def doit_numarray(L):
import numarray
L = numarray.array(L)
t0 = time.clock()
numarray.sort(L)
t1 = time.clock()
print "%6.2f" % (t1-t0),
fl()
def cumulative(input):
#print "max x = ",max(input)
x = N.sort(input)
#print "max x = ",max(x)
n = len(input)
y = N.arange(0., 1., (1. / float(n)))
return x, y
def mad(xdata, xmed):
# The XMAD subroutine calculates the Median Absolute Deviation from
# the sample median. The median, M , is subtracted from each
# ORDERED statistic and then the absolute value is taken. This new
# set of of statistics is then resorted so that they are ORDERED
# statistics. The MAD is then defined to be the median of this
# new set of statistics and is returned as XMADM. The MAD can
# be defined:
#
# XMADM = median{ abs(x[i] - M) }
#
# where the x[i] are the values passed in the array XDATA, and
# the median, M, is passed in the array XLETTER. The set of stats
# in the brackets is assumed to be resorted. For more information
# see page 408 in UREDA.
n = len(xdata)
dhalf, n1, n2 = (0.5, 1, 2)
xdata2 = N.absolute(xdata - xmed)
xdata2 = N.sort(xdata2,0)
if (float(n)/float(n2) - int(n/n2) == 0):
i1 = n/n2
i2 = n/n2 - n1
xmadm = dhalf*(xdata2[i1] + xdata2[i2])
else:
i1 = n/n2
xmadm = xdata2[i1]
return xmadm
def median(x, sorted=0):
if not sorted: x = N.sort(x,axis=0)
n = len(x)
n2 = n/2
if 2*n2 == n:
xmed = 0.5*(x[n2-1]+x[n2])
else:
xmed = x[n2]
return xmed
def quartiles(x, sorted=0):
if not sorted: x = N.sort(x)
n = len(x)
n2 = n/2
n4 = n/4
if 2*n2 == n:
xlq = 0.75*x[n4] + 0.25*x[n4+1]
xhq = 0.25*x[3*n4] + 0.75*x[3*n4+1]
else:
xlq = x[n4]
xhq = x[3*n4]
return (xlq, xhq)
def kuiper_test(x, y):
x = N.sort(x)
y = N.sort(y)
nx = len(x)
ny = len(y)
jx = jy = 0
fnx = fny = 0.0
fnx_arr = N.zeros(nx, N.Float32)
fny_arr = N.zeros(ny, N.Float32)
binx = N.zeros(nx, N.Float32)
biny = N.zeros(ny, N.Float32)
d1 = d2 = 0.0
while (jx < nx or jy < ny):
if jx < nx:
dx = x[jx]
else:
dx = x[-1]
if jy < ny:
dy = y[jy]
else:
dy = y[-1]
if (dx <= dy or jy >= ny-1) and jx < nx:
fnx = (jx+1)/float(nx)
binx[jx] = dx
fnx_arr[jx] = fnx
jx += 1
if (dy <= dx or jx >= nx-1) and jy < ny:
fny = (jy+1)/float(ny)
biny[jy] = dy
fny_arr[jy] = fny
jy += 1
dt1 = (fny-fnx)
dt2 = (fnx-fny)
d1 = max(d1, dt1)
d2 = max(d2, dt2)
v = d1 + d2
n = sqrt(nx*ny/(nx+ny))
prob = probkuiper((n+0.155+0.24/n)*v)
return v, prob, [binx, fnx_arr], [biny, fny_arr]
def ks_test(x, y):
x = N.sort(x)
y = N.sort(y)
nx = len(x)
ny = len(y)
jx = jy = 0
fnx = fny = 0.0
fnx_arr = N.zeros(nx, N.Float32)
fny_arr = N.zeros(ny, N.Float32)
binx = N.zeros(nx, N.Float32)
biny = N.zeros(ny, N.Float32)
d = 0.0
while (jx < nx or jy < ny):
if jx < nx:
dx = x[jx]
else:
dx = x[-1]
if jy < ny:
dy = y[jy]
else:
dy = y[-1]
if (dx <= dy or jy >= ny-1) and jx < nx:
fnx = (jx+1)/float(nx)
binx[jx] = dx
fnx_arr[jx] = fnx
jx += 1
if (dy <= dx or jx >= nx-1) and jy < ny:
fny = (jy+1)/float(ny)
biny[jy] = dy
fny_arr[jy] = fny
jy += 1
dt = abs(fny-fnx)
if dt > d:
d = dt
#print '%5.3f %5.3f %5.3f %5.3f %5.3f %5.3f'%(dx, dy, fnx, fny, dt, d)
n = sqrt(nx*ny/(nx+ny))
prob = probks((n+0.12+0.11/n)*d)
return d, prob, [binx, fnx_arr], [biny, fny_arr]
def zscale(data,contrast,min=100,max=60000):
"""Scale the data cube into the range 0-255"""
## pic 100 random elements along each dimension
## use zscale (see the IRAF display man page or
## http://iraf.net/article.php/20051205162333315
import random
x=[]
for i in random.sample(xrange(data.shape[0]),50):
for j in random.sample(xrange(data.shape[1]),50):
x.append(data[i,j])
yl=numarray.sort(numarray.clip(x,min,max))
n=len(yl)
ym=sum(yl)/float(n)
xl=numarray.array(range(n))
xm=sum(xl)/float(n)
ss_xx=sum((xl-xm)*(xl-xm))
ss_yy=sum((yl-ym)*(yl-ym))
ss_xy=sum((xl-xm)*(yl-ym))
b=ss_xy/ss_xx
a=ym-b*xm
z1=yl[n/2] + (b/contrast)*(1-n/2)
z2=yl[n/2] + (b/contrast)*(n-n/2)
## Now put the data inbetween Z1 and Z2
high=data-z1
z2=z2-z1
high=numarray.clip(high,0,z2)
## and change that to 0-255
high= 256-256*high/z2
### send back the scalled data
return high
def zscale(data,contrast,min=100,max=60000):
"""Scale the data cube into the range 0-255"""
## pic 100 random elements along each dimension
## use zscale (see the IRAF display man page or
## http://iraf.net/article.php/20051205162333315
import random
x=[]
for i in random.sample(xrange(data.shape[0]),50):
for j in random.sample(xrange(data.shape[1]),50):
x.append(data[i,j])
yl=numarray.sort(numarray.clip(x,min,max))
n=len(yl)
ym=sum(yl)/float(n)
xl=numarray.array(range(n))
xm=sum(xl)/float(n)
ss_xx=sum((xl-xm)*(xl-xm))
ss_yy=sum((yl-ym)*(yl-ym))
ss_xy=sum((xl-xm)*(yl-ym))
b=ss_xy/ss_xx
a=ym-b*xm
z1=yl[n/2] + (b/contrast)*(1-n/2)
z2=yl[n/2] + (b/contrast)*(n-n/2)
## Now put the data inbetween Z1 and Z2
high=data-z1
z2=z2-z1
high=numarray.clip(high,0,z2)
## and change that to 0-255
high= 256-256*high/z2
### send back the scalled data
return high
def median(x):
x = N.sort(x)
n = float(len(x))
nmed = int(n / 2.)
return x[nmed]
def cumulative(input):
x = N.sort(input)
n = len(input)
y = N.arange(0, 1, (1. / n))
return x, y
