def edf_keywords_completion(edfin, keywords, edfout=None):
"""This procedure takes an EDF file and completes it with the keywords given as CIF dictionnary
@param edfin: The name or the path of an EDF file to be read
@type edfin: Python String
@param keywords: dictionary containing the CIF-like data
@type keywords: Python dictionary
@param edfout: name or the path of an EDF file to be written. Caution, this WILL overwrite the file. If nothing is given, the name witll be edfin+keywords.edf
@type edfout: Python String
"""
if not edfout: edfout = os.path.splitext(edfin)[0] + "+keywords.edf"
print "processing file %s --->%s " % (edfin, edfout)
infile = EDF.EdfFile(edfin)
data = infile.GetData(0)
headers = infile.GetHeader(0)
for key in keywords:
if key.lower() == "loop_":
#As loops are not possible in EDF headers, we replace them by a kind of list
loops = keywords[key]
for oneLoop in loops:
oneLoopIdx = 0 #this is a index of the loop
for oneLoopdict in oneLoop[1]:
oneLoopIdx += 1
for loopKey in oneLoop[0]:
headers["%s[%i]" %
(loopKey, oneLoopIdx)] = oneLoopdict[loopKey]
else:
headers[key] = keywords[key]
outfile = EDF.EdfFile(edfout)
outfile.WriteImage(headers, data)
def getBGarray(self, bg_filename): if self.bg_path == '-' and bg_filename == '-': self.bg_combined = np.zeros((self.roi[3], self.roi[1])) else: bg_file_with_path = self.bg_path + '/' + self.bg_files[0] bg_class = EdfFile.EdfFile(bg_file_with_path) bg_img = bg_class.GetData(0).astype(np.int64)[ self.roi[2]:self.roi[3], self.roi[0]:self.roi[1]] self.bg_combined = np.zeros(np.shape(bg_img)) if self.rank == 0: print "Reading background files (ROI)..." for i in range(len(self.bg_files)): bg_file_with_path = self.bg_path + '/' + self.bg_files[i] bg_class = EdfFile.EdfFile(bg_file_with_path) self.bg_combined += bg_class.GetData(0).astype(np.int64)[ self.roi[2]:self.roi[3], self.roi[0]:self.roi[1]] self.bg_combined /= len(self.bg_files) bg_img_full = bg_class.GetData(0).astype(np.int64) self.bg_combined_full = np.zeros(np.shape(bg_img_full)) if self.rank == 0: print "Reading background files (Full)..." for i in range(len(self.bg_files)): bg_file_with_path = self.bg_path + '/' + self.bg_files[i] bg_class = EdfFile.EdfFile(bg_file_with_path) self.bg_combined_full += bg_class.GetData(0).astype(np.int64) self.bg_combined_full /= len(self.bg_files)
def testExecute(self):
"""
"""
self.run()
plugin = self.getPlugin()
################################################################################
# Compare XSDataResults
################################################################################
strExpectedOutput = self.readAndParseFile(
self.getReferenceDataOutputFile())
# strObtainedOutput = self.readAndParseFile (self.m_edObtainedOutputDataFile)
EDVerbose.DEBUG("Checking obtained result...")
