This file is indexed.

/usr/share/pyshared/fabio/fabioimage.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
 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
#!/usr/bin/env python 

"""

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
         
         and Jon Wright, Jerome Kieffer: ESRF

"""

import numpy, os, gzip, bz2, StringIO
import Image
import fabioutils


class fabioStream(StringIO.StringIO):
    """ 
    just an interface providing the name and mode property to a StringIO
    
    BugFix for MacOSX
    """
    def __init__(self, data, fname=None, mode="r"):
        StringIO.StringIO.__init__(self, data)
        self.closed = False
        if fname == None:
            self.name = "fabioStream"
        else:
            self.name = fname
        self.mode = mode


class fabioimage(object):
    """
    A common object for images in fable
    Contains a numpy array (.data) and dict of meta data (.header)
    """

    _need_a_seek_to_read = False
    _need_a_real_file = False

    def __init__(self, data=None , header=None):
        """
        Set up initial values
        """
        if type(data) == type("string"):
            raise Exception("fabioimage.__init__ bad argument - " + \
                            "data should be numpy array")
        self.data = data
        self.pilimage = None
        if header is None:
            self.header = {}
        else:
            self.header = header
        self.header_keys = self.header.keys() # holds key ordering
        if data is not None:
            self.dim1, self.dim2 = data.shape
        else:
            self.dim1 = self.dim2 = 0
        self.bytecode = None     # numpy typecode
        self.bpp = 2             # bytes per pixel
        # cache for image statistics
        self.mean = self.maxval = self.stddev = self.minval = None
        # Cache roi
        self.area_sum = None
        self.slice = None
        # New for multiframe files
        self.nframes = 1
        self.currentframe = 0
        self.filename = None

    def getframe(self, num):
        """ returns the file numbered 'num' in the series as a fabioimage """
        if self.nframes == 1:
            # single image per file
            import openimage
            return openimage.openimage(
                fabioutils.jump_filename(self.filename, num))
        raise Exception("getframe out of range")

    def previous(self):
        """ returns the previous file in the series as a fabioimage """
        import openimage
        return openimage.openimage(
            fabioutils.previous_filename(self.filename))

    def next(self):
        """ returns the next file in the series as a fabioimage """
        import openimage
        return openimage.openimage(
            fabioutils.next_filename(self.filename))

    def toPIL16(self, filename=None):
        """
        Convert to Python Imaging Library 16 bit greyscale image

        FIXME - this should be handled by the libraries now
        """
        if filename:
            self.read(filename)
        if self.pilimage is not None:
            return self.pilimage
        # mode map
        size = self.data.shape[:2][::-1]
        typmap = {
            'float32' : "F"     ,
            'int32'   : "F;32S" ,
            'uint32'  : "F;32"  ,
            'int16'   : "F;16S" ,
            'uint16'  : "F;16"  ,
            'int8'    : "F;8S"  ,
            'uint8'   : "F;8"  }
        if typmap.has_key(self.data.dtype.name):
            mode2 = typmap[ self.data.dtype.name ]
            mode1 = mode2[0]
        else:
            raise Exception("Unknown numpy type " + str(self.data.dtype.type))
        # 
        # hack for byteswapping for PIL in MacOS
        testval = numpy.array((1, 0), numpy.uint8).view(numpy.uint16)[0]
        if  testval == 1:
            dats = self.data.tostring()
        elif testval == 256:
            dats = self.data.byteswap().tostring()
        else:
            raise Exception("Endian unknown in fabioimage.toPIL16")

        self.pilimage = Image.frombuffer(mode1,
                                         size,
                                         dats,
                                         "raw",
                                         mode2,
                                         0,
                                         1)

        return self.pilimage

    def getheader(self):
        """ returns self.header """
        return self.header

    def getmax(self):
        """ Find max value in self.data, caching for the future """
        if self.maxval is None:
            self.maxval = self.data.max()
        return self.maxval

    def getmin(self):
        """ Find min value in self.data, caching for the future """
        if self.minval is None:
            self.minval = self.data.min()
        return self.minval

    def make_slice(self, coords):
        """
        Convert a len(4) set of coords into a len(2) 
        tuple (pair) of slice objects
        the latter are immutable, meaning the roi can be cached
        """
        assert len(coords) == 4
        if len(coords) == 4:
            # fabian edfimage preference
            if coords[0] > coords[2]:
                coords[0:3:2] = [coords[2], coords[0]]
            if coords[1] > coords[3]:
                coords[1:4:2] = [coords[3], coords[1]]
            #in fabian: normally coordinates are given as (x,y) whereas 
            # a matrix is given as row,col 
            # also the (for whichever reason) the image is flipped upside 
            # down wrt to the matrix hence these tranformations
            fixme = (self.dim2 - coords[3] - 1,
                     coords[0] ,
                     self.dim2 - coords[1] - 1,
                     coords[2])
        return (slice(int(fixme[0]), int(fixme[2]) + 1) ,
                 slice(int(fixme[1]), int(fixme[3]) + 1))


