def test_overflow() :
h = Histogram(axis=Axis(100, -50., 50.))
assert h.overflow() ==0
h.fill(100.)
h.fill(200.)
h.fill(10.)
assert h.overflow() == 2.
def test_fill():
h = Histogram(axis=Axis(100, -50, 50))
weight = 0.25
for i in xrange(-50,50) :
h.fill(i+0.5, weight)
assert h.entries() == 100
assert h.integral() == 100*weight
def test_integral() :
h = Histogram(axis=Axis(100, -50., 50.))
weight = 0.5
for x in xrange(10) :
h.fill(-49.99, weight) # should be 0th bin
h.fill(49.99, weight) # should be 100th bin, ie index 99
h.fill(0.1, 2*weight)
assert h.integral() == 10*4*weight
def test_binheight() :
h = Histogram(axis=Axis(100, -50., 50.))
weight = 0.5
h.fill(-49.99, weight) # should be 0th bin
h.fill(49.99, weight) # should be 100th bin, ie index 99
h.fill(0.1, 2*weight)
assert h.binHeight(0) == weight
assert h.binHeight(99) == weight
assert h.binHeight(50) == 2*weight
def test_mean() :
h0 = Histogram(axis=Axis(100, -50, 50))
h1 = Histogram(axis=Axis(100, -50, 50))
for x in xrange(-50, 50) :
h0.fill(x + 0.5)
if x + 0.5 > 0 :
h1.fill(x+0.5)
assert h1.mean() == 25.
assert h0.mean() == 0
def test_multiply_histograms() :
from random import gauss
h0 = Histogram(axis=Axis(100, -50., 50.))
h1 = Histogram(axis=Axis(100, -50., 50.))
dist = [gauss(0.,20) for x in xrange(100)]
h0.fill(dist)
h1.fill(dist, 2.5)
h2 = h0*h1
for bin in xrange(h2.axis.nbins):
assert h2.binHeight(bin) == h0.binHeight(bin) * h1.binHeight(bin)
def test_divide_histograms() :
from random import gauss
h0 = Histogram(axis=Axis(100, -50., 50.))
h1 = Histogram(axis=Axis(100, -50., 50.))
dist = [gauss(0.,10) for x in xrange(100)]
h0.fill(dist)
h1.fill(dist, 5.)
h2 = h0/h1
for bin in xrange(h2.axis.nbins):
if h1.binHeight(bin) != 0 :
assert h2.binHeight(bin) == h0.binHeight(bin) / h1.binHeight(bin)
def ntuple_column_histo(ntuple, tag, cut=None, bins=100):
'''
Create and return a histogram.Histogram object constructed from the
contents of an ntuple.NTuple column.
'''
from PyAna.pyhistogram.histogram import Histogram, Axis
col = ntuple.column(tag, cut)
min_range = min(col)
max_range = max(col)
width = (max_range - min_range) / bins
max_range += width / 2.
max_range -= width / 2.
histo = Histogram(axis=Axis(bins, min_range, max_range, label=tag))
for x in col:
histo.fill(x)
return histo
def test_underflow() :
h = Histogram(axis=Axis(100, -50., 50.))
assert h.underflow() == 0.
h.fill(-100.)
assert h.underflow() == 1.
def test_fill_with_weight():
import math
h = Histogram(axis=Axis(100, -50, 50))
for i in xrange(-50,50) :
w = math.sin(i)
h.fill(i, w)
def test_fill_from_list():
h = Histogram(axis=Axis(100, -50, 50))
x = [float(i) for i in xrange(-49,50)]
h.fill(x)
assert h.entries() == 99
def test_min_and_max() :
h0 = Histogram(axis=Axis(20, -10., 10.))
h0.fill([-5.5,6.5])
assert h0.min() == -6.0
assert h0.max() == 7.0
def test_sigma() :
h0 = Histogram(axis=Axis(100, -50., 50.))
h0.fill([5.5,-5.5,100.5,-100.5])
assert h0.sigma() == 5.5
def test_subtract_histograms() :
h0 = Histogram(axis=Axis(100, -50., 50.))
weight = 0.5
h0.fill([-49.99,49.99,-48.99,48.99], weight) # should be 0th bin
h0.fill(0.1, 4*weight)
h0.fill(-75., 32.)
h0.fill(100., 55)
h1 = Histogram(axis=Axis(100, -50., 50.))
h1.fill([-48.99,48.99], weight) # should be 0th bin
h1.fill(0.1, 2*weight)
h1.fill(-101.)
h1.fill(200.,2)
h2 = h0 - h1
assert h2.axis == h0.axis
assert h2.entries() == h0.entries() - h1.entries()
assert h2.overflow() == h0.overflow() - h1.overflow()
assert h2.underflow() == h0.underflow() - h1.underflow()
for bin in xrange(h2.axis.nbins):
assert h2.binHeight(bin) == h0.binHeight(bin) - h1.binHeight(bin)
def test_entries_does_not_count_out_of_range_entries() :
h = Histogram(axis=Axis(100, -50., 50.))
h.fill(45.5, 10.)
h.fill(100.)
h.fill(-100)
assert h.entries() == 3.