This file is indexed.

/usr/lib/python2.7/dist-packages/ginga/BaseImage.py is in python-ginga 2.6.1-2.

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
#
# BaseImage.py -- Abstraction of an generic data image.
#
# This is open-source software licensed under a BSD license.
# Please see the file LICENSE.txt for details.
#
from ginga.util.six.moves import map

import numpy
import logging

from ginga.misc import Bunch, Callback
from ginga import trcalc, AutoCuts


class ImageError(Exception):
    pass


class ViewerObjectBase(Callback.Callbacks):

    def __init__(self, metadata=None, logger=None, name=None):

        Callback.Callbacks.__init__(self)

        if logger is not None:
            self.logger = logger
        else:
            self.logger = logging.Logger('BaseImage')
        self.metadata = {}
        if metadata:
            self.update_metadata(metadata)
        # make sure an object has these attributes
        # TODO: this maybe should have a unique random string or something
        # but we'd have to fix a lot of code that is currently checking for
        # None
        self.metadata.setdefault('name', None)

        # For callbacks
        for name in ('modified', ):
            self.enable_callback(name)

    def get_metadata(self):
        return self.metadata.copy()

    def clear_metadata(self):
        self.metadata = {}

    def clear_all(self):
        self.clear_metadata()

    def update_metadata(self, map_like):
        for key, val in map_like.items():
            self.metadata[key] = val

    def get(self, kwd, *args):
        if kwd in self.metadata:
            return self.metadata[kwd]
        else:
            # return a default if there is one
            if len(args) > 0:
                return args[0]
            raise KeyError(kwd)

    def get_list(self, *args):
        return list(map(self.get, args))

    def __getitem__(self, kwd):
        return self.metadata[kwd]

    def update(self, kwds):
        self.metadata.update(kwds)

    def set(self, **kwds):
        self.update(kwds)

    def __setitem__(self, kwd, value):
        self.metadata[kwd] = value


class BaseImage(ViewerObjectBase):

    def __init__(self, data_np=None, metadata=None, logger=None, name=None):

        ViewerObjectBase.__init__(self, logger=logger, metadata=metadata,
                                  name=name)

        if data_np is None:
            data_np = numpy.zeros((1, 1))
        self._data = data_np
        self.order = ''
        self.name = name

        self._set_minmax()

        self.autocuts = AutoCuts.Histogram(self.logger)

    @property
    def shape(self):
        return self._get_data().shape

    @property
    def width(self):
        # NOTE: numpy stores data in column-major layout
        return self.shape[1]

    @property
    def height(self):
        # NOTE: numpy stores data in column-major layout
        return self.shape[0]

    @property
    def depth(self):
        return self.get_depth()

    @property
    def ndim(self):
        return len(self.shape)

    def get_size(self):
        return (self.width, self.height)

    def get_depth(self):
        shape = self.shape
        if len(shape) > 2:
            return shape[2]
        return 1

    def get_shape(self):
        return self.shape

    def get_center(self):
        wd, ht = self.get_size()
        ctr_x, ctr_y = wd // 2, ht // 2
        return (ctr_x, ctr_y)

    def get_data(self):
        return self._data

    def _get_data(self):
        return self._data

    def _get_fast_data(self):
        """
        Return an array similar to but possibly smaller than self._data,
        for fast calculation of the intensity distribution
        """
        return self._data

    def copy_data(self):
        data = self._get_data()
        return data.copy()

    def get_data_xy(self, x, y):
        assert (x >= 0) and (y >= 0), \
            ImageError("Indexes out of range: (x=%d, y=%d)" % (
                x, y))
        view = numpy.s_[y, x]
        return self._slice(view)

    def set_data(self, data_np, metadata=None, astype=None):
        """Use this method to SHARE (not copy) the incoming array.
        """
        if astype:
            data = data_np.astype(astype)
        else:
            data = data_np
        self._data = data

        if metadata:
            self.update_metadata(metadata)

        self._set_minmax()

        self.make_callback('modified')

    def clear_all(self):
        # clear metadata
        super(BaseImage, self).clear_all()