    def integrate_area(self, coords):
        """ 
        Sums up a region of interest 
        if len(coords) == 4 -> convert coords to slices
        if len(coords) == 2 -> use as slices
        floor -> ? removed as unused in the function.
        """
        if self.data == None:
            # This should return NAN, not zero ?
            return 0
        if len(coords) == 4:
            sli = self.make_slice(coords)
        elif len(coords) == 2 and isinstance(coords[0], slice) and \
                                  isinstance(coords[1], slice):
            sli = coords
        if sli == self.slice and self.area_sum is not None:
            return self.area_sum
        self.slice = sli
        self.area_sum = numpy.sum(
                            numpy.ravel(
                                self.data[ self.slice ].astype(numpy.float)))
        return self.area_sum

    def getmean(self):
        """ return the mean """
        if self.mean is None:
            self.mean = self.data.mean(dtype=numpy.double)
        return self.mean

    def getstddev(self):
        """ return the standard deviation """
        if self.stddev == None:
            self.stddev = self.data.std(dtype=numpy.double)
        return self.stddev

    def add(self, other):
        """
        Add another Image - warnign, does not clip to 16 bit images by default
        """
        if not hasattr(other, 'data'):
            print 'edfimage.add() called with something that ' + \
                'does not have a data field'
        assert self.data.shape == other.data.shape , \
                  'incompatible images - Do they have the same size?'
        self.data = self.data + other.data
        self.resetvals()


    def resetvals(self):
        """ Reset cache - call on changing data """
        self.mean = self.stddev = self.maxval = self.minval = None
        self.area_sum = None

    def rebin(self, x_rebin_fact, y_rebin_fact, keep_I=True):
        """ 
        Rebin the data and adjust dims 
        @param x_rebin_fact: x binning factor
        @param y_rebin_fact: y binning factor
        @param keep_I: shall the signal increase ?
        @type x_rebin_fact: int
        @type y_rebin_fact: int
        @type keep_I: boolean

        
        """
        if self.data == None:
            raise Exception('Please read in the file you wish to rebin first')

        if (self.dim1 % x_rebin_fact != 0) or (self.dim2 % y_rebin_fact != 0):
            raise RuntimeError('image size is not divisible by rebin factor - ' + \
                  'skipping rebin')
        else:
            dataIn = self.data.astype("float64")
            shapeIn = self.data.shape
            shapeOut = (shapeIn[0] / y_rebin_fact, shapeIn[1] / x_rebin_fact)
            binsize = y_rebin_fact * x_rebin_fact
            if binsize < 50: #method faster for small binning (4x4)
                out = numpy.zeros(shapeOut, dtype="float64")
                for j in range(x_rebin_fact):
                    for i in range(y_rebin_fact):
                        out += dataIn[i::y_rebin_fact, j::x_rebin_fact]
            else: #method faster for large binning (8x8)
                temp = self.data.astype("float64")
                temp.shape = (shapeOut[0], y_rebin_fact, shapeOut[1], x_rebin_fact)
                out = temp.sum(axis=3).sum(axis=1)
        self.resetvals()
        if keep_I:
            self.data = (out / (y_rebin_fact * x_rebin_fact)).astype(self.data.dtype)
        else:
            self.data = out.astype(self.data.dtype)

        self.dim1 = self.dim1 / x_rebin_fact
        self.dim2 = self.dim2 / y_rebin_fact

        #update header
        self.update_header()

    def write(self, fname):
        """
        To be overwritten - write the file
        """
        raise Exception("Class has not implemented readheader method yet")

    def readheader(self, filename):
        """
        Call the _readheader function...
        """
        # Override the needs asserting that all headers can be read via python modules
        save_state = self._need_a_real_file , self._need_a_seek_to_read
        self._need_a_real_file , self._need_a_seek_to_read = False, False
        fin = self._open(filename)
        self._readheader(fin)
        fin.close()
        self._need_a_real_file , self._need_a_seek_to_read = save_state

    def _readheader(self, fik_obj):
        """
        Must be overridden in classes
        """
        raise Exception("Class has not implemented _readheader method yet")

    def update_header(self , **kwds):
        """
        update the header entries
        by default pass in a dict of key, values.
        """
        self.header.update(kwds)

