/usr/share/pyshared/fabio/bruker100image.py is in python-fabio 0.0.8-1.
This file is owned by root:root, with mode 0o644.
The actual contents of the file can be viewed below.
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 | import numpy as N
import math
import Image
from brukerimage import brukerimage
import readbytestream
class bruker100image(brukerimage):
def toPIL16(self, filename=None):
# FIXME - why is this different for bruker100images?
if filename:
self.read(filename)
PILimage = Image.frombuffer("F",
(self.dim1, self.dim2),
self.data,
"raw",
"F;16", 0, -1)
return PILimage
else:
PILimage = Image.frombuffer("F",
(self.dim1, self.dim2),
self.data,
"raw",
"F;16", 0, -1)
return PILimage
def read(self, fname):
f = open(fname, "rb")
try:
self._readheader(f)
except:
raise
rows = int(self.header['NROWS'])
cols = int(self.header['NCOLS'])
npixelb = int(self.header['NPIXELB'][0])
# you had to read the Bruker docs to know this!
# We are now at the start of the image - assuming
# readbrukerheader worked
# size = rows * cols * npixelb
self.data = readbytestream(f, f.tell(), rows, cols, npixelb,
datatype="int", signed='n', swap='n')
noverfl = self.header['NOVERFL'].split() # now process the overflows
#read the set of "underflow pixels" - these will be completely
# disregarded for now
data = self.data
k = 0
while k < 2:#for the time being things - are done in 16 bits
datatype = {'1' : numpy.uint8,
'2' : numpy.uint16,
'4' : numpy.uint32 }[("%d" % 2 ** k)]
ar = numpy.array(numpy.fromstring(f.read(int(noverfl[k]) * (2 ** k)),
datatype), numpy.uint16)
#insert the the overflow pixels in the image array:
#this is probably a memory intensive way of doing this -
# might be done in a more clever way
lim = 2 ** (8 * k) - 1
#generate an array comprising of the indices into data.ravel()
# where its value equals lim.
M = numpy.compress(numpy.equal(data.ravel(), lim), numpy.arange(rows * cols))
#now put values from ar into those indices
numpy.put(data.ravel(), M, ar)
padding = 16 * int(math.ceil(int(noverfl[k]) * (2 ** k) / 16.)) - \
int(noverfl[k]) * (2 ** k)
f.seek(padding, 1)
print noverfl[k] + " bytes read + %d bytes padding" % padding
k = k + 1
f.close()
(self.dim1, self.dim2) = (rows, cols)
print self.dim1, self.dim2
self.resetvals()
return self
if __name__ == '__main__':
import sys, time
I = bruker100image()
b = time.clock()
while (sys.argv[1:]):
I.read(sys.argv[1])
r = I.toPIL16()
I.rebin(2, 2)
print sys.argv[1] + (": max=%d, min=%d, mean=%.2e, stddev=%.2e") % (
I.getmax(), I.getmin(), I.getmean(), I.getstddev())
print 'integrated intensity (%d %d %d %d) =%.3f' % (
10, 20, 20, 40, I.integrate_area((10, 20, 20, 40)))
sys.argv[1:] = sys.argv[2:]
e = time.clock()
print (e - b)
|