xsDataResultReference = XSDataResultBioSaxsMetadatav1_0.parseString(
strExpectedOutput)
xsDataResultObtained = plugin.getDataOutput()
EDAssert.strAlmostEqual(xsDataResultReference.marshal(),
xsDataResultObtained.marshal(),
"XSDataResult output are the same",
_strExcluded="bioSaxs")
################################################################################
# Compare dictionary:
################################################################################
edfRef = EdfFile.EdfFile(
xsDataResultObtained.getOutputImage().getPath().value)
edfObt = EdfFile.EdfFile(
os.path.join(self.getTestsDataImagesHome(), "bioSaxsMetadata.edf"))
########################################################################
# DEPRECATED PLUGIN => DEPREFCATED TESTS
########################################################################
# headerRef = edfRef.GetHeader(0)
# headerObt = edfObt.GetHeader(0)
# keysRef = headerRef.keys()
# keysObt = headerObt.keys()
# keysRef.sort()
# keysObt.sort()
# for key in ["HeaderID", "Image", 'EDF_BinarySize', "EDF_DataBlockID", "EDF_HeaderSize", "filename", "RasterOrientation", "Center_1", "Center_2", "Code", "Comments", "Concentration",
# "VersionNumber",'time_of_day', ]:
# if key in keysObt: keysObt.remove(key)
# if key in keysRef: keysRef.remove(key)
# EDAssert.equal(keysRef, keysObt, _strComment="Same keys in the header dictionary")
# for key in keysRef:
# EDAssert.strAlmostEqual(headerRef[key], headerObt[key], _strComment="header value %s are the same" % key, _strExcluded="bioSaxs")
################################################################################
# Compare images
################################################################################
outputData = edfObt.GetData(0)
referenceData = edfRef.GetData(0)
EDAssert.arraySimilar(outputData,
referenceData,
_fAbsMaxDelta=0.1,
_fScaledMaxDelta=0.05,
_strComment="Images-data are the same")
def getBGarray(self, bg_filename):
if self.bg_path == '-' and bg_filename == '-':
self.bg_combined = np.zeros((self.roi[3], self.roi[1]))
else:
bg_file_with_path = self.bg_path + '/' + self.bg_files[0]
bg_class = EdfFile.EdfFile(bg_file_with_path)
bg_img = bg_class.GetData(0).astype(
np.int64)[self.roi[2]:self.roi[3], self.roi[0]:self.roi[1]]
self.bg_combined = np.zeros(np.shape(bg_img))
if self.rank == 0:
print "Reading background files (ROI)..."
for i in range(len(self.bg_files)):
bg_file_with_path = self.bg_path + '/' + self.bg_files[i]
bg_class = EdfFile.EdfFile(bg_file_with_path)
self.bg_combined += bg_class.GetData(0).astype(
np.int64)[self.roi[2]:self.roi[3], self.roi[0]:self.roi[1]]
self.bg_combined /= len(self.bg_files)
bg_img_full = bg_class.GetData(0).astype(np.int64)
self.bg_combined_full = np.zeros(np.shape(bg_img_full))
if self.rank == 0:
print "Reading background files (Full)..."
for i in range(len(self.bg_files)):
bg_file_with_path = self.bg_path + '/' + self.bg_files[i]
bg_class = EdfFile.EdfFile(bg_file_with_path)
self.bg_combined_full += bg_class.GetData(0).astype(np.int64)
self.bg_combined_full /= len(self.bg_files)
bckg = self.bg_combined
bckg_entire = self.bg_combined_full
# Check background images
#fig = plt.figure()
#a=fig.add_subplot(1,2,1)
#plt.imshow(bckg)
#a.set_title('Background ROI')
#b=fig.add_subplot(1,2,2)
#plt.imshow(bckg_entire)
#b.set_title('All background')
#plt.show()
np.save('bckg_roi.npy', bckg)
np.save('bckg_all.npy', bckg_entire)
def testExecute(self):
"""
"""
self.run()
plugin = self.getPlugin()
################################################################################
# Compare XSDataResults
################################################################################
strExpectedOutput = self.readAndParseFile (self.getReferenceDataOutputFile())
# strObtainedOutput = self.readAndParseFile (self.m_edObtainedOutputDataFile)
EDVerbose.DEBUG("Checking obtained result...")
