def drawChi2Dist(title="", chi2DistDict=[], legend=[]):
#differnt injection
ext = ".pdf"
nList = []
binsList = []
pathcesList = []
color = ['g', 'b', 'g', 'r', 'c', 'm', 'y', 'k', 'w']
i = 0
with Canvas("Testchi2.pdf") as can:
for tag in legend:
pseudoTag = tag + "Pseudo"
print("chi2DistDict[tag]: ", chi2DistDict[pseudoTag])
print("color: ", color[i])
#n, bins, patches, = plt.hist(chi2DistDict[pseudoTag], 50, normed=1, facecolor=color[i], alpha=0.75, label=legend[i])
can.ax.hist(chi2DistDict[pseudoTag],
bins=20,
facecolor=color[i],
alpha=0.75,
label=legend[i])
can.ax.set_xlim(0, 3)
#bins = np.histogram(chi2DistDict[pseudotag], edges)
#nList.append(n)
#binsList.append(bins)
#patchesList.append(patches)
i = i + 1
can.ax.legend(framealpha=0)
can.save("testchi2.pdf")
def drawSignalGaussianFit(signalBkgDataSet,
asignalDataSet,
doLog=False,
title=""):
#drawing the signal
ext = ".pdf"
title = [s + "SignalGaussianFit" for s in title]
with Canvas(f'%s{ext}' % title, "Gaussian Fit Sig", "", "", 2) as can:
can.ax.errorbar(signalBkgDataSet.xData,
asignalDataSet.ySigData,
yerr=signalBkgDataSet.yerrData,
fmt='.g',
label="signal MC injected") # drawing the points
can.ax.set_ylim(0.1, 10000.0)
if doLog:
can.ax.set_yscale('log')
can.ax.plot(signalBkgDataSet.xData,
asignalDataSet.yGaussianFit,
'-r',
label="Signal Gaussian Fit")
can.ax.legend(framealpha=0)
##
#can.ratio.stem(signalBkgDataSet.xData, asignalDataSet.gaussianFitSignificance, markerfmt='.', basefmt=' ')
can.ratio[0].axhline(0, linewidth=1, alpha=0.5)
#can.ax.plot(xSB, ymuGP_KernBkg_SB, '-g', label="GP bkgnd kernel") #drawing
can.save(title)
def drawSignalSubtractionFit(signalBkgDataSet,
ySignalData,
ySignalFit,
ySignalFitSignificance,
doLog=False,
title=""):
title = title + "signalGPfit"
#drawing the signal
ext = ".pdf"
with Canvas(f'%s{ext}' % title, "GPSig-GPBkgFit Sig", "", "", 2) as can:
can.ax.errorbar(signalBkgDataSet.xData,
ySignalData,
yerr=signalBkgDataSet.yerrData,
fmt='.g',
label="signal MC injected") # drawing the points
can.ax.set_ylim(0.1, 10000.0)
if doLog:
can.ax.set_yscale('log')
can.ax.plot(signalBkgDataSet.xData,
ySignalFit,
'-r',
label="GP Sig + bkg Kernel fit - GP bkgnd kernel fit")
can.ax.legend(framealpha=0)
##
can.ratio[0].stem(signalBkgDataSet.xData,
ySignalFitSignificance,
markerfmt='.',
basefmt=' ')
can.ratio[0].axhline(0, linewidth=1, alpha=0.5)
#can.ax.plot(xSB, ymuGP_KernBkg_SB, '-g', label="GP bkgnd kernel") #drawing
can.save(title)
def drawAllSignalFit(signalBkgDataSet,
asignalDataSet,
doLog=False,
saveTxt=False,
title=""):
#drawing the signal
ext = ".pdf"
title = title + "allSignalFit"
with Canvas(f'%s{ext}' % title, "All Signal Fits", "", "", 2) as can:
can.ax.errorbar(signalBkgDataSet.xData,
asignalDataSet.ySigData,
yerr=signalBkgDataSet.yerrData,
fmt='.k',
label="signal MC injected") # drawing the points
can.ax.set_ylim(0.1, 15000.0)
if doLog:
can.ax.set_yscale('log')
can.ax.plot(signalBkgDataSet.xData,
asignalDataSet.yGaussianFit,
'-r',
label="Signal point Gaussian Fit(injected signal)")
can.ax.plot(signalBkgDataSet.xData,
asignalDataSet.yGPSubtractionFit,
'-g',
label="Signal GP Fit subtraction")
can.ax.plot(signalBkgDataSet.xData,
signalBkgDataSet.MAP_sig,
'-b',
label="Signal GP reconstructed Fit")
#can.axplot(signalBkgDataset.xData, asignalDataSet.yReconsturcted)
can.ax.legend(framealpha=0)
#can.ratio.stem(signalBkgDataSet.xData, asignalDataSet.gaussianFitSignificance, markerfmt='.', basefmt=' ')
can.ratio[0].axhline(0, linewidth=1, alpha=0.5)
#can.ax.plot(xSB, ymuGP_KernBkg_SB, '-g', label="GP bkgnd kernel") #drawing
can.save(title)
def drawAllSignalFitYvonne(signalBkgDataSet,
asignalDataSet,
doLog=False,
saveTxt=False,
title="",
significanceSig=None,
sig=None):
#drawing the signal
ext = ".pdf"
title = title + "allSignalFit"
with Canvas(f'%s{ext}' % title, "All Signal Fits", "", "", 2) as can:
can.ax.errorbar(signalBkgDataSet.xData,
asignalDataSet.ySigData,
yerr=signalBkgDataSet.yerrData,
fmt='.k',
label="signal MC injected") # drawing the points
can.ax.set_ylim(0.1, 15000.0)
if doLog:
