New systematics, new definition of M(3j) and new definition of the regions

I reoptimized the cuts, using the fit of the Pt of the neutrino coming from the Higgs and the redefinition of the invariant mass of the three jets M(3j). This mass uses the closest bjet coming from the top plus the two closest jets to this bjet. New theoretical uncertainties are taken into account and merged in the combine_rootfile.py code as two global qcdscale and pdfas systematics. There is probably a bug in the qcdscale systematic for the moment which needs to be resolved. The regions will maybe be redefined, where the tt2l and tt1l control regions will be used as measurement regions for the flip and fake leptons respectively. Two codes, compare_flip.py and compare_fake.py, are added to compare the shapes of these processes in all the regions. Mtt' is used as the observable in the ttX control region. Minor processes have been merged into one single 'others' process to create the datacard which includes the fake and flip leptons.