xsDataResultReference = XSDataResultBioSaxsAveragev1_0.parseString(strExpectedOutput)
xsDataResultObtained = plugin.getDataOutput()
EDAssert.strAlmostEqual(xsDataResultReference.marshal(), xsDataResultObtained.marshal(), "XSDataResult output are the same", _strExcluded="bioSaxs")
################################################################################
# Compare spectrum ascii Files
################################################################################
outputData = open(xsDataResultObtained.getAveragedCurve().getPath().value, "rb").read()
referenceData = open(os.path.join(self.getTestsDataImagesHome(), "bioSaxsAveraged.dat"), "rb").read()
EDAssert.strAlmostEqual(referenceData, outputData, _strComment="3-column ascii spectra files spectra are the same", _fRelError=0.1, _fAbsError=0.1, _strExcluded="bioSaxs")
################################################################################
# Compare images
################################################################################
edfObt = EdfFile.EdfFile(xsDataResultObtained.getAveragedImage().getPath().value)
edfRef = EdfFile.EdfFile(os.path.join(self.getTestsDataImagesHome(), "bioSaxsAveraged.edf"))
outputData = edfObt.GetData(0)
referenceData = edfRef.GetData(0)
EDAssert.arraySimilar(outputData, referenceData , _fAbsMaxDelta=0.1, _fScaledMaxDelta=0.05, _strComment="Averaged images are the same")
headerRef = edfRef.GetHeader(0)
headerObt = edfObt.GetHeader(0)
keysRef = headerRef.keys()
keysObt = headerObt.keys()
keysRef.sort()
keysObt.sort()
for key in ["HeaderID", "Image", 'EDF_BinarySize', "EDF_DataBlockID", "EDF_HeaderSize", "filename", "RasterOrientation", "History-1", "History-1~1" ]:
if key in keysObt: keysObt.remove(key)
if key in keysRef: keysRef.remove(key)
EDAssert.equal(keysRef, keysObt, _strComment="Same keys in the header dictionary for Corrected Images")
for key in keysRef:
EDAssert.strAlmostEqual(headerRef[key], headerObt[key], _strComment="header value in Averaged %s are the same" % key, _strExcluded="bioSaxs")
def headeredf(filename, imgn=0):
if isfile(filename):
f = EdfFile.EdfFile(filename)
return f.GetHeader(imgn)
else:
print "file ", filename, " does not exist!"
return
def __saveEDF(self, fileSelected):
if not fileSelected:
return
fileSelected = str(fileSelected)
data = self._saveEDFPlug._data()
edf_file = EdfFile.EdfFile(fileSelected)
edf_file.WriteImage({}, data)
def loadedf(filename, imgn=0):
if isfile(filename):
f = EdfFile.EdfFile(filename)
return f.GetData(imgn)
else:
print "file ", filename, " does not exist!"
return
def read(filename):
trcfile = EdfFile.EdfFile(filename)
trcheader = trcfile.GetHeader(0)
trcdata = asfarray(trcfile.GetData(0))
trcdata = trcdata / 2**16 * (float(trcheader['MaxValue']) - float(
trcheader['MinValue'])) + float(trcheader['MinValue'])
return trcdata
def __init__(self, filename):
try:
f = EdfFile.EdfFile(filename)
except:
raise UserInputException("""Unable to read in the file \
%s because the Object EdfFile that I am using to read in edf data \
raised an error when trying to read in the file. This probably means \
that there is something wrong \
with the file that you are trying to open""" % filename)
data = f.GetData(0)
if not alltrue(alltrue(less_equal(data, 2147483647))):
print """Warning, some of the data stored in the \
file %s has an intensity larger then 2^31-1 which is too big for this \
program to hold. Any of these large values were clipped to have a value
of 2^31-1.""" % filename
# clip any data that is too big
mask1 = data <= 2147483647
mask2 = data > 2147483647
masked_data = data * mask1 + (pow(2, 31) - 1) * mask2
masked_data = masked_data.astype(Numeric.Int32)
self.size = max(masked_data.shape[0], masked_data.shape[1])
# pad values if necessary - create an array to put everything in
self.data = Numeric.zeros((self.size, self.size), Numeric.Int32)
# copy the data into the padded array
self.data[0:masked_data.shape[0], 0:masked_data.shape[1]] = masked_data
self.data = Numeric.transpose(self.data)
def createAverageWfandDf(self):