        # unreference data array
        self._data = numpy.zeros((1, 1))

    def _slice(self, view):
        return self._get_data()[view]

    def get_slice(self, c):
        view = [slice(None)] * self.ndim
        view[-1] = self.order.index(c.upper())
        return self._slice(view)

    def has_slice(self, c):
        return c.upper() in self.order

    def get_array(self, order):
        order = order.upper()
        if order == self.order:
            return self._get_data()
        l = [self.get_slice(c) for c in order]
        return numpy.dstack(l)

    def set_order(self, order):
        self.order = order.upper()

    def get_order(self):
        return self.order

    def get_order_indexes(self, cs):
        cs = cs.upper()
        return [self.order.index(c) for c in cs]

    def has_valid_wcs(self):
        return hasattr(self, 'wcs') and self.wcs.has_valid_wcs()

    def _set_minmax(self):
        data = self._get_fast_data()
        try:
            self.maxval = numpy.nanmax(data)
            self.minval = numpy.nanmin(data)
        except Exception:
            self.maxval = 0
            self.minval = 0

        # TODO: see if there is a faster way to ignore infinity
        try:
            if numpy.isfinite(self.maxval):
                self.maxval_noinf = self.maxval
            else:
                self.maxval_noinf = numpy.nanmax(data[numpy.isfinite(data)])
        except:
            self.maxval_noinf = self.maxval

        try:
            if numpy.isfinite(self.minval):
                self.minval_noinf = self.minval
            else:
                self.minval_noinf = numpy.nanmin(data[numpy.isfinite(data)])
        except:
            self.minval_noinf = self.minval

    def get_minmax(self, noinf=False):
        if not noinf:
            return (self.minval, self.maxval)
        else:
            return (self.minval_noinf, self.maxval_noinf)

    def get_header(self):
        return self.get('header', Header())

    def transfer(self, other, astype=None):
        data = self._get_data()
        other.set_data(data, metadata=self.metadata, astype=astype)

    def copy(self, astype=None):
        data = self.copy_data()
        metadata = self.get_metadata()
        other = self.__class__(data_np=data, metadata=metadata)
        return other

    def cutout_data(self, x1, y1, x2, y2, xstep=1, ystep=1, astype=None):
        """cut out data area based on coords.
        """
        view = numpy.s_[y1:y2:ystep, x1:x2:xstep]
        data = self._slice(view)
        if astype:
            data = data.astype(astype)
        return data

    def cutout_adjust(self, x1, y1, x2, y2, xstep=1, ystep=1, astype=None):
        dx = x2 - x1
        dy = y2 - y1

        if x1 < 0:
            x1, x2 = 0, dx
        else:
            if x2 >= self.width:
                x2 = self.width
                x1 = x2 - dx

        if y1 < 0:
            y1, y2 = 0, dy
        else:
            if y2 >= self.height:
                y2 = self.height
                y1 = y2 - dy

        data = self.cutout_data(x1, y1, x2, y2, xstep=xstep, ystep=ystep,
                                astype=astype)
        return (data, x1, y1, x2, y2)

    def cutout_radius(self, x, y, radius, xstep=1, ystep=1, astype=None):
        return self.cutout_adjust(x-radius, y-radius,
                                  x+radius+1, y+radius+1,
                                  xstep=xstep, ystep=ystep,
                                  astype=astype)

    def cutout_cross(self, x, y, radius):
        """Cut two data subarrays that have a center at (x, y) and with
        radius (radius) from (image).  Returns the starting pixel (x0, y0)
        of each cut and the respective arrays (xarr, yarr).
        """
        n = radius
        wd, ht = self.get_size()
        x0, x1 = max(0, x - n), min(wd - 1, x + n)
        y0, y1 = max(0, y - n), min(ht - 1, y + n)

