This file is indexed.

/usr/share/pyshared/fabio/edfimage.py is in python-fabio 0.1.4-1.

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
 95
 96
 97
 98
 99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
505
506
507
508
509
510
511
512
513
514
515
516
517
518
519
520
521
522
523
524
525
526
527
528
529
530
531
532
533
534
535
536
537
538
539
540
541
542
543
544
545
546
547
548
549
550
551
552
553
554
555
556
557
558
559
560
561
562
563
564
565
566
567
568
569
570
571
572
573
574
575
576
577
578
579
580
581
582
583
584
585
586
587
588
589
590
591
592
593
594
595
596
597
598
599
600
601
602
603
604
605
606
607
608
609
610
611
612
613
614
615
616
617
618
619
620
621
622
623
624
625
626
627
628
629
630
631
632
633
634
635
636
637
638
639
640
641
642
643
644
645
646
647
648
649
650
651
652
653
654
655
656
657
658
659
660
661
662
663
664
665
666
667
668
669
670
671
672
673
674
675
676
677
678
679
680
681
682
683
684
685
686
687
688
689
690
691
692
693
694
695
696
697
698
699
700
701
702
703
704
705
706
707
708
709
710
711
712
713
714
715
716
717
718
719
720
721
722
723
724
725
726
727
728
729
730
731
732
733
734
735
736
737
738
739
740
741
742
743
744
745
746
747
748
749
750
751
752
753
754
755
756
757
758
759
760
761
762
763
764
765
766
767
768
769
770
771
772
773
774
775
776
777
778
779
780
781
782
783
784
785
786
787
788
789
790
791
792
793
794
795
796
797
798
799
800
801
802
803
804
805
806
807
808
809
810
811
812
813
814
815
816
817
818
819
820
821
822
823
824
825
826
827
828
829
830
831
832
833
834
835
836
837
838
839
840
841
842
843
844
845
846
847
848
849
850
851
852
853
854
855
856
857
858
859
860
861
862
863
864
865
866
867
868
869
870
871
872
873
874
875
876
877
878
879
880
881
882
883
884
885
886
887
888
889
890
891
892
893
894
895
896
897
898
899
900
901
902
903
904
905
906
907
908
909
910
911
912
913
914
915
916
917
918
919
920
921
922
923
924
925
926
927
928
929
930
931
932
933
934
#!/usr/bin/env python
# -*- coding: utf8 -*-
"""

License: GPLv2+

Authors:
........
* Henning O. Sorensen & Erik Knudsen:
  Center for Fundamental Research: Metal Structures in Four Dimensions;
  Risoe National Laboratory;
  Frederiksborgvej 399;
  DK-4000 Roskilde;
  email:erik.knudsen@risoe.dk
* Jon Wright & Jérôme Kieffer:
  European Synchrotron Radiation Facility;
  Grenoble (France)


"""
# get ready for python3
from __future__ import with_statement, print_function
import os, logging, types
logger = logging.getLogger("edfimage")
import numpy
from .fabioimage import fabioimage
from .fabioutils import isAscii, toAscii, nice_int
from .compression import decBzip2, decGzip, decZlib


BLOCKSIZE = 512
DATA_TYPES = {  "SignedByte"    :  numpy.int8,
                "Signed8"       :  numpy.int8,
                "UnsignedByte"  :  numpy.uint8,
                "Unsigned8"     :  numpy.uint8,
                "SignedShort"   :  numpy.int16,
                "Signed16"      :  numpy.int16,
                "UnsignedShort" :  numpy.uint16,
                "Unsigned16"    :  numpy.uint16,
                "UnsignedShortInteger" : numpy.uint16,
                "SignedInteger" :  numpy.int32,
                "Signed32"      :  numpy.int32,
                "UnsignedInteger":  numpy.uint32,
                "Unsigned32"    :  numpy.uint32,
                "SignedLong"    :  numpy.int32,
                "UnsignedLong"  :  numpy.uint32,
                "Signed64"      :  numpy.int64,
                "Unsigned64"    :  numpy.uint64,
                "FloatValue"    :  numpy.float32,
                "FLOATVALUE"    :  numpy.float32,
                "FLOAT"         :  numpy.float32, # fit2d
                "Float"         :  numpy.float32, # fit2d
                "FloatIEEE32"   :  numpy.float32,
                "Float32"       :  numpy.float32,
                "Double"        :  numpy.float64,
                "DoubleValue"   :  numpy.float64,
                "FloatIEEE64"   :  numpy.float64,
                "DoubleIEEE64"  :  numpy.float64}
try:
    DATA_TYPES["FloatIEEE128" ] =  numpy.float128
    DATA_TYPES["DoubleIEEE128" ] =  numpy.float128
    DATA_TYPES["QuadrupleValue" ] =  numpy.float128

