Fixes for fJVT implementation

[jdandoy]

• fJVT current uses CentralMaxPt, but should use ForwardMaxPt.
• ~It's not clear that JetForwardJvtTool is correct for PFlow jets, as there's also a JetForwardPFlowJvtTool. Currently following up with the ETmiss + pileup tagging subgroup.~ JetForwardJvtTool is the correct tool, JetForwardPFlowJvtTool performs a different task at the derivation stage.
• May want the option to use the derivation-level decoration, though note this may not have the correct pt threshold set, requiring an additional check before using.
• ~The fJVT tool currently doesn't use the correct maximum pt threshold, so we should set it correctly as a default option here.~ The R21.2 instance does set it correctly to 60 GeV.
• Could remove the hardcoding of the JVT likelihood file in anticipation of future updates.

[miholzbo]

FYI: In R22 we will (hopefully) handle this much more streamlined and use the JetJvtEfficiency as interface for all JVT-related taggers. My MR for athena is developing here and the associated branch to adjust the handling in xAH is in my xAH fork here.