can.ax.set_yscale('log')
can.ax.plot(signalBkgDataSet.xData,
asignalDataSet.yGaussianFit,
'-r',
label="Signal point Gaussian Fit(injected signal)")
can.ax.plot(signalBkgDataSet.xData,
asignalDataSet.yGPSubtractionFit,
'-g',
label="Signal GP Fit subtraction")
#can.ax.plot(signalBkgDataSet.xData, asignalDataSet.sig['Gaussian'],'-m', label="Signal GP Kernel reconstructed Gaussian Fit")
print(asignalDataSet.sig['custom'])
can.ax.plot(signalBkgDataSet.xData,
asignalDataSet.sig['custom'],
'-b',
label="Signal GP Kernel reconstructed signal template Fit")
#can.ax.plot(signalBkgDataSet.xData, asignalDataSet.sig['customTest'],'-k', label="Signal GP Kernel reconstructed signal template Fit default")
#accidentally broke gaussian signal
#print both signal
# if sig:
# for i in range(len(sig)):
# print("i", i)
# can.ratio[i].stem(signalBkgDataSet.xData, significanceSig[i], markerfmt='.', basefmt=' ')
# can.ratio[i].set_ylabel(sig[i])
# print("legend: ", sig[i])
# can.ratio[i].axhline(0, linewidth=1, alpha=0.5)
# can.ax.legend(framealpha=0)
can.ratio[1].stem(signalBkgDataSet.xData,
significanceSig[1],
markerfmt='.',
basefmt=' ')
can.ratio[1].set_ylabel(sig[1])
print("legend: ", sig[1])
can.ratio[1].axhline(0, linewidth=1, alpha=0.5)
can.ax.legend(framealpha=0)
#can.ratio.stem(signalBkgDataSet.xData, asignalDataSet.gaussianFitSignificance, markerfmt='.', basefmt=' ')
#can.ratio.axhline(0, linewidth=1, alpha=0.5)
#can.ax.plot(xSB, ymuGP_KernBkg_SB, '-g', label="GP bkgnd kernel") #drawing
can.save("results/" + title)
def run():
args = parse_args()
ext = args.output_file_extension
from pygp.canvas import Canvas
with File(args.input_file, 'r') as h5file:
x, y, xerr, yerr = get_xy_pts(
h5file['dijetgamma_g85_2j65']['Zprime_mjj_var'], FIT_RANGE)
if args.load_pars:
with open(args.load_pars, 'r') as pars_file:
best_fit = json.load(pars_file)
else:
lnProb = logLike_minuit(x, y, yerr)
_, best_fit = fit_gp_minuit(args.n_fits, lnProb)
if args.save_pars:
with open(args.save_pars, 'w') as pars_file:
json.dump(best_fit, pars_file, indent=2)
fit_pars = [best_fit[x] for x in FIT_PARS]
with Canvas(f'points{ext}') as can:
can.ax.errorbar(x, y, yerr=yerr, fmt='.')
can.ax.set_yscale('log')
kargs = {x: y for x, y in best_fit.items() if x not in FIT_PARS}
kernel_new = get_kernel(**kargs)
gp_new = george.GP(kernel_new, mean=Mean(fit_pars), fit_mean=True)
ext = args.output_file_extension
gp_new.compute(x, yerr)
plot_gp(x, y, xerr, yerr, gp_new, name=f'spectrum{ext}')
if not args.signal_file:
return
with File(args.signal_file, 'r') as h5file:
x_sig, y_sig, xerr_sig, yerr_sig = get_xy_pts(
h5file['dijetgamma_g85_2j65']['Zprime_mjj_var'], FIT_RANGE)
y_sig *= args.signal_multiplier
yerr_sig *= args.signal_multiplier
tot_error = (yerr**2 + yerr_sig**2)**(1 / 2)
plot_gp(x_sig,
y + y_sig,
xerr_sig,
tot_error,
gp_new,
name=f'with-signal{ext}',
y_bg=y,
yerr_bg=yerr,
signal_multiplier=args.signal_multiplier)
def plot_gp(x,
y,
xerr,
yerr,
gp_new,
name,
y_bg=None,
yerr_bg=None,
signal_multiplier=1.0):
from pygp.canvas import Canvas
t = np.linspace(np.min(x), np.max(x), 500)
mu, cov = gp_new.predict(y, t)
mu_x, cov_x = gp_new.predict(y, x)
signif = (y - mu_x) / np.sqrt(np.diag(cov_x) + yerr**2)
fit_mean_smooth = gp_new.mean.get_value(t)
std = np.sqrt(np.diag(cov))
with Canvas(name) as can:
sm = signal_multiplier
sig_label = f'sig * {sm:.3g} + bg' if sm != 1.0 else 'sig + bg'
can.ax.errorbar(x, y, yerr=yerr, fmt='.', label=sig_label)
if y_bg is not None:
can.ax.errorbar(x, y_bg, yerr=yerr_bg, fmt='.', label='bg')
can.ax.set_yscale('log')
can.ax.plot(t, mu, '-r')
# can.ax.plot(t, fit_mean_smooth, '-k', label='fit function')
can.ax.fill_between(t,
mu - std,
mu + std,
facecolor=(0, 1, 0, 0.5),
zorder=5,
label=r'GP error = $1\sigma$')
can.ax.legend(framealpha=0)
can.ax.set_ylabel('events')
can.ratio.stem(x, signif, markerfmt='.', basefmt=' ')
can.ratio.axhline(0, linewidth=1, alpha=0.5)
can.ratio.set_xlabel(r'$m_{jj}$ [GeV]', ha='right', x=0.98)
can.ratio.set_ylabel('significance')
def drawFit2(xData=None,
yerr=None,
yData=None,
yFit=None,