except AttributeError:
    # not in your numpy
    pass

NUMPY_EDF_DTYPE = {"int8"       :"SignedByte",
                   "int16"      :"SignedShort",
                   "int32"      :"SignedInteger",
                   "int64"      :"Signed64",
                   "uint8"      :"UnsignedByte",
                   "uint16"     :"UnsignedShort",
                   "uint32"     :"UnsignedInteger",
                   "uint64"     :"Unsigned64",
                   "float32"    :"FloatValue",
                   "float64"    :"DoubleValue",
                   "float128"   :"QuadrupleValue",
             }

MINIMUM_KEYS = ['HEADERID',
                'IMAGE',
                'BYTEORDER',
                'DATATYPE',
                'DIM_1',
                'DIM_2',
                'SIZE'] # Size is thought to be essential for writing at least

DEFAULT_VALUES = {
                  # I do not define default values as they will be calculated at write time
                  # JK20110415
                  }

class Frame(object):
    """
    A class representing a single frame in an EDF file
    """
    def __init__(self, data=None, header=None, header_keys=None, number=None):
        if header is None:
            self.header = {}
        else:
            self.header = dict(header)

        if header_keys is None:
            self.header_keys = self.header.keys()
        else:
            self.header_keys = header_keys[:]
            for key in header_keys:
                if key not in self.header:
                    logger.warning("Header key %s, in header_keys is not in header dictionary, poping !!!" % key)
                    self.header_keys.remove(key)

        self.capsHeader = {}
        for key in self.header_keys:
            try:
                self.capsHeader[key.upper()] = key
            except AttributeError:
                logger.warning("Header key %s is not a string" % key)
        self._data = data
        self.dims = []
        self.dim1 = 0
        self.dim2 = 0
        self.start = None # Position of start of raw data in file
        self.size = None  # size of raw data in file
        self.file = None  # opened file object with locking capabilities !!!
        self.bpp = None
        self._bytecode = None
        if (number is not None) and isinstance(number, int):
            self.iFrame = number
        else:
            self.iFrame = 0

    def parseheader(self, block):
        """
        Parse the header in some EDF format from an already open file

        @param block: string representing the header block
        @type block: string, should be full ascii
        @return: size of the binary blob
        """
        #reset values ...
        self.header = {}
        self.capsHeader = {}
        self.header_keys = []
        self.size = None
        calcsize = 1
        self.dims = []

        for line in block.split(';'):
            if '=' in line:
                key, val = line.split('=' , 1)
                # Why would someone put null bytes in a header?
                key = key.replace("\x00"," ").strip()
                self.header[key] = val.replace("\x00"," ").strip()
                self.capsHeader[key.upper()] = key
                self.header_keys.append(key)

        # Compute image size
        if "SIZE" in self.capsHeader:
            try:
                self.size = nice_int(self.header[self.capsHeader["SIZE"]])
            except ValueError:
                logger.warning("Unable to convert to integer : %s %s " % (self.capsHeader["SIZE"], self.header[self.capsHeader["SIZE"]]))
        if "DIM_1" in self.capsHeader:
            try:
                dim1 = nice_int(self.header[self.capsHeader['DIM_1']])
            except ValueError:
                logger.error("Unable to convert to integer Dim_1: %s %s" % (self.capsHeader["DIM_1"], self.header[self.capsHeader["DIM_1"]]))
            else:
                calcsize *= dim1
                self.dims.append(dim1)
        else:
            logger.error("No Dim_1 in headers !!!")
        if "DIM_2" in self.capsHeader:
            try:
                dim2 = nice_int(self.header[self.capsHeader['DIM_2']])
            except ValueError:
                logger.error("Unable to convert to integer Dim_3: %s %s" % (self.capsHeader["DIM_2"], self.header[self.capsHeader["DIM_2"]]))
            else:
                calcsize *= dim2
                self.dims.append(dim2)
        else:
            logger.error("No Dim_2 in headers !!!")
        iDim = 3
        # JON: this appears to be for nD images, but we don't treat those
        while iDim is not None:
            strDim = "DIM_%i" % iDim
            if strDim in self.capsHeader:
                try:
                    dim3 = nice_int(self.header[self.capsHeader[strDim]])
                except ValueError:
                    logger.error("Unable to convert to integer %s: %s %s"
                                  % (strDim, self.capsHeader[strDim], self.header[self.capsHeader[strDim]]))
                    dim3 = None
                    iDim = None
                else:
                    if dim3 > 1:
                        # Otherwise treat dim3==1 as a 2D image
                        calcsize *= dim3
                        self.dims.append(dim3)
                    iDim += 1