    def read(self, filename):
        """
        To be overridden - fill in self.header and self.data
        """
        raise Exception("Class has not implemented read method yet")


    def _open(self, fname, mode="rb"):
        """
        Try to handle compressed files, streams, shared memory etc
        Return an object which can be used for "read" and "write" 
        ... FIXME - what about seek ? 
        """
        fileObject = None
        self.filename = fname
        if hasattr(fname, "read") and hasattr(fname, "write"):
            # It is already something we can use
            return fname
        if isinstance(fname, (str, unicode)):
            self.header["filename"] = fname
            if os.path.splitext(fname)[1] == ".gz":
                fileObject = self._compressed_stream(fname,
                                       fabioutils.COMPRESSORS['.gz'],
                                       gzip.GzipFile,
                                       mode)
            elif os.path.splitext(fname)[1] == '.bz2':
                fileObject = self._compressed_stream(fname,
                                       fabioutils.COMPRESSORS['.bz2'],
                                       bz2.BZ2File,
                                       mode)
            #
            # Here we return the file even though it may be bzipped or gzipped
            # but named incorrectly...
            #
            # FIXME - should we fix that or complain about the daft naming?
            else:
                fileObject = open(fname, mode)
            if "name" not in dir(fileObject):
                    fileObject.name = fname

        return fileObject

    def _compressed_stream(self,
                           fname,
                           system_uncompress,
                           python_uncompress,
                           mode='rb'):
        """
        Try to transparently handle gzip / bzip without always getting python 
        performance
        """
        # assert that python modules are always OK based on performance benchmark
        # Try to fix the way we are using them?
        fobj = None
        if self._need_a_real_file and mode[0] == "r":
            fo = python_uncompress(fname, mode)
            fobj = os.tmpfile()
            fobj.write(fo.read())
            fo.close()
            fobj.seek(0)
        elif self._need_a_seek_to_read and mode[0] == "r":
            fo = python_uncompress(fname, mode)
            fobj = fabioStream(fo.read(), fname, mode)
        else:
            fobj = python_uncompress(fname, mode)
        return fobj


def test():
    """
    check some basic fabioimage functionality
    """
    import time
    start = time.time()

    dat = numpy.ones((1024, 1024), numpy.uint16)
    dat = (dat * 50000).astype(numpy.uint16)
    assert dat.dtype.char == numpy.ones((1), numpy.uint16).dtype.char
    hed = {"Title":"50000 everywhere"}
    obj = fabioimage(dat, hed)

    assert obj.getmax() == 50000
    assert obj.getmin() == 50000
    assert obj.getmean() == 50000 , obj.getmean()
    assert obj.getstddev() == 0.

    dat2 = numpy.zeros((1024, 1024), numpy.uint16, savespace=1)
    cord = [ 256, 256, 790, 768 ]
    slic = obj.make_slice(cord)
    dat2[slic] = dat2[slic] + 100

    obj = fabioimage(dat2, hed)

    # New object, so...
    assert obj.maxval is None
    assert obj.minval is None

    assert obj.getmax() == 100, obj.getmax()
    assert obj.getmin() == 0 , obj.getmin()
    npix = (slic[0].stop - slic[0].start) * (slic[1].stop - slic[1].start)
    obj.resetvals()
    area1 = obj.integrate_area(cord)
    obj.resetvals()
    area2 = obj.integrate_area(slic)
    assert area1 == area2
    assert obj.integrate_area(cord) == obj.integrate_area(slic)
    assert obj.integrate_area(cord) == npix * 100, obj.integrate_area(cord)


    def clean():
        """ clean up the created testfiles"""
        for name in ["testfile", "testfile.gz", "testfile.bz2"]:
            try:
                os.remove(name)
            except:
                continue


    clean()

    gzip.open("testfile.gz", "wb").write("{ hello }")
    fout = obj._open("testfile.gz")
    readin = fout.read()
    assert readin == "{ hello }", readin + " gzipped file"


    bz2.BZ2File("testfilebz", "wb").write("{ hello }")
    fout = obj._open("testfile.bz2")
    readin = fout.read()
    assert readin == "{ hello }", readin + " bzipped file"

    ftest = open("testfile", "wb")
    ftest.write("{ hello }")
    assert ftest == obj._open(ftest)
    ftest.close()
    fout = obj._open("testfile")
    readin = fout.read()
    assert readin == "{ hello }", readin + "plain file"
    fout.close()
    ftest.close()
    clean()

    print "Passed in", time.time() - start, "s"

if __name__ == '__main__':
    test()