        xview = numpy.s_[y, x0:x1 + 1]
        yview = numpy.s_[y0:y1 + 1, x]

        xarr = self._slice(xview)
        yarr = self._slice(yview)

        return (x0, y0, xarr, yarr)

    def get_shape_mask(self, shape_obj):
        """
        Return full mask where True marks pixels within the given shape.
        """
        wd, ht = self.get_size()
        yi = numpy.mgrid[:ht].reshape(-1, 1)
        xi = numpy.mgrid[:wd].reshape(1, -1)
        contains = shape_obj.contains_arr(xi, yi)
        return contains

    def get_shape_view(self, shape_obj, avoid_oob=True):
        """
        Calculate a bounding box in the data enclosing `shape_obj` and
        return a view that accesses it and a mask that is True only for
        pixels enclosed in the region.

        If `avoid_oob` is True (default) then the bounding box is clipped
        to avoid coordinates outside of the actual data.
        """
        x1, y1, x2, y2 = map(int, shape_obj.get_llur())

        if avoid_oob:
            # avoid out of bounds indexes
            wd, ht = self.get_size()
            x1, x2 = max(0, x1), min(x2, wd-1)
            y1, y2 = max(0, y1), min(y2, ht-1)

        # calculate pixel containment mask in bbox
        yi = numpy.mgrid[y1:y2+1].reshape(-1, 1)
        xi = numpy.mgrid[x1:x2+1].reshape(1, -1)
        contains = shape_obj.contains_arr(xi, yi)

        view = numpy.s_[y1:y2+1, x1:x2+1]
        return (view, contains)

    def cutout_shape(self, shape_obj):
        """
        Cut out and return a portion of the data corresponding to `shape_obj`.
        A masked numpy array is returned, where the pixels not enclosed in
        the shape are masked out.
        """

        view, mask = self.get_shape_view(shape_obj)

        # cutout our enclosing (possibly shortened) bbox
        data = self._slice(view)

        # mask non-containing members
        mdata = numpy.ma.array(data, mask=numpy.logical_not(mask))
        return mdata

    def get_scaled_cutout_wdht(self, x1, y1, x2, y2, new_wd, new_ht,
                               method='basic'):
        """Extract a region of the image defined by corners (x1, y1) and
        (x2, y2) and resample it to fit dimensions (new_wd, new_ht).

        `method` describes the method of interpolation used, where the
        default "basic" is nearest neighbor.
        """

        if method in ('basic', 'view'):
            shp = self.shape

            (view, (scale_x, scale_y)) = \
                trcalc.get_scaled_cutout_wdht_view(shp, x1, y1, x2, y2,
                                                   new_wd, new_ht)
            newdata = self._slice(view)

        else:
            data_np = self._get_data()
            (newdata, (scale_x, scale_y)) = \
                trcalc.get_scaled_cutout_wdht(data_np, x1, y1, x2, y2,
                                              new_wd, new_ht,
                                              interpolation=method,
                                              logger=self.logger)

        res = Bunch.Bunch(data=newdata, scale_x=scale_x, scale_y=scale_y)
        return res

    def get_scaled_cutout_basic(self, x1, y1, x2, y2, scale_x, scale_y,
                                method='basic'):
        """Extract a region of the image defined by corners (x1, y1) and
        (x2, y2) and scale it by scale factors (scale_x, scale_y).

        `method` describes the method of interpolation used, where the
        default "basic" is nearest neighbor.
        """

        new_wd = int(round(scale_x * (x2 - x1 + 1)))
        new_ht = int(round(scale_y * (y2 - y1 + 1)))

        return self.get_scaled_cutout_wdht(x1, y1, x2, y2, new_wd, new_ht,
                                           # TODO:
                                           # this causes a problem for the
                                           # current Glue plugin--update that
                                           #method=method
                                           )

    def get_scaled_cutout(self, x1, y1, x2, y2, scale_x, scale_y,
                          method='basic', logger=None):
        if method == 'basic':
            return self.get_scaled_cutout_basic(x1, y1, x2, y2,
                                                scale_x, scale_y)