            else:
                logger.debug("No Dim_3 -> it is a 2D image")
                iDim = None
        if self._bytecode is None:
            if "DATATYPE" in self.capsHeader:
                self._bytecode = DATA_TYPES[self.header[self.capsHeader['DATATYPE']]]
            else:
                self._bytecode = numpy.uint16
                logger.warning("Defaulting type to uint16")
        self.bpp = len(numpy.array(0, self._bytecode).tostring())
        calcsize *= self.bpp
        if (self.size is None):
            self.size = calcsize
        elif (self.size != calcsize):
            if ("COMPRESSION" in self.capsHeader) and (self.header[self.capsHeader['COMPRESSION']].upper().startswith("NO")):
                logger.info("Mismatch between the expected size %s and the calculated one %s" % (self.size, calcsize))
                self.size = calcsize

        for i, n in enumerate(self.dims):
            setattr(self, "dim%i" % (i + 1), n)

        return self.size


    def swap_needed(self):
        """
        Decide if we need to byteswap
        """
        if ('Low'  in self.header[self.capsHeader['BYTEORDER']] and numpy.little_endian) or \
           ('High' in self.header[self.capsHeader['BYTEORDER']] and not numpy.little_endian):
            return False
        if ('High'  in self.header[self.capsHeader['BYTEORDER']] and numpy.little_endian) or \
           ('Low' in self.header[self.capsHeader['BYTEORDER']] and not numpy.little_endian):
            if self.bpp in [2, 4, 8]:
                return True
            else:
                return False


    def getData(self):
        """
        Unpack a binary blob according to the specification given in the header

        @return: dataset as numpy.ndarray
        """
        data = None
        if self._data is not None:
            data = self._data
        elif self.file is None:
            data = self._data
        else:
            if self._bytecode is None:
                if "DATATYPE" in self.capsHeader:
                    self._bytecode = DATA_TYPES[self.header[self.capsHeader["DATATYPE"]]]
                else:
                    self._bytecode = numpy.uint16
            dims = self.dims[:]
            dims.reverse()
            with self.file.lock:
                if self.file.closed:
                    logger.error("file: %s from %s is closed. Cannot read data." % (self.file, self.file.filename))
                    return
                else:
                    self.file.seek(self.start)
                    fileData = self.file.read(self.size)

            if ("COMPRESSION" in self.capsHeader):
                compression = self.header[self.capsHeader["COMPRESSION"]].upper()
                uncompressed_size = self.bpp
                for i in dims:
                    uncompressed_size *= i
                if "OFFSET" in compression :
                    try:
                        import byte_offset#IGNORE:F0401
                    except ImportError, error:
                        logger.error("Unimplemented compression scheme:  %s (%s)" % (compression, error))
                    else:
                        myData = byte_offset.analyseCython(fileData, size=uncompressed_size)
                        rawData = myData.astype(self._bytecode).tostring()
                        self.size = uncompressed_size
                elif compression == "NONE":
                    rawData = fileData
                elif "GZIP" in compression:
                    rawData = decGzip(fileData)
                    self.size = uncompressed_size
                elif "BZ" in compression :
                    rawData = decBzip2(fileData)
                    self.size = uncompressed_size
                elif "Z" in compression :
                    rawData = decZlib(fileData)
                    self.size = uncompressed_size
                else:
                    logger.warning("Unknown compression scheme %s" % compression)
                    rawData = fileData