if not (os.path.exists(self.dirname + '/refHST0000.edf')):
print('NO REF ')
refBeg = glob.glob(self.dirname + '/ref*_0000.edf')
print(refBeg)
if len(refBeg) > 0:
self.outputFileFFNameBeg = self.dirname + '/refForHST0000.edf'
vBeg = Averager.Averager(refBeg,
self.outputFileFFNameBeg,
option=0)
if (self.ref_on < self.numberOfProjections):
cpt = self.ref_on
while cpt < self.numberOfProjections:
textref = '%4.4d' % cpt
refBeg = glob.glob(self.dirname + '/ref*_' + textref +
'.edf')
if len(refBeg > 0):
self.outputFileNameFFEnd = self.dirname + '/refForHST' + textref + '.edf'
vBeg = Averager.Averager(refBeg,
self.outputFileNameFFEnd,
option=1)
cpt += self.ref_on
else:
textref = '%4.4d' % self.ref_on
refBeg = glob.glob(self.dirname + '/ref*_' + textref + '.edf')
if len(refBeg) > 0:
self.outputFileNameFFEnd = self.dirname + '/refForHST' + textref + '.edf'
vBeg = Averager.Averager(refBeg,
self.outputFileNameFFEnd,
option=0)
darkFile = self.dirname + '/darkend0000.edf'
data = edf.EdfFile(darkFile).GetData(0)
data = np.divide(data, self.numberOfDarkFields)
self.darkOutputFile = self.dirname + '/darkForHST0000.edf'
filetoWrite = edf.EdfFile(self.darkOutputFile, access='wb+')
filetoWrite.WriteImage({}, data, Append=0, DataType='FloatValue')
else:
self.outputFileFFNameBeg = self.dirname + '/refHST0000.edf'
self.outputFileNameFFEnd = self.dirname + '/refHST0000.edf'
self.darkOutputFile = self.dirname + '/dark.edf'
self.averageDone = True
def save3DImage(self, outputName):
self.nbSlices, self.width, self.height = self.data.shape
for k in range(self.nbSlices):
textNumSlice = '%4.4d' % k
finalOutputName = outputName + textNumSlice + '.edf'
filetoWrite = edf.EdfFile(finalOutputName, access='wb+')
dataToStore = self.data[k, :, :].squeeze()
filetoWrite.WriteImage({}, dataToStore)
def saveSino(self, outputName):
for i in range(self.width):
textNumSlice = '%4.4d' % i
finalOutputName = outputName + textNumSlice + '.edf'
filetoWrite = edf.EdfFile(finalOutputName, access='wb+')
dataToStore = self.data[:, i, :].squeeze()
filetoWrite.WriteImage({}, dataToStore)
def createAverageWfandDfPbm(self):
refBeg = glob.glob(self.dirname + '/RefA*.edf')
print(refBeg)
if len(refBeg) > 0:
self.outputFileFFNameBeg = self.dirname + '/refForHST0000.edf'
vBeg = Averager.Averager(refBeg,
self.outputFileFFNameBeg,
option=0)
ref = edf.EdfFile(refBeg[0], access='r')
ref = ref.GetData(0)
self.darkOutputFile = self.dirname + '/darkForHST0000.edf'
data = np.ones(ref.shape)
filetoWrite = edf.EdfFile(self.darkOutputFile, access='wb+')
filetoWrite.WriteImage({}, data, Append=0, DataType='FloatValue')
self.averageDone = True
def saveedf(filename, data, imgn=0):
try:
newf = EdfFile.EdfFile(filename)
newf.WriteImage({}, data, imgn)
print "file is saved to ", filename
return
except:
print "file is not saved!"
return
def makeDarkMean(Darkfiedls):
nbslices, height, width = Darkfiedls.shape
meanSlice = np.mean(Darkfiedls, axis=0)
print ('----------------------- mean Dark calculation done ------------------------- ')
OutputFileName = '/Users/helene/PycharmProjects/spytlab/meanDarkTest.edf'
outputEdf = edf.EdfFile(OutputFileName, access='wb+')
outputEdf.WriteImage({}, meanSlice)
return meanSlice
def readImage(filename):
if filename.endswith('.edf') or filename.endswith('.tiff') :
edfFile = edf.EdfFile(filename, access='rb')
im2D = edfFile.GetData(0)
if filename.endswith('.tiff') or filename.endswith('.tif') or filename.endswith('.png') or filename.endswith('.TIF') or filename.endswith('.TIFF') :
img=Image.open(filename)
im2D= np.array(img)
return im2D
def openImage(filename):
filename = str(filename)
if filename.endswith('.edf'):
im = edf.EdfFile(filename, access='rb')
imarray = im.GetData(0)
else:
imarray = Image.open(filename)
imarray = np.array(imarray)
return imarray
def read_edf_slices( imageDimensions, base_dir, edf_base_name, digits, extension, slices_range, crop=None ):
"""
This reads EDF slices, and returns a 3D volume
INPUT:
- image_size size of RAW files.
- base_dir the directory inside which the raw files are.
