def main():
f = TFile(sys.argv[1])
jetmass = f.Get("jetm_qg")
# jetmass = f.Get("jetmass_avg")
jetmass_cut = get_jetmass_cut(jetmass)
fcns, fit_histo = slice_by_slice_fit(jetmass_cut, get_1d_fit_fcn)
diff_histo = jetmass_cut - fit_histo
# # NOTE: this is now done by utils.slice_by_slice_fit, but this makes a stackplot too
# # Fit the 1d function to each slice and store the fit parameters
# # xpoints = array.array('d')
# fcns = []
# # fitparams = [array.array('d') for i in range(NPARAM)]
# # fitparam_errs = [array.array('d') for i in range(NPARAM)]
# mjj_stack = THStack("mjjs", "") # Stack for each m_j, normalized to 1
# for x in range(jetmass_cut.GetNbinsX()):
# sliced_histo = slice_histo_y(jetmass_cut, x+1, "slice_%s" % (x+1), "M_jj, slice %s" % (x+1))
# fit_fcn = get_1d_fit_fcn(sliced_histo, x+1)
# sliced_histo.Fit(fit_fcn)
# fcns.append(fit_fcn)
# print "Slice %s chisq = %s" % (x+1, fit_fcn.GetChisquare())
# print
# print
# # import pdb; pdb.set_trace()
# # This doesn't do what I'd hoped:
# # fit_fcn.SetParameter(0, fit_fcn.GetParameter(0)/sliced_histo.Integral())
# sliced_histo.Scale(1.0/sliced_histo.Integral())
# mjj_stack.Add(sliced_histo)
# graph fit parameters
fitparams = [
array.array('d', [fit_fcn.GetParameter(i) for fit_fcn in fcns])
for i in range(NPARAM)
]
fitparam_errs = [
array.array('d', [fit_fcn.GetParError(i) for fit_fcn in fcns])
for i in range(NPARAM)
]
xpoints = array.array('d', [float(mj_fit_range[0] + mj_binsize * (i+0.5))
for i in range(len(fitparams[0]))])
no_errors = array.array('d', [0.0 for i in range(len(fitparams[0]))])
fitparam_graphs = [
TGraphErrors(len(fitparams[i]), xpoints, fitparams[i], no_errors, fitparam_errs[i])
for i in range(NPARAM)
]
print "total chisquare = %s" % sum(f.GetChisquare() for f in fcns)
print "total ndf = %s" % (jetmass_cut.GetNbinsX() * (jetmass_cut.GetNbinsY() - NPARAM))
import pdb; pdb.set_trace()