            else:
                rawData = fileData

            expected = self.size
            obtained = len(rawData)
            if expected > obtained:
                logger.error("Data stream is incomplete: %s < expected %s bytes" % (obtained, expected))
                rawData += "\x00" * (expected - obtained)
            elif expected < len(rawData):
                logger.info("Data stream contains trailing junk : %s > expected %s bytes" % (obtained, expected))
                rawData = rawData[:expected]
            if self.swap_needed():
                data = numpy.fromstring(rawData, self._bytecode).byteswap().reshape(tuple(dims))
            else:
                data = numpy.fromstring(rawData, self._bytecode).reshape(tuple(dims))
            self._data = data
            self._bytecode = data.dtype.type
        return data
    def setData(self, npa=None):
        """Setter for data in edf frame"""
        self._data = npa
    data = property(getData, setData, "property: (edf)frame.data, uncompress the datablock when needed")
    def getByteCode(self):
        if self._bytecode is None:
            self._bytecode = self.data.dtype.type

        return self._bytecode
    def setByteCode(self, _iVal):
        self._bytecode = _iVal
    bytecode = property(getByteCode, setByteCode)

    def getEdfBlock(self, force_type=None, fit2dMode=False):
        """
        @param force_type: type of the dataset to be enforced like "float64" or "uint16"
        @type force_type: string or numpy.dtype
        @param fit2dMode: enforce compatibility with fit2d and starts countimg number of images at 1
        @type fit2dMode: boolean
        @return: ascii header block
        @rtype: python string with the concatenation of the ascii header and the binary data block
        """
        if force_type is not None:
            data = self.data.astype(force_type)
        else:
            data = self.data
        fit2dMode = bool(fit2dMode)
        for key in self.header:
            KEY = key.upper()
            if KEY not in self.capsHeader:
                self.capsHeader[KEY] = key
            if key not in self.header_keys:
                self.header_keys.append(key)

        header = self.header.copy()
        header_keys = self.header_keys[:]
        capsHeader = self.capsHeader.copy()

        listHeader = ["{\n"]
#        First of all clean up the headers:
        for i in capsHeader:
            if "DIM_" in i:
                header.pop(capsHeader[i])
                header_keys.remove(capsHeader[i])
        for KEY in ["SIZE", "EDF_BINARYSIZE", "EDF_HEADERSIZE", "BYTEORDER", "DATATYPE", "HEADERID", "IMAGE"]:
            if KEY in capsHeader:
                header.pop(capsHeader[KEY])
                header_keys.remove(capsHeader[KEY])
        if "EDF_DATABLOCKID" in capsHeader:
            header_keys.remove(capsHeader["EDF_DATABLOCKID"])
            #but do not remove the value from dict, instead reset the key ...
            if capsHeader["EDF_DATABLOCKID"] != "EDF_DataBlockID":
                header["EDF_DataBlockID"] = header.pop(capsHeader["EDF_DATABLOCKID"])
                capsHeader["EDF_DATABLOCKID"] = "EDF_DataBlockID"

#            Then update static headers freshly deleted
        header_keys.insert(0, "Size")
        header["Size"] = len(data.tostring())
        header_keys.insert(0, "HeaderID")
        header["HeaderID"] = "EH:%06d:000000:000000" % (self.iFrame + fit2dMode)
        header_keys.insert(0, "Image")
        header["Image"] = str(self.iFrame + fit2dMode)

        dims = list(data.shape)
        nbdim = len(dims)
        for i in dims:
            key = "Dim_%i" % nbdim
            header[key] = i
            header_keys.insert(0, key)
            nbdim -= 1
        header_keys.insert(0, "DataType")
        header["DataType"] = NUMPY_EDF_DTYPE[str(numpy.dtype(data.dtype))]
        header_keys.insert(0, "ByteOrder")
        if numpy.little_endian:
            header["ByteOrder"] = "LowByteFirst"
        else:
            header["ByteOrder"] = "HighByteFirst"
        approxHeaderSize = 100
        for key in header:
            approxHeaderSize += 7 + len(key) + len(str(header[key]))
        approxHeaderSize = BLOCKSIZE * (approxHeaderSize // BLOCKSIZE + 1)
        header_keys.insert(0, "EDF_HeaderSize")
        header["EDF_HeaderSize"] = "%5s" % (approxHeaderSize)
        header_keys.insert(0, "EDF_BinarySize")
        header["EDF_BinarySize"] = len(data.tostring())
        header_keys.insert(0, "EDF_DataBlockID")
        if not "EDF_DataBlockID" in header:
            header["EDF_DataBlockID"] = "%i.Image.Psd" % (self.iFrame + fit2dMode)
        preciseSize = 4 #2 before {\n 2 after }\n
        for key in header_keys:
            #Escape keys or values that are no ascii
            strKey = str(key)
            if not isAscii(strKey, listExcluded=["}", "{"]):
                logger.warning("Non ascii key %s, skipping" % strKey)
                continue
            strValue = str(header[key])
            if not isAscii(strValue, listExcluded=["}", "{"]):
                logger.warning("Non ascii value %s, skipping" % strValue)
                continue
            line = strKey + " = " + strValue + " ;\n"
            preciseSize += len(line)
            listHeader.append(line)
        if preciseSize > approxHeaderSize:
            logger.error("I expected the header block only at %s in fact it is %s" % (approxHeaderSize, preciseSize))
            for  idx, line in enumerate(listHeader[:]):
                if line.startswith("EDF_HeaderSize"):
                    headerSize = BLOCKSIZE * (preciseSize // BLOCKSIZE + 1)
                    newline = "EDF_HeaderSize = %5s ;\n" % headerSize
                    delta = len(newline) - len(line)
                    if (preciseSize // BLOCKSIZE) != ((preciseSize + delta) // BLOCKSIZE):
                        headerSize = BLOCKSIZE * ((preciseSize + delta) // BLOCKSIZE + 1)
                        newline = "EDF_HeaderSize = %5s ;\n" % headerSize
                    preciseSize = preciseSize + delta
                    listHeader[idx] = newline
                    break
        else:
            headerSize = approxHeaderSize
        listHeader.append(" "*(headerSize - preciseSize) + "}\n")
        return "".join(listHeader) + data.tostring()



