic-selection/reconstruction level that can be really harmful for strange decays (like p_T requirements on tracks, which might completely kill the signals).
• Understand and fix the issues with the monitoring of certain lines (like for K_S^0 \rightarrow e^+e^- which is not being done for unknown reasons)
• Understand the null efficiency of the K_S^0 \rightarrow e^+ e^- e^+ e^- line