Keep any good tracks associated to AntiKt4EMPFlowJets in EXOT8 thinning

This merge request affects the EXOT8 derivation in the DerivationFramework.

In the track thinning, tracks associated to AntiKt4EMPFlowJets but not AntiKt4EMTopoJets nor AntiKt2LCTopoJets were discarded. Such tracks are now kept, provided they satisfy some track quality requirements. Since there is large overlap in the catchment areas of AntiKt4EMTopo jets and AntiKt4EMPFlow jets, both being clustered with R = 0.4, the number of additional tracks is small.

This change was requested by @jveatch in a discussion regarding MR #40942. It would allow investigating the feasability of constructing TAR jets using PFlow jets as inputs.

The size increase of the derivation as a result of the change was tested in files from 5 different MC samples. The size on disk increases by less than 0.5% w.r.t. the size in 21.2 (c.f. size_changes.txt). The output of the ART data test run (test_data18EXOT8.sh) increases from 66273709 bytes to 66558257 bytes (an 0.4% increase) due to the change.

Tagging @guescini for review by the Exotics Derivation contacts, @kkrizka, @mswiatlo, and @jveatch for EXOT8.