yFit2=None,
sig=None,
signiLegend=None,
title=None,
saveTxt=False,
saveTxtDir=None):
# draw the data set using diffrent fits
ext = ".pdf"
with Canvas(f'%s{ext}' % title, "UA2", "", 2) as can:
can.ax.errorbar(xData, yData, yerr, fmt='.g',
label="datapoints") # drawing the points
can.ax.set_yscale('log')
can.ax.plot(xData, yFit, '-g', label="bkgndKernelFit") #drawing
can.ax.plot(xData, yFit2, '-r', label="GP background+signal") #drawing
if saveTxtDir == None:
saveTxtDir = ""
with open(saveTxtDir + "dataPoints.txt", "w") as f0:
writer = csv.writer(f0, delimiter="\t")
writer.writerows(zip(xData, yData))
with open(saveTxtDir + "bkgndReconstructed.txt", "w") as f0:
writer = csv.writer(f0, delimiter="\t")
writer.writerows(zip(xData, yFit))
with open(saveTxtDir + "bkgndSigReconstructed.txt", "w") as f0:
writer = csv.writer(f0, delimiter="\t")
writer.writerows(zip(xData, yFit2))
can.ax.legend(framealpha=0)
if sig:
for i in range(len(sig)):
print("i", i)
can.ratio[i].stem(xData, sig[i], markerfmt='.', basefmt=' ')
if signiLegend != None:
can.ratio[i].set_ylabel(signiLegend[i])
can.ratio[i].axhline(0, linewidth=1, alpha=0.5)
can.save("results/" + title)
def drawFit(xData=None,
yerr=None,
yData=None,
yFit=None,
sig=None,
title=None,
saveTxt=False,
saveTxtDir=None):
# draw the data set using diffrent fits
ext = ".pdf"
with Canvas(f'%s{ext}' % title, "UA2", "", 2) as can:
can.ax.errorbar(xData, yData, yerr, fmt='.g',
label="datapoints") # drawing the points
can.ax.set_yscale('log')
can.ax.plot(xData, yFit, '-g', label="UA2Fit") #drawing
#ratio plot:
print("x: ", xData)
print("sig: ", sig)
if sig:
can.ratio[0].stem(xData, sig, markerfmt='.', basefmt=' ')
can.ratio[0].set_ylabel("significance")
can.ratio[0].axhline(0, linewidth=1, alpha=0.5)
can.save(title)
def drawFit3(xData=None,
yerr=None,
yData=None,
yFit=None,
yFit2=None,
yFit3=None,
legend=[],
sigList=None,
title=None,
saveTxt=False,
saveTxtDir=None):
# draw the data set using diffrent fits
ext = ".pdf"
with Canvas(f'%s{ext}' % title, "UA2", "", 2) as can:
can.ax.errorbar(xData, yData, yerr, fmt='.g',
label="datapoints") # drawing the points
can.ax.set_yscale('log')
if legend == []:
legend = [
"bkgKernelFit", "GP bakcground +signal kernel fit",
"GP backgtound+signal fit -custom signal template reconsturct"
]
can.ax.plot(xData, yFit, '-g', label=legend[0]) #drawing
can.ax.plot(xData, yFit2, '-r', label=legend[1]) #drawing
can.ax.plot(xData, yFit3, '-b', label=legend[2]) #drawing
#ratio plot:
#print("x: ", xData)
#print("sig: ", sig)
can.ax.legend(framealpha=0)
i = 0
for sig in sigList:
can.ratio[i].stem(xData, sig[i], markerfmt='.', basefmt=' ')
can.ratio[i].set_ylabel("significance")
can.ratio[i].axhline(0, linewidth=1, alpha=0.5)
can.save(title)
def drawSignalReconstructed(signalBkgDataSet,
asignalDataSet,
doLog=False,
title=""):
#drawing the signal with meghan's method
ext = args.output_file_extension
title = title + "signalonlyMeghan"
with Canvas(f'%s{ext}' % title, "GPSig-GPSBFit Sig", "", "", 2) as can:
can.ax.errorbar(xSB,
ySig,
yerr=yerrSB,
fmt='.g',
label="signal MC injected") # drawing the points
can.ax.set_ylim(0.1, 6000000)
can.ax.set_yscale('log')
can.ax.plot(xSB, MAP_bkg, '-r', label="GP Sig Kernel Only fit")
can.ax.legend(framealpha=0)
can.ratio[0].stem(xSB,
GPSigOnlySignificanceMeg,
markerfmt='.',
basefmt=' ')
can.ratio[0].axhline(0, linewidth=1, alpha=0.5)
#can.ax.plot(xSB, ymuGP_KernBkg_SB, '-g', label="GP bkgnd kernel") #drawing
can.save(title)
def run_vs_3paramFit():
args = parse_args()
ext = args.output_file_extension
from pygp.canvas import Canvas
xMinFit = 300
# Getting data points
xRaw, yRaw, xerrRaw, yerrRaw = getDataPoints(args.input_file,
'dijetgamma_g85_2j65',
'Zprime_mjj_var')
xSig, ySig, xerrSig, yerrSig = getDataPoints(args.signal_file,
'dijetgamma_g85_2j65',
'Zprime_mjj_var')
#Data processing, cutting out the not desired range
x, y, xerr, yerr = dataCut(xMinFit, 1500, 0, xRaw, yRaw, xerrRaw, yerrRaw)
xFit, yFit, xerrFit, yerrFit = dataCut(xMinFit, 1500, 0, xRaw, yRaw,
xerrRaw, yerrRaw)