- edf_base_name the "base" name of the raw files
- digits pre-padding with zeros in file number
- extension extension of raw files
- slices_range tuple or list of number of slices to read
- crop list of lists: [ [ x_min, x_max ], [ y_min, y_max ] ]
"""
import numpy
import EdfFile
# ET: there should be a +1 here because if I wanto to load slices from 1 to 5 this give 5 slices and not 4
numberOfSlices = slices_range[1] - slices_range[0] + 1
# Define array for image loading
if crop == None:
# If we don't have a crop, we're going to use the whole slice size.
outputVolume = numpy.ones( ( numberOfSlices, imageDimensions[1], imageDimensions[0] ), dtype='<f4' ) * numpy.nan
else:
# Slice dimensions from crop
outputVolume = numpy.ones( ( numberOfSlices, crop[1][1] - crop[1][0], crop[0][1] - crop[0][0] ), dtype='<f4' ) * numpy.nan
# Load all images into big array
for sliceNumber in range( numberOfSlices ):
try:
# 2016-04-30 ET: if worrking in 2D digits == 0 and the name in constructed differently
if digits ==0:
filename = "%s/%s%s"%( base_dir, edf_base_name, extension )
else:
filename = "%s/%s%0*i%s"%( base_dir, edf_base_name, digits, sliceNumber + slices_range[0], extension )
currentImage = numpy.array( EdfFile.EdfFile( filename ).GetData( 0 ) )
if crop == None:
outputVolume[ sliceNumber ] = currentImage
else:
outputVolume[ sliceNumber ] = currentImage[ crop[1][0]:crop[1][1], crop[0][0]:crop[0][1] ]
except :
try: logging.log.warning( "read_edf_slices(): File %s not found "%filename )
except: print "read_edf_slices(): File %s not found "%filename
pass
try: logging.log.debug( "read_edf_pil_slices(): Volume mean value = %s"%( outputVolume.mean() ) )
except: print "read_edf_pil_slices(): Volume mean value = %s"%( outputVolume.mean() )
return outputVolume
def write(filename, trcdata):
trcfile = EdfFile.EdfFile(filename)
maxtrc = trcdata.max()
mintrc = trcdata.min()
trcdata = (trcdata - mintrc) / (maxtrc - mintrc) * 2**16
trcfile.WriteImage({
'MaxValue': maxtrc,
'MinValue': mintrc
},
trcdata,
0,
DataType='UnsignedShort')
def edf(self):
"""
Read data from a edf file.
Returns
-------
ndarray
Data.
"""
f = EdfFile.EdfFile(self.fname, access='r')
d = f.GetStaticHeader(0)
arr = np.empty((f.NumImages, int(d['Dim_2']), int(d['Dim_1'])))
for (i, ar) in enumerate(arr):
arr[i::] = f.GetData(i)
arr = self._slice_array(arr)
return arr
def getImage(self, index, full): file_with_path = self.path + '/' + self.data_files[index] img = EdfFile.EdfFile(file_with_path) roi = self.roi if full: im = img.GetData(0).astype(np.int64) - self.bg_combined_full else: im = img.GetData(0).astype(np.int64)[ roi[2]:roi[3], roi[0]:roi[1]] - self.bg_combined im = self.cleanImage(im) return im
def prepareFromFile(file_path, annotations):
edf_file = EdfFile.EdfFile(file_path, annotations)
epochs = edf_file.signals_list[1].getEpochs()
outputMatrix = edf_file.createOutput(epochs)
for e, o in zip(epochs, outputMatrix):
if max(o) == 0:
epochs.remove(e)
outputMatrix.remove(o)
inputMatrix = edf_file.createInput(epochs, True)
if (len(inputMatrix) != len(outputMatrix)):
print("shapes dont match in file %s. input: %d, output: %d" %
(file_path, len(inputMatrix), len(outputMatrix)))
return np.array([]), np.array([])
return inputMatrix, np.array(outputMatrix)
def writeImage(filename, data):
if filename.endswith('.edf') or filename.endswith('.tiff') :
edfFile = edf.EdfFile(filename, access='wb-')
edfFile.WriteImage({}, data)
else:
typeImage=data.dtype
if typeImage==np.uint8 :
scipy.misc.imsave(filename, data)
if typeImage==np.bool :
toStore=np.zeros(data.shape,dtype=np.uint8)
toStore[data==True]=255
scipy.misc.imsave(filename, toStore)
if typeImage==np.float32 or typeImage==np.float16 or typeImage==np.float64 :