class edfimage(fabioimage):
    """ Read and try to write the ESRF edf data format """

    def __init__(self, data=None , header=None, header_keys=None, frames=None):
        self.currentframe = 0
        self.filesize = None
        try:
            dim = len(data.shape)
        except Exception, error: #IGNORE:W0703
            logger.debug("Data don't look like a numpy array (%s), resetting all!!" % error)
            data = None
            dim = 0
            fabioimage.__init__(self, data, header)
        if dim == 2:
            fabioimage.__init__(self, data, header)
        elif dim == 1 :
            data.shape = (0, len(data))
            fabioimage.__init__(self, data, header)
        elif dim == 3 :
            fabioimage.__init__(self, data[0, :, :], header)
        elif dim == 4 :
            fabioimage.__init__(self, data[0, 0, :, :], header)
        elif dim == 5 :
            fabioimage.__init__(self, data[0, 0, 0, :, :], header)

        if frames is None:
            frame = Frame(data=self.data, header=self.header,
                          header_keys=header_keys ,
                          number=self.currentframe)
            self.__frames = [frame]
        else:
            self.__frames = frames

    @staticmethod
    def checkHeader(header=None):
        """
        Empty for fabioimage but may be populated by others classes
        """
        if type(header) != types.DictionaryType:
            return {}
        new = {}
        for key, value in header.items():
            new[toAscii(key, ";{}")] = toAscii(value, ";{}")
        return new

    @staticmethod
    def _readHeaderBlock(infile):
        """
        Read in a header in some EDF format from an already open file

        @param infile: file object open in read mode
        @return: string (or None if no header was found.
        """
        MAX_HEADER_SIZE = BLOCKSIZE * 20
        block = infile.read(BLOCKSIZE)
        if len(block) < BLOCKSIZE:
            logger.debug("Under-short header: only %i bytes in %s" % (len(block), infile.name))
            return
        if (block.find("{") < 0) :
            # This does not look like an edf file
            logger.warning("no opening {. Corrupt header of EDF file %s" % infile.name)
            return
        if "EDF_HeaderSize" in block:
            start = block.index("EDF_HeaderSize")
            chunk = block[start:].split("=")[1].strip()
            try:
                new_max_header_size = int(chunk.split(";")[0].strip())
            except Exception:
                logger.warning("Unable to read header size in %s" % chunk)
            else:
                if new_max_header_size > MAX_HEADER_SIZE:
                    logger.info("Redefining MAX_HEADER_SIZE to %s" % new_max_header_size)
                    MAX_HEADER_SIZE = new_max_header_size
        while '}' not in block:
            block = block + infile.read(BLOCKSIZE)
            if len(block) > MAX_HEADER_SIZE:
                logger.warning("Runaway header in EDF file MAX_HEADER_SIZE: %s \n%s" % (MAX_HEADER_SIZE, block))
                return
        start = block.find("{") + 1
        end = block.find("}")

