source: sasview/src/sas/qtgui/Plotting/PlotUtilities.py @ b0b09b9

ESS_GUIESS_GUI_DocsESS_GUI_batch_fittingESS_GUI_bumps_abstractionESS_GUI_iss1116ESS_GUI_iss879ESS_GUI_iss959ESS_GUI_openclESS_GUI_orderingESS_GUI_sync_sascalc
Last change on this file since b0b09b9 was b0b09b9, checked in by Piotr Rozyczko <rozyczko@…>, 7 years ago

Initial changes to make SasView? run with python3

  • Property mode set to 100755
File size: 9.4 KB
Line 
1import sys
2import numpy
3import string
4
5from collections import OrderedDict
6
7# MPL shapes dictionary with some extra styles rendered internally.
8# Ordered for consistent display in combo boxes
9SHAPES = OrderedDict([
10        ('Circle' , 'o'),
11        ('Point' , '.'),
12        ('Pixel' , ','),
13        ('Triangle Down' , 'v'),
14        ('Triangle Up' , '^'),
15        ('Triangle Left' , '<'),
16        ('Triangle Right' , '>'),
17        ('Octagon' , '8'),
18        ('Square' , 's'),
19        ('Pentagon' , 'p'),
20        ('Star' , '*'),
21        ('Hexagon1' , 'h'),
22        ('Hexagon2' , 'H'),
23        ('Cross +' , 'p'),
24        ('Cross X ' , 'x'),
25        ('Diamond' , 'D'),
26        ('Thin Diamond' , 'd'),
27        ('Line' , '-'),
28        ('Dash' , '--'),
29        ('Vline' , 'vline'),
30        ('Step' , 'step'),
31])
32
33# MPL Colors dictionary. Ordered for consistent display
34COLORS = OrderedDict([
35        ('Blue', 'b'),
36        ('Green', 'g'),
37        ('Red', 'r'),
38        ('Cyan', 'c'),
39        ('Magenta', 'm'),
40        ('Yellow', 'y'),
41        ('Black', 'k'),
42        ('Custom', 'x'),
43])
44
45def build_matrix(data, qx_data, qy_data):
46    """
47    Build a matrix for 2d plot from a vector
48    Returns a matrix (image) with ~ square binning
49    Requirement: need 1d array formats of
50    data, qx_data, and qy_data
51    where each one corresponds to z, x, or y axis values
52
53    """
54    # No qx or qy given in a vector format
55    if qx_data is None or qy_data is None \
56            or qx_data.ndim != 1 or qy_data.ndim != 1:
57        return data
58
59    # maximum # of loops to fillup_pixels
60    # otherwise, loop could never stop depending on data
61    max_loop = 1
62    # get the x and y_bin arrays.
63    x_bins, y_bins = get_bins(qx_data, qy_data)
64    # set zero to None
65
66    #Note: Can not use scipy.interpolate.Rbf:
67    # 'cause too many data points (>10000)<=JHC.
68    # 1d array to use for weighting the data point averaging
69    #when they fall into a same bin.
70    weights_data = numpy.ones([data.size])
71    # get histogram of ones w/len(data); this will provide
72    #the weights of data on each bins
73    weights, xedges, yedges = numpy.histogram2d(x=qy_data,
74                                                y=qx_data,
75                                                bins=[y_bins, x_bins],
76                                                weights=weights_data)
77    # get histogram of data, all points into a bin in a way of summing
78    image, xedges, yedges = numpy.histogram2d(x=qy_data,
79                                              y=qx_data,
80                                              bins=[y_bins, x_bins],
81                                              weights=data)
82    # Now, normalize the image by weights only for weights>1:
83    # If weight == 1, there is only one data point in the bin so
84    # that no normalization is required.
85    image[weights > 1] = image[weights > 1] / weights[weights > 1]
86    # Set image bins w/o a data point (weight==0) as None (was set to zero
87    # by histogram2d.)
88    image[weights == 0] = None
89
90    # Fill empty bins with 8 nearest neighbors only when at least
91    #one None point exists
92    loop = 0
93
94    # do while loop until all vacant bins are filled up up
95    #to loop = max_loop
96    while not(numpy.isfinite(image[weights == 0])).all():
97        if loop >= max_loop:  # this protects never-ending loop
98            break
99        image = fillupPixels(image=image, weights=weights)
100        loop += 1
101
102    return image
103
104def get_bins(qx_data, qy_data):
105    """
106    get bins
107    return x_bins and y_bins: 1d arrays of the index with
108    ~ square binning
109    Requirement: need 1d array formats of
110    qx_data, and qy_data
111    where each one corresponds to  x, or y axis values
112    """
113    # No qx or qy given in a vector format
114    if qx_data is None or qy_data is None \
115            or qx_data.ndim != 1 or qy_data.ndim != 1:
116        return data
117
118    # find max and min values of qx and qy
119    xmax = qx_data.max()
120    xmin = qx_data.min()
121    ymax = qy_data.max()
122    ymin = qy_data.min()
123
124    # calculate the range of qx and qy: this way, it is a little
125    # more independent
126    x_size = xmax - xmin
127    y_size = ymax - ymin
128
129    # estimate the # of pixels on each axes
130    npix_y = int(numpy.floor(numpy.sqrt(len(qy_data))))
131    npix_x = int(numpy.floor(len(qy_data) / npix_y))
132
133    # bin size: x- & y-directions
134    xstep = x_size / (npix_x - 1)
135    ystep = y_size / (npix_y - 1)
136
137    # max and min taking account of the bin sizes
138    xmax = xmax + xstep / 2.0
139    xmin = xmin - xstep / 2.0
140    ymax = ymax + ystep / 2.0
141    ymin = ymin - ystep / 2.0
142
143    # store x and y bin centers in q space
144    x_bins = numpy.linspace(xmin, xmax, npix_x)
