By default only the btag reshape uncertainties related to the JES are dumped in the output, while the purity and statistical uncertainties are not dumped (see [1] for the btag reshape uncertainties recommendations).
the inclusion of the full set of btag uncertainties is straightforward [2]. This is already used in the resonant and non-resonant(HHcombination) HH->bbgg analyses but it was not propagated to the main flashgg branch.
from [1]: "It should be noted that JES is not an uncertainty intrinsic to the shape calibration method. The JES-varied scale factors are only provided to propagate the effect of various JES sources to the central b-tag SF measurement."
By default only the btag reshape uncertainties related to the JES are dumped in the output, while the purity and statistical uncertainties are not dumped (see [1] for the btag reshape uncertainties recommendations).
[1] https://twiki.cern.ch/twiki/bin/viewauth/CMS/BTagShapeCalibration#Systematic_uncertainties [2] https://github.com/fabio-mon/flashgg/commit/1f89c0aa64e4e4f607031811d5d22e6ca75b2741#diff-2bd0683a094a6649ac76f5e408715a0385a19da946214376abdec3de36a16f33