def test_pdf():
effPdf = Models.PolyPos_pdf('B', xvar=x, power=6)
eff2 = Efficiency1D('E2', effPdf, cut=acc)
r2 = eff2.fitTo(ds)
f2 = eff2.draw(ds)
print r2
for p in points:
print ' Point/Eff %.1f %s%%' % (p, (
100 * eff2(p, error=True)).toString('(%5.2f+-%4.2f)'))
def test_poly4():
logger = getLogger('test_poly4')
logger.info("Test Poly(4)-Distribution")
model = Models.PolyPos_pdf('P4', x, power=3)
result, f = model.fitTo(dataset, silent=True, refit=5)
with wait(1), use_canvas('test_poly4'):
model.draw(dataset)
title = 'Positive polynomial'
logger.info('%s \n%s' % (title, result.table(title=title, prefix='# ')))
models.add(model)
def test_poly4():
logger.info("Test Poly(4)-Distribution")
model = Models.PolyPos_pdf('P4', x, power=4)
result, f = model.fitTo(dataset, silent=True)
if 0 != result.status() or 3 != result.covQual():
logger.warning('Fit is not perfect MIGRAD=%d QUAL=%d ' %
(result.status(), result.covQual()))
print(result)
else:
for phi in model.phis:
logger.info("\tPoly4: phi= %-17s " % phi.ve())
models.add(model)
def test_simfit5():
logger = getLogger('test_simfit5')
# =========================================================================
## resoltuion for the left peak
# =========================================================================
reso1 = Models.ResoGauss('RG1',
xvar=dm1,
sigma=(sigma1, sigma1 / 2, sigma1 * 2),
mean=(0, -1, 1))
# fit DM1 dataset for resolution
r1, f1 = reso1.fitTo(dset1, silent=True)
r1, f1 = reso1.fitTo(dset1, silent=True, draw=True, nbins=100)
# =========================================================================
## resoltuion for the central peak
# =========================================================================
reso2 = Models.ResoGauss('RG2',
xvar=dm2,
sigma=(sigma2, sigma2 / 2, sigma2 * 2),
mean=(0, -1, 1))
# fit DM2 dataset for resolution
r2, f2 = reso2.fitTo(dset2, silent=True)
r2, f2 = reso2.fitTo(dset2, silent=True, draw=True, nbins=100)
# =========================================================================
## resoltuion for the right peak
# =========================================================================
reso3 = Models.ResoGauss('RG3',
xvar=dm3,
sigma=(sigma3, sigma3 / 2, sigma3 * 2),
mean=(0, -1, 1))
# fit DM2 dataset for resolution
r3, f3 = reso3.fitTo(dset3, silent=True)
r3, f3 = reso3.fitTo(dset3, silent=True, draw=True, nbins=100)
# =========================================================================
# model for data fit
# =========================================================================
signal1 = Models.Gauss_pdf('G1',
xvar=mass,
sigma=reso1.sigma,
mean=(mean1, mean1 - 1, mean1 + 1))
signal2 = Models.Gauss_pdf('G2',
xvar=mass,
sigma=reso2.sigma,
mean=(mean2, mean2 - 1, mean2 + 1))
signal3 = Models.Gauss_pdf('G3',
xvar=mass,
sigma=reso3.sigma,
mean=(mean3, mean3 - 1, mean3 + 1))
background = Models.PolyPos_pdf('Bkg', xvar=mass, power=1)
model = Models.Fit1D(signals=[signal1, signal2, signal3],
background=background)
model.S = NS1, NS2, NS3
model.B = NB
# =========================================================================
## Fit data (without resolution samples)
# =========================================================================
r0, f0 = model.fitTo(dataset, silent=True)
r0, f0 = model.fitTo(dataset, silent=True, draw=True, nbins=100)
logger.info('Fit result (only data):\n%s' % r0.table(prefix="# "))
## combine PDFs
model_sim = Models.SimFit(category, {
'data': model,
'dm1': reso1,
'dm2': reso2,
'dm3': reso3
},
name='X')
# =========================================================================
## Simultanegous fit data with resolution samples)
# =========================================================================
r_1, f_1 = model_sim.fitTo(cdataset1, silent=True)
r_1, f_1 = model_sim.fitTo(cdataset1, silent=True)
logger.info('Simultaneous fit result (unbinned reso):\n%s' %
r_1.table(prefix="# "))
with use_canvas('test_simfit5'), wait(1):
with wait(1):
fdm1 = model_sim.draw('dm1', cdataset1, nbins=100)
with wait(1):
fdm2 = model_sim.draw('dm2', cdataset1, nbins=100)
with wait(1):
fdm3 = model_sim.draw('dm3', cdataset1, nbins=100)
with wait(1):
fd = model_sim.draw('data', cdataset1, nbins=100)
# =========================================================================
## Simultanegous fit data with resolution samples)
# =========================================================================
r_2, f_2 = model_sim.fitTo(cdataset2, silent=True, sumw2=True)
r_2, f_2 = model_sim.fitTo(cdataset2, silent=True, sumw2=True)
logger.info('Simultaneous fit result (binned reso):\n%s' %
r_2.table(prefix="# "))
with use_canvas('test_simfit5'), wait(1):
with wait(1):
fdm1 = model_sim.draw('dm1', cdataset2, nbins=100)
with wait(1):
fdm2 = model_sim.draw('dm2', cdataset2, nbins=100)
with wait(1):
fdm3 = model_sim.draw('dm3', cdataset2, nbins=100)
with wait(1):
fd = model_sim.draw('data', cdataset2, nbins=100)
import ROOT
import ostap.fitting.models as Models
from ostap.core.core import cpp
#variables
imc = ROOT.RooRealVar('imc', 'imc', 1., 3.)
pk = ROOT.RooRealVar('pk', 'peak', 0.)
#signal
bw = cpp.Ostap.Math.BreitWigner(pk)
sig_bw = Models.BreitWigner_pdf('sig_bw', breitwigner=bw, xvar=imc)
sig_bw.gamma.setVal(0.20)
sig_bw.gamma.setError(0.01)
sig_bw.mean.setVal(2.10)
sig_bw.mean.setError(0.10)
#background
bkg0 = Models.PolyPos_pdf('bkg', imc, power=1)
bkg0.phis[0].setVal(2.72)
bkg0.phis[0].setError(0.02)
#compound model
model = Models.Fit1D(signal=sig_bw, background=bkg0)
model.draw()