# make an evently spaced x
t = np.linspace(np.min(x), np.max(x), 500)
#calculating the log-likihood and minimizing for the gp
lnProb = logLike_minuit(x, y, xerr)
#
# idecay = np.random.random() * 0.64
# ilength = np.random.random() * 5e5
# ipower = np.random.random() * 1.0
# isub = np.random.random() * 1.0
# 'amp': 5701461179.0,
# 'p0': 0.23,
# 'p1': 0.46,
# 'p2': 0.89
#
#dan crap initial guess->returns infiinity
print("prob:", lnProb(5701461179.0, 0.64, 5e5, 1.0, 1.0, 0.23, 0.46, 0.89))
print(
"prob:",
lnProb(7894738685.23, 91.507809530688036, 152892.47486888882,
0.77936430405302681, 393.57106174440003, 0.46722747265429021,
0.92514297129196166, 1.7803224928740065))
#def __call__(self, Amp, decay, length, power, sub, p0, p1, p2):
min_likelihood, best_fit = fit_gp_minuit(10, lnProb)
fit_pars = [best_fit[x] for x in FIT3_PARS]
#making the GP
kargs = {x: y for x, y in best_fit.items() if x not in FIT3_PARS}
kernel_new = get_kernel(**kargs)
print(kernel_new.get_parameter_names())
#making the kernel
gp_new = george.GP(kernel_new, mean=Mean(fit_pars), fit_mean=True)
gp_new.compute(x, yerr)
mu, cov = gp_new.predict(y, t)
mu_x, cov_x = gp_new.predict(y, x)
# calculating the fit function value
#GP compute minimizes the log likelihood of the best_fit function
best = [best_fit[x] for x in FIT3_PARS]
print("best param meghan GP minmization:", best)
meanFromGPFit = Mean(best)
fit_meanM = meanFromGPFit.get_value(
x, xerr) #so this is currently shit. can't get the xErr thing working
print("fit_meanM:", fit_meanM)
#fit_mean_smooth = gp_new.mean.get_value(t)
#----3 param fit function in a different way
lnProb = logLike_3ff(xFit, yFit, xerrFit)
minimumLLH, best_fit_params = fit_3ff(100, lnProb)
fit_mean = model_3param(xFit, best_fit_params, xerrFit)
##----4 param fit function
#lnProb = logLike_4ff(xFit,yFit,xerrFit)
#minimumLLH, best_fit_params = fit_4ff(100, lnProb)
#fit_mean = model_4param(xFit, best_fit_params, xerrFit)
#calculating significance
signif = significance(mu_x, y, cov_x, yerr)
initialCutPos = np.argmax(x > xMinFit)
#sigFit = (fit_mean - y[initialCutPos:]) / np.sqrt(np.diag(cov_x[initialCutPos:]) + yerr[initialCutPos:]**2)
sigFit = significance(fit_mean, y[initialCutPos:], cov_x[initialCutPos:],
yerr[initialCutPos:])
#sigit = (mu_x)
std = np.sqrt(np.diag(cov))
ext = args.output_file_extension
title = "test"
with Canvas(f'%s{ext}' % title) as can:
can.ax.errorbar(x, y, yerr=yerr, fmt='.')
can.ax.set_yscale('log')
# can.ax.set_ylim(1, can.ax.get_ylim()[1])
can.ax.plot(t, mu, '-r')
can.ax.plot(xFit, fit_mean, '-b')
#this only works with the xErr part of Mean commented out
#can.ax.plot(x, fit_meanM, '.g')
# can.ax.plot(t, fit_mean_smooth, '--b')
#can.ax.fill_between(t, mu - std, mu + std,
#facecolor=(0, 1, 0, 0.5),
#zorder=5, label='err = 1')
#can.ratio.stem(x, signif, markerfmt='.', basefmt=' ')
can.ratio.stem(xFit, sigFit, markerfmt='.', basefmt=' ')
can.ratio.axhline(0, linewidth=1, alpha=0.5)
can.ratio2.stem(x, signif, markerfmt='.', basefmt=' ')
can.save(title)
def run_bkgndVswithSig_GP():
args = parse_args()
ext = args.output_file_extension
from pygp.canvas import Canvas
mass = 750
# Getting data points
xRaw, yRaw, xerrRaw, yerrRaw = getDataPoints(args.input_file,
'dijetgamma_g85_2j65',
'Zprime_mjj_var')
xSig, ySig, xerrSig, yerrSig = getDataPoints(args.signal_file,
'dijetgamma_g85_2j65',
'Zprime_mjj_var')
#Data processing, cutting out the not desired range
xBkg, yBkg, xerrBkg, yerrBkg = dataCut(300, 1500, 0, xRaw, yRaw, xerrRaw,
yerrRaw)
xSig, ySig, xerrSig, yerrSig = dataCut(300, 1500, 0, xSig, ySig, xerrSig,
yerrSig)
# make an evently spaced x
t = np.linspace(np.min(xBkg), np.max(xBkg), 500)
#Drawing the raw data of just the bkgnd and Sig + bkgnd
with Canvas(f'compareRawBkgndAndBkgndWithSig{ext}') as can:
can.ax.errorbar(xBkg, yBkg, yerr=yerrBkg, fmt='.g')
can.ax.errorbar(xSig, ySig, yerr=yerrSig, fmt='.r')
can.ax.set_yscale('log')
#bkgnd calculating the log-likihood and minimizing for the gp
lnProbBkg = logLike_minuit(xBkg, yBkg, xerrBkg)
min_likelihood, best_fit = fit_gp_minuit(100, lnProbBkg)
