Source code for fabio.fit2dmaskimage

# Get ready for python3:
from __future__ import with_statement, print_function
__doc__ = """
Author: Andy Hammersley, ESRF
Translation into python/fabio: Jon Wright, ESRF.
Writer: Jérôme Kieffer
"""

__authors__ = ["Jon Wright", "Jérôme Kieffer"]
__contact__ = "Jerome.Kieffer@esrf.fr"
__license__ = "GPLv3+"
__copyright__ = "European Synchrotron Radiation Facility, Grenoble, France"
__version__ = "28 Oct 2013"

import numpy
import sys
from fabioimage import fabioimage
if sys.version_info < (3.0):
    bytes = str


class fit2dmaskimage(fabioimage):
    """ Read and try to write Andy Hammersley's mask format """

[docs] def _readheader(self, infile): """ Read in a header from an already open file """ # 1024 bytes gives 256x32 bit integers header = infile.read(1024) for i, j in [ ("M", 0), ("A", 4), ("S", 8), ("K", 12) ]: if header[j] != i: raise Exception("Not a fit2d mask file") fit2dhdr = numpy.fromstring(header, numpy.int32) self.dim1 = fit2dhdr[4] # 1 less than Andy's fortran self.dim2 = fit2dhdr[5] def read(self, fname, frame=None): """ Read in header into self.header and
[docs] the data into self.data """ fin = self._open(fname) self._readheader(fin) # Compute image size self.bytecode = numpy.uint8 self.bpp = len(numpy.array(0, self.bytecode).tostring()) # integer division num_ints = (self.dim1 + 31) // 32 total = self.dim2 * num_ints * 4 data = fin.read(total) assert len(data) == total fin.close() # Now to unpack it data = numpy.fromstring(data, numpy.uint8) data = numpy.reshape(data, (self.dim2, num_ints * 4)) result = numpy.zeros((self.dim2, num_ints * 4 * 8), numpy.uint8) # Unpack using bitwise comparisons to 2**n bits = numpy.ones((1), numpy.uint8) for i in range(8): temp = numpy.bitwise_and(bits, data) result[:, i::8] = temp.astype(numpy.uint8) bits = bits * 2 # Extra rows needed for packing odd dimensions spares = num_ints * 4 * 8 - self.dim1 if spares == 0: self.data = numpy.where(result == 0, 0, 1) else: self.data = numpy.where(result[:, :-spares] == 0, 0, 1) # Transpose appears to be needed to match edf reader (scary??) # self.data = numpy.transpose(self.data) self.data = numpy.reshape(self.data.astype(numpy.uint16), (self.dim2, self.dim1)) self.pilimage = None return self def write(self, fname): """ Try to write a file
[docs] """ header = numpy.zeros(1024, dtype=numpy.uint8) header[0] = 77 #M header[4] = 65 #A header[8] = 83 #S header[12] = 75 #K header[24] = 1 #1 str1 = numpy.array([self.dim1], numpy.uint32).tostring() str2 = numpy.array([self.dim2], numpy.uint32).tostring() if not numpy.little_endian: str1 = str1[-1::-1] str2 = str2[-1::-1] for i, c in zip(range(16, 20), str1): header[i] = ord(c) for i, c in zip(range(20, 24), str2): header[i] = ord(c) compact_array = numpy.zeros((self.dim2, ((self.dim1 + 31) // 32) * 4), dtype=numpy.uint8) large_array = numpy.zeros((self.dim2, ((self.dim1 + 31) // 32) * 32), dtype=numpy.uint8) large_array[:self.dim2, :self.dim1] = (self.data != 0) order = 1 for i in range(8): compact_array += large_array[:, i::8] * order order *= 2 with self._open(fname, mode="wb") as outfile: outfile.write(header) outfile.write(compact_array.tostring()) @staticmethod def checkData(data=None): if data is None:
return None
[docs] else: return data.astype(int)