        # Now it is essential to go to the start of the binary part
        if block[end: end + 3] == "}\r\n":
            offset = end + 3 - len(block)
        elif block[end: end + 2] == "}\n":
            offset = end + 2 - len(block)
        else:
            logger.error("Unable to locate start of the binary section")
            offset = None
        if offset is not None:
            infile.seek(offset, os.SEEK_CUR)
        return block[start:end]


    def _readheader(self, infile):
        """
        Read all headers in a file and populate self.header
        data is not yet populated
        @type infile: file object open in read mode
        """
        self.__frames = []
        bContinue = True
        while bContinue:
            block = self._readHeaderBlock(infile)
            if block is None:
                bContinue = False
                break
            frame = Frame(number=self.nframes)
            size = frame.parseheader(block)
            frame.file = infile
            frame.start = infile.tell()
            frame.size = size
            self.__frames += [frame]
            try:
                infile.seek(size, os.SEEK_CUR)
            except Exception, error:
                logger.warning("infile is %s" % infile)
                logger.warning("Position is %s" % infile.tell())
                logger.warning("size is %s" % size)
                logger.error("It seams this error occurs under windows when reading a (large-) file over network: %s ", error)
                raise Exception(error)

            if  frame.start + size > infile.size:
                logger.warning("Non complete datablock: got %s, expected %s" % (infile.size - frame.start, size))
                bContinue = False
                break

        for i, frame in enumerate(self.__frames):
            missing = []
            for item in MINIMUM_KEYS:
                if item not in frame.capsHeader:
                    missing.append(item)
            if len(missing) > 0:
                logger.info("EDF file %s frame %i misses mandatory keys: %s " % (self.filename, i, " ".join(missing)))
        self.currentframe = 0


    def read(self, fname, frame=None):
        """
        Read in header into self.header and
            the data   into self.data
        """
        self.resetvals()
        self.filename = fname

        infile = self._open(fname, "rb")
        self._readheader(infile)
        if frame is None:
            pass
        elif frame < self.nframes:
            self = self.getframe(frame)
        else:
            logger.error("Reading file %s You requested frame %s but only %s frames are available", fname, frame, self.nframes)
        self.resetvals()
        # ensure the PIL image is reset
        self.pilimage = None
        return self

    def swap_needed(self):
        """
        Decide if we need to byteswap
        """
        if ('Low'  in self.header[self.capsHeader['BYTEORDER']] and numpy.little_endian) or \
           ('High' in self.header[self.capsHeader['BYTEORDER']] and not numpy.little_endian):
            return False
        if ('High'  in self.header[self.capsHeader['BYTEORDER']] and numpy.little_endian) or \
           ('Low' in self.header[self.capsHeader['BYTEORDER']] and not numpy.little_endian):
            if self.bpp in [2, 4, 8]:
                return True
            else:
                return False

    def unpack(self):
        """
        Unpack a binary blob according to the specification given in the header and return the dataset

        @return: dataset as numpy.ndarray
        """
        return self.__frames[self.currentframe].getData()


    def getframe(self, num):
        """ returns the file numbered 'num' in the series as a fabioimage """
        newImage = None
        if self.nframes == 1:
            logger.debug("Single frame EDF; having fabioimage default behavour: %s" % num)
            newImage = fabioimage.getframe(self, num)
        elif num in xrange(self.nframes):
            logger.debug("Multi frame EDF; having edfimage specific behavour: %s/%s" % (num, self.nframes))
            newImage = edfimage(frames=self.__frames)
            newImage.currentframe = num
            newImage.filename = self.filename
        else:
            txt = "Cannot access frame: %s/%s" % (num, self.nframes)
            logger.error(txt)
            raise ValueError("edfimage.getframe:" + txt)
        return newImage


    def previous(self):
        """ returns the previous file in the series as a fabioimage """
        newImage = None
        if self.nframes == 1:
            newImage = fabioimage.previous(self)
        else:
            newFrameId = self.currentframe - 1
            newImage = self.getframe(newFrameId)
        return newImage


    def next(self):
        """ returns the next file in the series as a fabioimage """
        newImage = None
        if self.nframes == 1:
            newImage = fabioimage.next(self)
        else:
            newFrameId = self.currentframe + 1
            newImage = self.getframe(newFrameId)
        return newImage


    def write(self, fname, force_type=None, fit2dMode=False):
        """
        Try to write a file
        check we can write zipped also
        mimics that fabian was writing uint16 (we sometimes want floats)