print(best_fit)
fit_pars = [best_fit[x] for x in FIT3_PARS]
#making the GP
kargs = {
xBkg: yBkg
for xBkg, yBkg in best_fit.items() if xBkg not in FIT3_PARS
}
kernel_Bkg = get_kernel(**kargs)
print(kernel_Bkg.get_parameter_names())
#making the kernel
gp_Bkg = george.GP(kernel_Bkg, mean=Mean(fit_pars), fit_mean=True)
gp_Bkg.compute(xBkg, yerrBkg)
muBkg, covBkg = gp_Bkg.predict(yBkg, t)
mu_xBkg, cov_xBkg = gp_Bkg.predict(yBkg, xBkg)
#signal calulating the log_likihood and minizming for the GP
# calculating the fit function value
lnProbSig = logLike_gp_fitgpsig(xSig, ySig, xerrSig)
min_likelihoodSig, best_fitSig = fit_gp_fitgpsig_minuit(lnProbSig)
fit_pars = [best_fit[xSig] for xSig in FIT3_PARS]
Args = kargs + best_fitSig
kernel_Sig = get_kernelXtreme(**Arg)
print(kernel_Sig.get_parameter_names())
#making the kernel
gp_sig = george.GP(kernel_Sig, mean=Mean(fit_pars), fit_mean=True)
gp_sig.compute(xSig, yerrSig)
muSig, covSig = gp_sig.predict(ySig, t)
mu_xSig, cov_xSig = gp_Sig.predict(ySig, xSig)
#GP compute minimizes the log likelihood of the best_fit function
best = [best_fit[x] for x in FIT3_PARS]
#calculating significance
signBkg = significance(mu_xBkg, yBkg, cov_xBkg, yerrBkg)
#sigFit = (fit_mean - y[initialCutPos:]) / np.sqrt(np.diag(cov_x[initialCutPos:]) + yerr[initialCutPos:]**2)
signSig = significance(mu_xSig, ySig, cov_xSig, yerrSig)
#sigit = (mu_x)
std = np.sqrt(np.diag(cov_xBkg))
ext = args.output_file_extension
with Canvas(f'compareGPvsSignal{ext}') as can:
can.ax.errorbar(xBkg, yBkg, yerr=yerrBkg, fmt='.g')
can.ax.set_yscale('log')
# can.ax.set_ylim(1, can.ax.get_ylim()[1])
can.ax.plot(xBkg, muSig, '-g')
can.ax.plot(xSig, muSig, '-r')
#this only works with the xErr part of Mean commented out
#can.ax.plot(x, fit_meanM, '.g')
# can.ax.plot(t, fit_mean_smooth, '--b')
#can.ax.fill_between(t, mubkg - std, mu + std,
#facecolor=(0, 1, 0, 0.5),
#zorder=5, label='err = 1')
can.ratio.stem(xBkg, signBkg, markerfmt='.', basefmt=' ')
can.ratio2.stem(xSig, signSig, markerfmt='.', basefmt=' ')
can.ratio.axhline(0, linewidth=1, alpha=0.5)
def run_bkgnd():
xMin = 300
xMax = 1500
xMinFit = 300
xMaxFit = 1500
args = parse_args()
ext = args.output_file_extension
from pygp.canvas import Canvas
# Getting data points
xRaw, yRaw, xerrRaw, yerrRaw = getDataPoints(args.input_file,
'dijetgamma_g85_2j65',
'Zprime_mjj_var')
#Data processing, cutting out the not desired range
xBkg, yBkg, xerrBkg, yerrBkg = dataCut(
xMin, xMax, 0, xRaw, yRaw, xerrRaw,
yerrRaw) # for GP bkgnd kernel and signal kernel
xBkgFit, yBkgFit, xerrBkgFit, yerrBkgFit = dataCut(
xMinFit, xMaxFit, 0, xRaw, yRaw, xerrRaw, yerrRaw) # for fit function
# make an evently spaced x
t = np.linspace(np.min(xBkg), np.max(xBkg), 500)
#Drawing the raw data of just the bkgnd and Sig + bkgnd
with Canvas(
f'RawBkgnd') as can: # just to check and see everything is okay
can.ax.errorbar(xBkg, yBkg, yerr=yerrBkg, fmt='.g')
can.ax.set_yscale('log')
#GP bkgnd calculating the log-likihood and minimizing for the gp
lnProbBkg = logLike_minuit(xBkg, yBkg, xerrBkg)
min_likelihood, best_fit = fit_gp_minuit(10, lnProbBkg)
print(best_fit)
fit_pars = [best_fit[x] for x in FIT3_PARS]
#making the GP
kargs = {
xBkg: yBkg
for xBkg, yBkg in best_fit.items() if xBkg not in FIT3_PARS
}
kernel_Bkg = get_kernel(**kargs)
print(kernel_Bkg.get_parameter_names())
#making the kernel
gp_Bkg = george.GP(kernel_Bkg, mean=Mean(fit_pars), fit_mean=True)
gp_Bkg.compute(xBkg, yerrBkg)
#muBkg, covBkg = gp_Bkg.predict(yBkg, t)
mu_xBkg, cov_xBkg = gp_Bkg.predict(yBkg, xBkg)
#finding the fit y values
fit_mean = y_bestFit3Params(xBkgFit, yBkgFit, xerrBkgFit, 10)
ext = args.output_file_extension
title = "test"
with Canvas(f'%s{ext}' % title, "Fit Function", "GP bkgnd kernel",
"GP signal+bkgnd kernel", 3) as can:
can.ax.errorbar(xBkg, yBkg, yerr=yerrBkg,
fmt='.g') # drawing the points
can.ax.set_yscale('log')
can.ax.plot(xBkg, mu_xBkg, '-g') #drawing
can.ax.plot(xBkgFit, fit_mean, '-r')
#can.ax.xticks(xBkg)
#this only works with the xErr part of Mean commented out