scipy.misc.imsave(filename, data)
return im2D
def getDatasFromFile(filepath, fromIndex=0, toIndex=-1):
returnDatas = []
try:
f = EdfFile.EdfFile(filepath)
if toIndex < 0:
toIndex = f.GetNumImages()
for i in range(fromIndex, toIndex):
a = f.GetData(i)
header = f.GetHeader(i)
rData = Core.Processlib.Data()
rData.buffer = a
try:
rData.header = header
except TypeError:
pass
returnDatas.append(rData)
except:
import traceback
traceback.print_exc()
finally:
return returnDatas
def getHeader(self, filenumber):
file_with_path = self.path + '/' + self.data_files[filenumber]
img = EdfFile.EdfFile(file_with_path)
header = img.GetHeader(0)
mot_array = header['motor_mne'].split(' ')
motpos_array = header['motor_pos'].split(' ')
try:
det_array = header['counter_mne'].split(' ')
detpos_array = header['counter_pos'].split(' ')
except KeyError:
det_array = []
detpos_array = []
try:
srcur = float(header['machine current'].split(' ')[0])
except KeyError:
srcur = 0
return mot_array, motpos_array, det_array, detpos_array, srcur
def getFullHeader(self, filenumber):
file_with_path = self.path + '/' + self.data_files[filenumber]
img = EdfFile.EdfFile(file_with_path)
header = img.GetHeader(0)
metalist = []
for ind in header.keys():
if ind != 'motor_pos'\
and ind != 'motor_mne'\
and ind != 'counter_pos'\
and ind != 'counter_mne':
metalist.append(header[ind])
try:
metalist.extend(header['motor_pos'].split(' '))
metalist.extend(header['counter_pos'].split(' '))
except KeyError:
pass
return metalist
def getIndexList(self):
file_with_path = self.path + '/' + self.data_files[0]
img = EdfFile.EdfFile(file_with_path)
header = img.GetHeader(0)
indexlist = []
for ind in header.keys():
if ind != 'motor_pos'\
and ind != 'motor_mne'\
and ind != 'counter_pos'\
and ind != 'counter_mne':
indexlist.append(ind)
try:
indexlist.extend(header['motor_mne'].split(' '))
indexlist.extend(header['counter_mne'].split(' '))
except KeyError:
pass
self.indexlist = indexlist
def Iq(input_file, avg):
info = get_input(input_file)
qtot = qpattern(info)
geometry = info['geometry']
if geometry == 'gisaxs':
qtot = qtot[1]
wavelength = float(info['wavelength'])
distance = float(info['detector sample distance'])
detector = info['detector']
if detector == 'princeton' or detector == 'andor 22.5micron':
pix_size = 0.0225
if detector == 'medipix':
pix_size = 0.055
if detector == 'andor 13micron' or detector == 'andor':
pix_size = 0.013
if detector == 'xbpm':
pix_size = 0.001
deltaq = 4 * pi / wavelength * sin(arctan(2 * pix_size / distance) / 2)
mask_file = info['mask file']
tot = EdfFile.EdfFile(mask_file)
totmask = tot.GetData(0) + tot.GetData(1)
q = qtot[totmask == 0]
indq = argsort(q)
q = q[indq]
qr = arange(min(q), max(q) + deltaq, deltaq)
m = avg[totmask == 0]
m = m[indq]
lqv = len(qr)
radi = zeros((lqv, 2))
ini = 0
hh, bins = histogram(q, lqv, new=True)
radi[:, 0] = bins[:-1] + deltaq / 2
for i in xrange(lqv):
radi[i, 1] = mean(m[ini:ini + hh[i]])
ini = ini + hh[i]
return radi
def _get_from_file(self, image_nb):
for ref_data in self.ref_data:
values = self._get_filenames(ref_data, image_nb)
filename, path_in_file, image_index, file_format = values[0]
if file_format in ('EDF', 'EDFGZ', 'EDFConcat'):
if file_format == 'EDFConcat':
image_index = 0
if EdfFile is not None:
f = EdfFile.EdfFile(filename)
return f.GetData(image_index)
else:
raise RuntimeError("EdfFile module is not available, "
"cannot return image data.")
elif file_format == 'HDF5':
if h5py is not None:
with h5py.File(filename) as f:
dataset = f[path_in_file]
return dataset[image_index]
else:
raise RuntimeError("Format net yet managed")
else:
raise RuntimeError("Can't retrieved image %d from file" %
image_nb)