        @param force_type: can be numpy.uint16 or simply "float"
        @return: None

        """

        with self._open(fname, mode="wb") as outfile:
            for i, frame in enumerate(self.__frames):
                frame.iFrame = i
                outfile.write(frame.getEdfBlock(force_type=force_type, fit2dMode=fit2dMode))


    def appendFrame(self, frame=None, data=None, header=None):
        """
        Method used add a frame to an EDF file
        @param frame: frame to append to edf image
        @type frame: instance of Frame
        @return: None
        """
        if isinstance(frame, Frame):
            self.__frames.append(frame)
        else:
            self.__frames.append(Frame(data, header))


    def deleteFrame(self, frameNb=None):
        """
        Method used to remove a frame from an EDF image. by default the last one is removed.
        @param frameNb: frame number to remove, by  default the last.
        @type frameNb: integer
        @return: None
        """
        if frameNb is None:
            self.__frames.pop()
        else:
            self.__frames.pop(frameNb)

    def fastReadData(self, filename=None):
        """
        This is a special method that will read and return the data from another file ...
        The aim is performances, ... but only supports uncompressed files. 
         
        @return: data from another file using positions from current edfimage
        """
        if (filename is None) or not os.path.isfile(filename):
            raise RuntimeError("edfimage.fastReadData is only valid with another file: %s does not exist" % (filename))
        data = None
        frame = self.__frames[self.currentframe]
        with open(filename, "rb")as f:
            f.seek(frame.start)
            raw = f.read(frame.size)
        try:
            data = numpy.fromstring(raw, dtype=self.bytecode)
            data.shape = self.data.shape
        except Exception, err :
            logger.error("unable to convert file content to numpy array: %s", err)
        return data

    def fastReadROI(self, filename, coords=None):
        """
        Method reading Region of Interest of another file  based on metadata available in current edfimage.
        The aim is performances, ... but only supports uncompressed files.
        
        @return: ROI-data from another file using positions from current edfimage
        @rtype: numpy 2darray
        """
        if (filename is None) or not os.path.isfile(filename):
            raise RuntimeError("edfimage.fastReadData is only valid with another file: %s does not exist" % (filename))
        data = None
        frame = self.__frames[self.currentframe]

        if len(coords) == 4:
            slice1 = self.make_slice(coords)
        elif len(coords) == 2 and isinstance(coords[0], slice) and \
                                  isinstance(coords[1], slice):
            slice1 = coords
        else:
            logger.warning('readROI: Unable to understand Region Of Interest: got %s', coords)
            return
        d1 = self.data.shape[-1]
        start0 = slice1[0].start
        start1 = slice1[1].start
        slice2 = (slice(0, slice1[0].stop - start0, slice1[0].step),
                   slice(0, slice1[1].stop - start1, slice1[1].step))
        start = frame.start + self.bpp * (d1 * start0 + start1)
        size = self.bpp * ((slice2[0].stop) * d1)
        with open(filename, "rb")as f:
            f.seek(start)
            raw = f.read(size)
        try:
            data = numpy.fromstring(raw, dtype=self.bytecode)
            data.shape = -1, d1
        except Exception, err :
            logger.error("unable to convert file content to numpy array: %s", err)
        return data[slice2]


################################################################################
# Properties definition for header, data, header_keys and capsHeader
################################################################################
    def getNbFrames(self):
        """
        Getter for number of frames
        """
        return len(self.__frames)
    def setNbFrames(self, val):
        """
        Setter for number of frames ... should do nothing. Here just to avoid bugs
        """
        if val != len(self.__frames):
            logger.warning("trying to set the number of frames ")
    nframes = property(getNbFrames, setNbFrames, "property: number of frames in EDF file")


    def getHeader(self):
        """
        Getter for the headers. used by the property header,
        """
        return self.__frames[self.currentframe].header
    def setHeader(self, _dictHeader):
        """
        Enforces the propagation of the header to the list of frames
        """
        try:
            self.__frames[self.currentframe].header = _dictHeader
        except AttributeError:
            self.__frames = [Frame(header=_dictHeader)]
        except IndexError:
            if self.currentframe < len(self.__frames):
                self.__frames.append(Frame(header=_dictHeader))
    def delHeader(self):
        """
        Deleter for edf header
        """
        self.__frames[self.currentframe].header = {}
    header = property(getHeader, setHeader, delHeader, "property: header of EDF file")