#can.ax.plot(x, fit_meanM, '.g')
# can.ax.plot(t, fit_mean_smooth, '--b')
#can.ax.fill_between(t, mubkg - std, mu + std,
#facecolor=(0, 1, 0, 0.5),
#zorder=5, label='err = 1')
#can.ratio.stem(xBkg, signBkg, markerfmt='.', basefmt=' ')
#can.ratio2.stem(xSig, signSig, markerfmt='.', basefmt=' ')
can.ratio.axhline(0, linewidth=1, alpha=0.5)
can.save(title)
def run_vs_4paramFit():
args = parse_args()
ext = args.output_file_extension
from pygp.canvas import Canvas
# Getting data points
xRaw, yRaw, xerrRaw, yerrRaw = getDataPoints(args.input_file,
'dijetgamma_g85_2j65',
'Zprime_mjj_var')
xSig, ySig, xerrSig, yerrSig = getDataPoints(args.signal_file,
'dijetgamma_g85_2j65',
'Zprime_mjj_var')
#Data processing, cutting out the not desired range
x, y, xerr, yerr = dataCut(0, 1500, 0, xRaw, yRaw, xerrRaw, yerrRaw)
xMinFit = 300
xFit, yFit, xerrFit, yerrFit = dataCut(xMinFit, 1500, 0, xRaw, yRaw,
xerrRaw, yerrRaw)
# make an evently spaced x
t = np.linspace(np.min(x), np.max(x), 500)
#calculating the log-likihood and minimizing for the gp
lnProb = logLike_minuit(x, y, xerr)
min_likelihood, best_fit = fit_gp_minuit(20, lnProb)
fit_pars = [best_fit[x] for x in FIT3_PARS]
#making the GP
kargs = {x: y for x, y in best_fit.items() if x not in FIT3_PARS}
kernel_new = get_kernel(**kargs)
print(kernel_new.get_parameter_names())
#making the kernel
gp_new = george.GP(kernel_new, mean=Mean(fit_pars), fit_mean=True)
gp_new.compute(x, yerr)
mu, cov = gp_new.predict(y, t)
mu_x, cov_x = gp_new.predict(y, x)
##----4 param fit function
lnProb = logLike_4ff(xFit, yFit, xerrFit)
minimumLLH, best_fit_params = fit_4ff(100, lnProb)
fit_mean = model_4param(xFit, best_fit_params, xerrFit)
#calculating significance
signif = significance(mu_x, y, cov_x, yerr)
initialCutPos = np.argmax(x > xMinFit)
sigFit = significance(fit_mean, y[initialCutPos:], cov_x[initialCutPos:],
yerr[initialCutPos:])
std = np.sqrt(np.diag(cov))
ext = args.output_file_extension
with Canvas(f'compareGPvs4Param{ext}', 'GP vs 4 param fit') as can:
can.ax.errorbar(x, y, yerr=yerr, fmt='.')
can.ax.set_yscale('log')
# can.ax.set_ylim(1, can.ax.get_ylim()[1])
can.ax.plot(t, mu, '-r')
can.ax.plot(xFit, fit_mean, '-g')
can.ax.fill_between(t,
mu - std,
mu + std,
facecolor=(0, 1, 0, 0.5),
zorder=5,
label='err = 1')
can.ratio.stem(x, signif, markerfmt='.', basefmt=' ')
can.ratio2.stem(xFit, sigFit, markerfmt='.', basefmt=' ')
can.ratio.axhline(0, linewidth=1, alpha=0.5)
def run_bkgnd():
xMin = 300
xMax = 1500
xMinFit = 300
xMaxFit = 1500
args = parse_args()
ext = args.output_file_extension
from pygp.canvas import Canvas
# Getting data points
xRaw, yRaw, xerrRaw, yerrRaw = getDataPoints(args.input_file,
'dijetgamma_g85_2j65',
'Zprime_mjj_var')
#data points from official fit function
xFitRaw, yFitRaw, xerrFitRaw, yerrFitRaw = getDataPoints(
args.fitFromOfficial, "", "basicBkgFrom4ParamFit")
print("xFit: ", xFitRaw)
#Data processing, cutting out the not desired range
xBkg, yBkg, xerrBkg, yerrBkg = dataCut(
xMin, xMax, 0, xRaw, yRaw, xerrRaw,
yerrRaw) # for GP bkgnd kernel and signal kernel
xBkgFit, yBkgFit, xerrBkgFit, yerrBkgFit = dataCut(
xMinFit, xMaxFit, 0, xRaw, yRaw, xerrRaw, yerrRaw) # for fit function
xFit, yFit, xerrFit, yerrFit = dataCut(xMinFit, xMaxFit, 0, xFitRaw,
yFitRaw, xerrFitRaw, yerrRaw)
# calculate the min log likelihood of the official fit function
logLikeOff = logLike_3ffOff(xBkg, yBkg, xFit, yFit, xerrBkg)
# make an evently spaced x
t = np.linspace(np.min(xBkg), np.max(xBkg), 500)
#Drawing the raw data of just the bkgnd and Sig + bkgnd
with Canvas(
f'RawBkgnd') as can: # just to check and see everything is okay
can.ax.errorbar(xBkg, yBkg, yerr=yerrBkg, fmt='.g')
can.ax.set_yscale('log')
#GP bkgnd: finidng the mean and covaraicne
mu_xBkg, cov_xBkg, bestFitBkg = y_bestFitGP(xBkg,
yBkg,
xerrBkg,
yerrBkg,
30,
kernelType="bkg")
# GP Signal: finind the mean of covariance
mu_xSig, cov_xSig, bestFitSig = y_bestFitGP(xBkg,
yBkg,
xerrBkg,
yerrBkg,
30,
kernelType="sig")
#finding the fit y values
fit_mean = y_bestFit3Params(xBkgFit, yBkgFit, xerrBkgFit, 1)