        data = self._get_data()
        newdata, (scale_x, scale_y) = trcalc.get_scaled_cutout_basic(
            data, x1, y1, x2, y2, scale_x, scale_y, interpolation=method)

        res = Bunch.Bunch(data=newdata, scale_x=scale_x, scale_y=scale_y)
        return res

    def get_thumbnail(self, length):
        wd, ht = self.get_size()
        if ht == 0:
            width, height = 1, 1
        elif wd > ht:
            width, height = length, int(length * float(ht) / wd)
        else:
            width, height = int(length * float(wd) / ht), length

        res = self.get_scaled_cutout_wdht(0, 0, wd, ht, width, height)
        return res.data

    def get_pixels_on_line(self, x1, y1, x2, y2, getvalues=True):
        """Uses Bresenham's line algorithm to enumerate the pixels along
        a line.
        (see http://en.wikipedia.org/wiki/Bresenham%27s_line_algorithm)

        If `getvalues`==False then it will return tuples of (x, y) coordinates
        instead of pixel values.
        """
        # NOTE: seems to be necessary or we get a non-terminating result
        x1, y1, x2, y2 = int(x1), int(y1), int(x2), int(y2)

        dx = abs(x2 - x1)
        dy = abs(y2 - y1)
        if x1 < x2:
            sx = 1
        else:
            sx = -1
        if y1 < y2:
            sy = 1
        else:
            sy = -1
        err = dx - dy

        res = []
        x, y = x1, y1
        while True:
            if getvalues:
                try:
                    val = self.get_data_xy(x, y)
                except Exception:
                    val = numpy.NaN
                res.append(val)
            else:
                res.append((x, y))
            if (x == x2) and (y == y2):
                break
            e2 = 2 * err
            if e2 > -dy:
                err = err - dy
                x += sx
            if e2 < dx:
                err = err + dx
                y += sy

        return res

    def info_xy(self, data_x, data_y, settings):
        # Get the value under the data coordinates
        try:
            value = self.get_data_xy(int(data_x), int(data_y))

        except Exception as e:
            value = None

        info = Bunch.Bunch(itype='base', data_x=data_x, data_y=data_y,
                           x=data_x, y=data_y,
                           value=value)
        return info


class Header(dict):

    def __init__(self, *args, **kwdargs):
        super(Header, self).__init__(*args, **kwdargs)
        self.keyorder = []

    def __getitem__(self, key):
        bnch = super(Header, self).__getitem__(key)
        return bnch.value

    def __setitem__(self, key, value):
        try:
            bnch = super(Header, self).__getitem__(key)
            bnch.value = value
        except KeyError:
            bnch = Bunch.Bunch(key=key, value=value, comment='')
            self.keyorder.append(key)
            super(Header, self).__setitem__(key, bnch)
        return bnch

    def __delitem__(self, key):
        super(Header, self).__delitem__(key)
        self.keyorder.remove(key)

    def get_card(self, key):
        bnch = super(Header, self).__getitem__(key)
        return bnch

    def set_card(self, key, value, comment=None):
        try:
            bnch = super(Header, self).__getitem__(key)
            bnch.value = value
            if not (comment is None):
                bnch.comment = comment
        except KeyError:
            if comment is None:
                comment = ''
            bnch = Bunch.Bunch(key=key, value=value, comment=comment)
            self.keyorder.append(key)
            super(Header, self).__setitem__(key, bnch)
        return bnch

    def get_keyorder(self):
        return self.keyorder

    def keys(self):
        return self.keyorder

    def items(self):
        return [(key, self[key]) for key in self.keys()]

    def get(self, key, alt=None):
        try:
            return self.__getitem__(key)
        except KeyError:
            return alt

    def update(self, map_kind):
        for key, value in map_kind.items():
            self.__setitem__(key, value)

    def asdict(self):
        return dict([(key, self[key]) for key in self.keys()])

#END