    def getHeaderKeys(self):
        """
        Getter for edf header_keys
        """
        return self.__frames[self.currentframe].header_keys
    def setHeaderKeys(self, _listtHeader):
        """
        Enforces the propagation of the header_keys to the list of frames
        @param _listtHeader: list of the (ordered) keys in the header
        @type _listtHeader: python list
        """
        try:
            self.__frames[self.currentframe].header_keys = _listtHeader
        except AttributeError:
            self.__frames = [Frame(header_keys=_listtHeader)]
        except IndexError:
            if self.currentframe < len(self.__frames):
                self.__frames.append(Frame(header_keys=_listtHeader))
    def delHeaderKeys(self):
        """
        Deleter for edf header_keys
        """
        self.__frames[self.currentframe].header_keys = []
    header_keys = property(getHeaderKeys, setHeaderKeys, delHeaderKeys, "property: header_keys of EDF file")

    def getData(self):
        """
        getter for edf Data
        @return: data for current frame
        @rtype: numpy.ndarray
        """
        npaData = None
        try:
            npaData = self.__frames[self.currentframe].data
        except AttributeError:
            self.__frames = [Frame()]
            npaData = self.__frames[self.currentframe].data
        except IndexError:
            if self.currentframe < len(self.__frames):
                self.__frames.append(Frame())
                npaData = self.__frames[self.currentframe].data
        return npaData

    def setData(self, _data):
        """
        Enforces the propagation of the data to the list of frames
        @param _data: numpy array representing data
        """
        try:
            self.__frames[self.currentframe].data = _data
        except AttributeError:
            self.__frames = [Frame(data=_data)]
        except IndexError:
            if self.currentframe < len(self.__frames):
                self.__frames.append(Frame(data=_data))
    def delData(self):
        """
        deleter for edf Data
        """
        self.__frames[self.currentframe].data = None
    data = property(getData, setData, delData, "property: data of EDF file")

    def getCapsHeader(self):
        """
        getter for edf headers keys in upper case
        @return: data for current frame
        @rtype: dict
        """
        return self.__frames[self.currentframe].capsHeader
    def setCapsHeader(self, _data):
        """
        Enforces the propagation of the header_keys to the list of frames
        @param _data: numpy array representing data
        """
        self.__frames[self.currentframe].capsHeader = _data
    def delCapsHeader(self):
        """
        deleter for edf capsHeader
        """
        self.__frames[self.currentframe].capsHeader = {}
    capsHeader = property(getCapsHeader, setCapsHeader, delCapsHeader, "property: capsHeader of EDF file, i.e. the keys of the header in UPPER case.")

    def getDim1(self):
        return self.__frames[self.currentframe].dim1
    def setDim1(self, _iVal):
        try:
            self.__frames[self.currentframe].dim1 = _iVal
        except AttributeError:
            self.__frames = [Frame()]
        except IndexError:
            if self.currentframe < len(self.__frames):
                self.__frames.append(Frame())
                self.__frames[self.currentframe].dim1 = _iVal
    dim1 = property(getDim1, setDim1)
    def getDim2(self):
        return self.__frames[self.currentframe].dim2
    def setDim2(self, _iVal):
        try:
            self.__frames[self.currentframe].dim2 = _iVal
        except AttributeError:
            self.__frames = [Frame()]
        except IndexError:
            if self.currentframe < len(self.__frames):
                self.__frames.append(Frame())
                self.__frames[self.currentframe].dim2 = _iVal
    dim2 = property(getDim2, setDim2)

    def getDims(self):
        return self.__frames[self.currentframe].dims
    dims = property(getDims)
    def getByteCode(self):
        return self.__frames[self.currentframe].bytecode
    def setByteCode(self, _iVal):
        try:
            self.__frames[self.currentframe].bytecode = _iVal
        except AttributeError:
            self.__frames = [Frame()]
        except IndexError:
            if self.currentframe < len(self.__frames):
                self.__frames.append(Frame())
                self.__frames[self.currentframe].bytecode = _iVal
    bytecode = property(getByteCode, setByteCode)
    def getBpp(self):
        return self.__frames[self.currentframe].bpp
    def setBpp(self, _iVal):
        try:
            self.__frames[self.currentframe].bpp = _iVal
        except AttributeError:
            self.__frames = [Frame()]
        except IndexError:
            if self.currentframe < len(self.__frames):
                self.__frames.append(Frame())
                self.__frames[self.currentframe].bpp = _iVal
    bpp = property(getBpp, setBpp)