#Signal only fit
ySigFit = mu_xSig - mu_xBkg
print("ySigFit: ", ySigFit)
#finding signifiance
GPSignificance, chi2 = res_significance(yBkg, mu_xBkg)
fitSignificance, chi2fit = res_significance(yBkgFit, fit_mean)
GPSigSignificance, chi2SignalFit = res_significance(yBkg, mu_xSig)
FitSigFromOff, chi2FitOff = res_significance(yBkgFit, yFit)
#drawing the result
ext = args.output_file_extension
title = "test"
with Canvas(f'%s{ext}' % title, "Fit Function official ",
"GP bkgnd kernel", "GP signal+bkgnd kernel", 3) as can:
can.ax.errorbar(xBkg, yBkg, yerr=yerrBkg,
fmt='.g') # drawing the points
can.ax.set_yscale('log')
#can.ax.plot(xBkgFit, fit_mean, '-r', label="fit function")
can.ax.plot(xFit, yFit, '-r', label="fit function official")
can.ax.plot(xBkg, mu_xBkg, '-g', label="GP bkg kernel") #drawing
can.ax.plot(xBkg, mu_xSig, '-b', label="GP signal kernel")
can.ax.legend(framealpha=0)
can.ratio.stem(xFit, FitSigFromOff, markerfmt='.', basefmt=' ')
can.ratio.set_ylabel("significance")
can.ratio2.stem(xBkg, GPSignificance, markerfmt='.', basefmt=' ')
can.ratio2.set_ylabel("significance")
can.ratio3.set_ylabel("significance")
can.ratio3.stem(xBkg, GPSigSignificance, markerfmt='.', basefmt=' ')
can.ratio.axhline(0, linewidth=1, alpha=0.5)
can.ratio2.axhline(0, linewidth=1, alpha=0.5)
can.ratio3.axhline(0, linewidth=1, alpha=0.5)
can.save(title)
fitChi2List, GPChi2List = psuedoTest(100, yBkgFit, yerrBkgFit, fit_mean,
yBkg, yerrBkg, mu_xBkg)
print("fitChi2List", fitChi2List)
print("GPChi2List", GPChi2List)
#n, bins, patches = plt.hist(fitChi2List, 50, normed=1, facecolor='green', alpha=0.75)
#plt.plot(n, drawstyle="steps")
#plt.savefig("chi2.pdf")
value, edges = np.histogram(fitChi2List)
print("value %r, edges %r" % (value, edges))
binCenters = (edges[1:] + edges[:-1]) / 2
valueGP, edgesGP = np.histogram(GPChi2List)
print("value %r, edges %r" % (valueGP, edgesGP))
binCentersGP = (edgesGP[1:] + edgesGP[:-1]) / 2
#plt.step(binCenters,value)
#plt.savefig("chi2.pdf")
n = binCenters.size - 1
nGP = binCentersGP.size - 1
title = "chi2"
with Canvas(f'%s{ext}' % title, "Fit Function", "GP bkgnd kernel", "",
2) as can:
can.ax.step(binCenters / n, value, label="Fit Function", where="mid")
can.ax.step(binCentersGP / nGP, valueGP, label="GP bkgnd", where="mid")
can.ax.legend(framealpha=0)
can.save(title)
print("min LL of off: ", logLikeOff)
#drawing the signal only
ext = args.output_file_extension
title = "signalonlybkg"
with Canvas(f'%s{ext}' % title, "GPSig-GPBkgFit Sig", "", "", 2) as can:
#can.ax.errorbar(xBkg, ySig, yerr=yerrBkg, fmt='.g', label="signal MC injected") # drawing the points
can.ax.set_ylim(-1000, 1000.0)
#can.ax.set_yscale('log')
can.ax.plot(xBkg,
ySigFit,
'-r',
label="GP Sig + bkg Kernel fit - GP bkgnd kernel fit")
can.ax.legend(framealpha=0)
##
#can.ratio.stem(xBkg, GPSigOnlySignificance, markerfmt='.', basefmt=' ')
can.ratio.axhline(0, linewidth=1, alpha=0.5)
#can.ax.plot(xBkg, mu_xBkg, '-g', label="GP bkgnd kernel") #drawing
can.save(title)
def drawFitDataSet(dataSet, title, saveTxt=False, saveTxtDir=None):
# draw the data set using diffrent fits
ext = ".pdf"
with Canvas(f'%s{ext}' % title, "GP bkgnd kernel",
"GP signal+bkgnd kernel", 2) as can:
can.ax.errorbar(dataSet.xData,
dataSet.yData,
yerr=dataSet.yerrData,
fmt='.g',
label="datapoints") # drawing the points
can.ax.set_yscale('log')
#can.ax.plot(dataSet.x_simpleFit, dataSet.yFit_simpleFit, '-r', label="fit function")
#can.ax.plot(dataSet.xOffFit, dataSet.yFit_officialFit, '-m', label="fit function official")
can.ax.plot(dataSet.xData,
dataSet.y_GPBkgKernelFit,
'-g',
label="GP bkgnd kernel") #drawing
#can.ax.plot(dataSet.xData, dataSet.y_GPSigPlusBkgKernelFit, '-b', label="GP signal kernel")
if saveTxtDir:
saveTxtDir = saveTxtDir + "/"
if saveTxt:
with open(saveTxtDir + "dataPoints.txt", "w") as f0:
writer = csv.writer(f0, delimiter="\t")
writer.writerows(zip(dataSet.xData, dataSet.yData))
print("check: ", dataSet.xData, " ", dataSet.yData)
with open(saveTxtDir + "simpleFit.txt", "w") as f1:
writer = csv.writer(f1, delimiter="\t")
