Update for 2.5 compatibility

Fix bug in thcovmat construction

Seaquest data

Add option to use same trvl seed for all replicas

Added missing uncertainty in the CMS dijet 7 TeV NNLO K factor

Use a higher absolute and especially relative precision while still
being reasonable.

Use numpy's assert_allclose which has better error reporting.

Added cut to remove the first data point of the ATLAS W+jet pT measur…

Added cut to remove the first data point of the ATLAS W+jet pT measurement where resummation effects may be large

Add initialization of preproc ranges to docs

Add limits to preprocessing ranges in `vp-nextfitruncard`