145    y_bins = numpy.linspace(ymin, ymax, npix_y)
146
147    #set x_bins and y_bins
148    return x_bins, y_bins
149
150def fillupPixels(image=None, weights=None):
151    """
152    Fill z values of the empty cells of 2d image matrix
153    with the average over up-to next nearest neighbor points
154
155    :param image: (2d matrix with some zi = None)
156
157    :return: image (2d array )
158
159    :TODO: Find better way to do for-loop below
160
161    """
162    # No image matrix given
163    if image == None or numpy.ndim(image) != 2 \
164            or numpy.isfinite(image).all() \
165            or weights == None:
166        return image
167    # Get bin size in y and x directions
168    len_y = len(image)
169    len_x = len(image[1])
170    temp_image = numpy.zeros([len_y, len_x])
171    weit = numpy.zeros([len_y, len_x])
172    # do for-loop for all pixels
173    for n_y in range(len(image)):
174        for n_x in range(len(image[1])):
175            # find only null pixels
176            if weights[n_y][n_x] > 0 or numpy.isfinite(image[n_y][n_x]):
177                continue
178            else:
179                # find 4 nearest neighbors
180                # check where or not it is at the corner
181                if n_y != 0 and numpy.isfinite(image[n_y - 1][n_x]):
182                    temp_image[n_y][n_x] += image[n_y - 1][n_x]
183                    weit[n_y][n_x] += 1
184                if n_x != 0 and numpy.isfinite(image[n_y][n_x - 1]):
185                    temp_image[n_y][n_x] += image[n_y][n_x - 1]
186                    weit[n_y][n_x] += 1
187                if n_y != len_y - 1 and numpy.isfinite(image[n_y + 1][n_x]):
188                    temp_image[n_y][n_x] += image[n_y + 1][n_x]
189                    weit[n_y][n_x] += 1
190                if n_x != len_x - 1 and numpy.isfinite(image[n_y][n_x + 1]):
191                    temp_image[n_y][n_x] += image[n_y][n_x + 1]
192                    weit[n_y][n_x] += 1
193                # go 4 next nearest neighbors when no non-zero
194                # neighbor exists
195                if n_y != 0 and n_x != 0 and\
196                        numpy.isfinite(image[n_y - 1][n_x - 1]):
197                    temp_image[n_y][n_x] += image[n_y - 1][n_x - 1]
198                    weit[n_y][n_x] += 1
199                if n_y != len_y - 1 and n_x != 0 and \
200                    numpy.isfinite(image[n_y + 1][n_x - 1]):
201                    temp_image[n_y][n_x] += image[n_y + 1][n_x - 1]
202                    weit[n_y][n_x] += 1
203                if n_y != len_y and n_x != len_x - 1 and \
204                    numpy.isfinite(image[n_y - 1][n_x + 1]):
205                    temp_image[n_y][n_x] += image[n_y - 1][n_x + 1]
206                    weit[n_y][n_x] += 1
207                if n_y != len_y - 1 and n_x != len_x - 1 and \
208                    numpy.isfinite(image[n_y + 1][n_x + 1]):
209                    temp_image[n_y][n_x] += image[n_y + 1][n_x + 1]
210                    weit[n_y][n_x] += 1
211
212    # get it normalized
213    ind = (weit > 0)
214    image[ind] = temp_image[ind] / weit[ind]
215
216    return image
217
218def rescale(lo, hi, step, pt=None, bal=None, scale='linear'):
219    """
220    Rescale (lo,hi) by step, returning the new (lo,hi)
221    The scaling is centered on pt, with positive values of step
222    driving lo/hi away from pt and negative values pulling them in.
223    If bal is given instead of point, it is already in [0,1] coordinates.
224
225    This is a helper function for step-based zooming.
226    """
227    # Convert values into the correct scale for a linear transformation
228    # TODO: use proper scale transformers
229    loprev = lo
230    hiprev = hi
231    if scale == 'log':
232        assert lo > 0
233        if lo > 0:
234            lo = numpy.log10(lo)
235        if hi > 0:
236            hi = numpy.log10(hi)
237        if pt is not None:
238            pt = numpy.log10(pt)
239
240    # Compute delta from axis range * %, or 1-% if persent is negative
241    if step > 0:
242        delta = float(hi - lo) * step / 100
243    else:
244        delta = float(hi - lo) * step / (100 - step)
245
246    # Add scale factor proportionally to the lo and hi values,
247    # preserving the
248    # point under the mouse
249    if bal is None:
250        bal = float(pt - lo) / (hi - lo)
251    lo = lo - (bal * delta)
252    hi = hi + (1 - bal) * delta
253
254    # Convert transformed values back to the original scale
255    if scale == 'log':
256        if (lo <= -250) or (hi >= 250):
257            lo = loprev
258            hi = hiprev
259        else:
260            lo, hi = numpy.power(10., lo), numpy.power(10., hi)
261    return (lo, hi)
262
263def getValidColor(color):
264    '''
265    Returns a valid matplotlib color
266    '''
267
268    if color is None:
269        return color
270
271    # Check if it's an int
272    if isinstance(color, int):
273        # Check if it's within the range
274        if 0 <= color <=6:
275            color = list(COLORS.values())[color]
276    # Check if it's an RGB string
277    elif isinstance(color, str):
278        # Assure the correctnes of the string
279        assert(color[0]=="#" and len(color) == 7)
280        assert(all(c in string.hexdigits for c in color[1:]))
281    else:
282        raise AttributeError
283
284    return color
Note: See TracBrowser for help on using the repository browser.