def test_fit_bins():
import ROOT
x = ROOT.RooRealVar('mbc', '', 0, 0, 1)
pdf = Gauss(x)
df = get_test_df(100)
pdf.fit(df, nbins=5)
assert isinstance(pdf.roo_pdf, ROOT.RooAbsPdf)
def test_Convolution():
import ROOT
df = get_test_df()
assert isinstance(df, pd.DataFrame)
bkg = Chebychev(('mbc', 0, 1))
sig = Gauss(('mbc', 0, 1))
pdf = Convolution(bkg, sig)
assert isinstance(pdf, Convolution)
assert isinstance(pdf.roo_pdf, ROOT.RooAbsPdf)
def test_ProdPdf():
import ROOT
df = get_test_df()
assert isinstance(df, pd.DataFrame)
bkg = Chebychev(('mbc', 0, 1))
sig = Gauss(('mbc', 0, 1))
pdf = sig * bkg
assert isinstance(pdf, ProdPdf)
assert isinstance(pdf.roo_pdf, ROOT.RooAbsPdf)
def test_AddPdf_fit():
import ROOT
df = get_test_df()
assert isinstance(df, pd.DataFrame)
bkg = Chebychev(('mbc', 0, 1))
sig = Gauss(('mbc', 0, 1))
pdf = sig + bkg
pdf.fit(df)
#pdf.plot('test2.pdf')
assert isinstance(pdf, AddPdf)
assert isinstance(pdf.roo_pdf, ROOT.RooAbsPdf)
# -*- coding: utf-8 -*-
""" Simple fit to a gaussian distribution (binned)
In this example a fit is performed to a simple gaussion distribution.
Observables can be initialised by a list with the column name / variable name as first argument, followed
by the range and/or with the initial value and range:
x = ('x', -3, 3)
x = ('mass', -3, 0.02, 3)
Parameters are initialised with a tuple: sigma=(0,1) or again including a starting parameter: sigma=(0.01, 0, 1)
The order here is not important.
All parameters and observables can also be initialised by a ROOT.RooRealVar.
"""
from pyroofit.models import Gauss
import numpy as np
data = np.random.normal(0, 1, 1000)
pdf = Gauss(('x', -3, 3), mean=(-1, 0, 1))
pdf.fit(data, nbins=10)
pdf.plot('simple_fit_binning.pdf',)
pdf.get()
import ROOT
df_mixed = {
'mass':
np.append(
np.random.random_sample(1000) * 7 - 3.5,
np.random.normal(0, 0.5, 1000))
}
df_mixed = pd.DataFrame(df_mixed)
df_bkg = {'mass': np.random.random_sample(2000) * 7 - 3.5}
df_bkg = pd.DataFrame(df_bkg)
x = ROOT.RooRealVar('mass', 'M', 0, -3, 3, 'GeV')
pdf_sig = Gauss(x, mean=(-1, 1))
pdf_bkg = Chebychev(x, n=1)
pdf = pdf_sig + pdf_bkg
sf = SimFit(pdf, pdf_bkg)
sf.use_extended = True # bug
sf.use_minos = True
sf.fit([(pdf, df_mixed), (pdf_bkg, df_bkg)])
#pdf_bkg.plot('simultaneous_fit_bkg.pdf', df_bkg)
pdf.plot('simultaneous_fit.pdf',
df_mixed,
nbins=20,
extra_info=[["Legend"], ["More Legend"],
['#mu', *pdf_sig.get('mean')],
"""
from pyroofit.models import Gauss, Chebychev
from pyroofit.composites import AddPdf
import numpy as np
import pandas as pd
import ROOT
dist_signal = np.append(np.random.normal(750, 1.5, 500),
np.random.normal(750, .5, 1000))
dist_background = np.random.random_sample(1000) * 10 + 745,
df = pd.DataFrame({'mass': np.append(dist_signal, dist_background)})
x = ROOT.RooRealVar('mass', 'M', 750, 745, 755,
'GeV') # or x = ('mass', 745, 755)
pdf_sig1 = Gauss(x, mean=(745, 755), sigma=(0.1, 1, 2), name="Gauss1")
# We want to furthermore constrain the mean to be the same
pdf_sig2 = Gauss(x,
mean=pdf_sig1.parameters.mean,
sigma=(0.1, 1, 2),
name="Gauss2")
pdf_bkg = Chebychev(x, n=1, name="bkg")
pdf = AddPdf([pdf_sig1, pdf_sig2, pdf_bkg], name="model")
pdf.fit(df)
pdf.plot('multiple_pdfs.pdf', legend=True)
pdf.get()
# -*- coding: utf-8 -*-
""" Simple product pdf
"""
from pyroofit.models import Gauss
import numpy as np
import pandas as pd
# Creating some pseudo data
mean = (1, 2)
cov = [[1, -.1], [-.1, 1]]
x = np.random.multivariate_normal(mean, cov, 300)
df = pd.DataFrame({'x': x[:, 0], 'y': x[:, 1]})
# Initialize the PDF
pdf_x = Gauss(['x', -1, 5], mean=[0, 3], sigma=[0.1, 2], name='gx')
pdf_y = Gauss(['y', -1, 5],
mean=[0, 3],
sigma=pdf_x.parameters.sigma,
name='gy')
pdf = pdf_x * pdf_y
# Fit
pdf.fit(df)
pdf.get()
The order here is not important.
All parameters and observables can also be initialised by a ROOT.RooRealVar.
"""
from pyroofit.models import Gauss, Chebychev
import numpy as np
import pandas as pd
import ROOT
df = {
'mass':
np.append(
np.random.random_sample(1000) * 10 + 745,
np.random.normal(750, 1, 1000))
}
df = pd.DataFrame(df)
x = ROOT.RooRealVar('mass', 'M', 750, 745, 755,
'GeV') # or x = ('mass', 745, 755)
pdf_sig = Gauss(x, mean=(745, 755), sigma=(0.1, 1, 2), title="Signal")
pdf_bkg = Chebychev(x, n=1, title="Background")
pdf = pdf_sig + pdf_bkg
pdf.fit(df)
pdf.plot('signal_and_background.pdf', legend=True)
pdf.get()
def test_PDF_Gauss():
import ROOT
x = ROOT.RooRealVar('mbc', '', 0, 0, 1)
pdf = Gauss(x)
assert isinstance(pdf.roo_pdf, ROOT.RooAbsPdf)