writer.writerows(
zip(dataSet.x_simpleFit, dataSet.yFit_simpleFit))
with open(saveTxtDir + "fitFunctionOfficial.txt", "w") as f2:
writer = csv.writer(f2, delimiter="\t")
writer.writerows(zip(dataSet.xOffFit,
dataSet.yFit_officialFit))
with open(saveTxtDir + "GPBkgndKernel.txt", "w") as f3:
writer = csv.writer(f3, delimiter="\t")
writer.writerows(zip(dataSet.xData, dataSet.y_GPBkgKernelFit))
with open(saveTxtDir + "GPSigPlusBkgKernel.txt", "w") as f4:
writer = csv.writer(f4, delimiter="\t")
writer.writerows(
zip(dataSet.xData, dataSet.y_GPSigPlusBkgKernelFit))
can.ax.legend(framealpha=0)
#can.ratio.stem(dataSet.x_simpleFit, dataSet.significance_simpleFit, markerfmt='.', basefmt=' ')
#can.ratio.stem(xSBFit, fitSignificance, markerfmt='.', basefmt=' ')
#can.ratio.stem(xSBFit, testsig, markerfmt='.', basefmt=' ')
#can.ratio.set_ylabel("significance")
can.ratio[0].stem(dataSet.xData,
dataSet.significance_GPBkgKernelFit,
markerfmt='.',
basefmt=' ')
can.ratio[0].set_ylabel("significance")
can.ratio[1].set_ylabel("significance")
can.ratio[1].stem(dataSet.xData,
dataSet.significance_GPSigPlusBkgKernelFit,
markerfmt='.',
basefmt=' ')
can.ratio[0].axhline(0, linewidth=1, alpha=0.5)
can.save(title)
def run_vs_SearchPhase_UA2Fit():
args = parse_args()
ext = args.output_file_extension
from pygp.canvas import Canvas
# Getting data points
xRaw, yRaw, xerrRaw, yerrRaw = getDataPoints(args.input_file,
'dijetgamma_g85_2j65',
'Zprime_mjj_var')
#Data processing, cutting out the not desired range
x, y, xerr, yerr = dataCut(0, 1500, 0, xRaw, yRaw, xerrRaw, yerrRaw)
xMinFit = 303
xMaxFit = 1500
xFit, yFit, xerrFit, yerrFit = dataCut(xMinFit, xMaxFit, 0, xRaw, yRaw,
xerrRaw, yerrRaw)
# make an evently spaced x
t = np.linspace(np.min(x), np.max(x), 500)
#calculating the log-likihood and minimizing for the gp
lnProb = logLike_minuit(x, y, xerr)
min_likelihood, best_fit = fit_gp_minuit(20, lnProb)
fit_pars = [best_fit[x] for x in FIT3_PARS]
#making the GP
kargs = {x: y for x, y in best_fit.items() if x not in FIT3_PARS}
kernel_new = get_kernel(**kargs)
print(kernel_new.get_parameter_names())
#making the kernel
gp_new = george.GP(kernel_new, mean=Mean(fit_pars), fit_mean=True)
gp_new.compute(x, yerr)
mu, cov = gp_new.predict(y, t)
mu_x, cov_x = gp_new.predict(y, xFit)
# calculating the fit function value
#GP compute minimizes the log likelihood of the best_fit function
best = [best_fit[x] for x in FIT3_PARS]
print("best param meghan GP minmization:", best)
meanFromGPFit = Mean(best)
fit_meanM = meanFromGPFit.get_value(
x, xerr) #so this is currently shit. can't get the xErr thing working
print("fit_meanM:", fit_meanM)
#fit results from search phase
best_fit_params = (1.0779, 2.04035, 58.5769, -157.945)
fit_mean = model_UA2(xFit, best_fit_params, xerrFit)
##----4 param fit function
#lnProb = logLike_4ff(xFit,yFit,xerrFit)
#minimumLLH, best_fit_params = fit_4ff(100, lnProb)
#fit_mean = model_4param(xFit, best_fit_params, xerrFit)
#calculating significance
signif = significance(mu_x, y, cov_x, yerr)
initialCutPos = np.argmax(x > xMinFit)
finalCutPos = np.argmin(x < xMaxFit)
#sigFit = (fit_mean - y[initialCutPos:]) / np.sqrt(np.diag(cov_x[initialCutPos:]) + yerr[initialCutPos:]**2)
sigFit = significance(fit_mean, y[initialCutPos:], cov_x[initialCutPos:],
yerr[initialCutPos:])
#sigit = (mu_x)
#std = np.sqrt(np.diag(cov))
ext = args.output_file_extension
title = "compareGPvsSearchPhaseUA2"
with Canvas(f'%s{ext}' % title) as can:
can.ax.errorbar(x, y, yerr=yerr, fmt='.')
can.ax.set_yscale('log')
# can.ax.set_ylim(1, can.ax.get_ylim()[1])
can.ax.plot(t, mu, '-r')
can.ax.plot(xFit, fit_mean, '-b')
#this only works with the xErr part of Mean commented out
#can.ax.plot(x, fit_meanM, '.g')
# can.ax.plot(t, fit_mean_smooth, '--b')
#can.ax.fill_between(t, mu - std, mu + std,
#facecolor=(0, 1, 0, 0.5),
#zorder=5, label='err = 1')
#can.ratio.stem(x, signif, markerfmt='.', basefmt=' ')
can.ratio.stem(xFit, sigFit, markerfmt='.', basefmt=' ')
can.ratio.axhline(0, linewidth=1, alpha=0.5)
can.save(title)