1 | import sys |
---|
2 | import numpy |
---|
3 | import string |
---|
4 | |
---|
5 | from collections import OrderedDict |
---|
6 | |
---|
7 | # MPL shapes dictionary with some extra styles rendered internally. |
---|
8 | # Ordered for consistent display in combo boxes |
---|
9 | SHAPES = 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 |
---|
34 | COLORS = 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 | |
---|
45 | def 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 | |
---|
104 | def 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 | |
---|
150 | def 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 is None or numpy.ndim(image) != 2 \ |
---|
164 | or numpy.isfinite(image).all() \ |
---|
165 | or weights is 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 | |
---|
218 | def 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 | |
---|
263 | def 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 | elif isinstance(color, str): |
---|
277 | # This could be a one letter code |
---|
278 | if len(color) == 1: |
---|
279 | if not color in list (COLORS.values()): |
---|
280 | raise AttributeError |
---|
281 | else: |
---|
282 | # or an RGB string |
---|
283 | assert(color[0]=="#" and len(color) == 7) |
---|
284 | assert(all(c in string.hexdigits for c in color[1:])) |
---|
285 | else: |
---|
286 | raise AttributeError |
---|
287 | |
---|
288 | return color |
---|