1 | #!/usr/bin/env python |
---|
2 | r""" |
---|
3 | Show numerical precision of $2 J_1(x)/x$. |
---|
4 | """ |
---|
5 | from __future__ import division, print_function |
---|
6 | |
---|
7 | import sys |
---|
8 | import os |
---|
9 | sys.path.insert(0, os.path.abspath(os.path.join(os.path.dirname(__file__), '..'))) |
---|
10 | |
---|
11 | import numpy as np |
---|
12 | from numpy import pi, inf |
---|
13 | import scipy.special |
---|
14 | try: |
---|
15 | from mpmath import mp |
---|
16 | except ImportError: |
---|
17 | # CRUFT: mpmath split out into its own package |
---|
18 | from sympy.mpmath import mp |
---|
19 | #import matplotlib; matplotlib.use('TkAgg') |
---|
20 | import pylab |
---|
21 | |
---|
22 | from sasmodels import core, data, direct_model, modelinfo |
---|
23 | |
---|
24 | class Comparator(object): |
---|
25 | def __init__(self, name, mp_function, np_function, ocl_function, xaxis, limits): |
---|
26 | self.name = name |
---|
27 | self.mp_function = mp_function |
---|
28 | self.np_function = np_function |
---|
29 | self.ocl_function = ocl_function |
---|
30 | self.xaxis = xaxis |
---|
31 | self.limits = limits |
---|
32 | |
---|
33 | def __repr__(self): |
---|
34 | return "Comparator(%s)"%self.name |
---|
35 | |
---|
36 | def call_mpmath(self, vec, bits=500): |
---|
37 | """ |
---|
38 | Direct calculation using mpmath extended precision library. |
---|
39 | """ |
---|
40 | with mp.workprec(bits): |
---|
41 | return [self.mp_function(mp.mpf(x)) for x in vec] |
---|
42 | |
---|
43 | def call_numpy(self, x, dtype): |
---|
44 | """ |
---|
45 | Direct calculation using numpy/scipy. |
---|
46 | """ |
---|
47 | x = np.asarray(x, dtype) |
---|
48 | return self.np_function(x) |
---|
49 | |
---|
50 | def call_ocl(self, x, dtype, platform='ocl'): |
---|
51 | """ |
---|
52 | Calculation using sasmodels ocl libraries. |
---|
53 | """ |
---|
54 | x = np.asarray(x, dtype) |
---|
55 | model = core.build_model(self.ocl_function, dtype=dtype) |
---|
56 | calculator = direct_model.DirectModel(data.empty_data1D(x), model) |
---|
57 | return calculator(background=0) |
---|
58 | |
---|
59 | def run(self, xrange="log", diff=True): |
---|
60 | r""" |
---|
61 | Compare accuracy of different methods for computing f. |
---|
62 | |
---|
63 | *xrange* is log=[10^-3,10^5], linear=[1,1000], zoom[1000,1010], |
---|
64 | or neg=[-100,100]. |
---|
65 | |
---|
66 | *diff* is False if showing function value rather than relative error. |
---|
67 | |
---|
68 | *x_bits* is the precision with which the x values are specified. The |
---|
69 | default 23 should reproduce the equivalent of a single precisio |
---|
70 | """ |
---|
71 | linear = xrange != "log" |
---|
72 | if xrange == "zoom": |
---|
73 | lin_min, lin_max, lin_steps = 1000, 1010, 2000 |
---|
74 | elif xrange == "neg": |
---|
75 | lin_min, lin_max, lin_steps = -100.1, 100.1, 2000 |
---|
76 | else: |
---|
77 | lin_min, lin_max, lin_steps = 1, 1000, 2000 |
---|
78 | lin_min = max(lin_min, self.limits[0]) |
---|
79 | lin_max = min(lin_max, self.limits[1]) |
---|
80 | log_min, log_max, log_steps = -3, 5, 400 |
---|
81 | with mp.workprec(500): |
---|
82 | if linear: |
---|
83 | qrf = np.linspace(lin_min, lin_max, lin_steps, dtype='single') |
---|
84 | qr = [mp.mpf(float(v)) for v in qrf] |
---|
85 | #qr = mp.linspace(lin_min, lin_max, lin_steps) |
---|
86 | else: |
---|
87 | qrf = np.logspace(log_min, log_max, log_steps, dtype='single') |
---|
88 | qr = [mp.mpf(float(v)) for v in qrf] |
---|
89 | #qr = [10**v for v in mp.linspace(log_min, log_max, log_steps)] |
---|
90 | |
---|
91 | target = self.call_mpmath(qr, bits=500) |
---|
92 | pylab.subplot(121) |
---|
93 | self.compare(qr, 'single', target, linear, diff) |
---|
94 | pylab.legend(loc='best') |
---|
95 | pylab.subplot(122) |
---|
96 | self.compare(qr, 'double', target, linear, diff) |
---|
97 | pylab.legend(loc='best') |
---|
98 | pylab.suptitle(self.name + " compared to 500-bit mpmath") |
---|
99 | |
---|
100 | def compare(self, x, precision, target, linear=False, diff=True): |
---|
101 | r""" |
---|
102 | Compare the different computation methods using the given precision. |
---|
103 | """ |
---|
104 | if precision == 'single': |
---|
105 | #n=11; plotdiff(x, target, self.call_mpmath(x, n), 'mp %d bits'%n, diff=diff) |
---|
106 | #n=23; plotdiff(x, target, self.call_mpmath(x, n), 'mp %d bits'%n, diff=diff) |
---|
107 | pass |
---|
108 | elif precision == 'double': |
---|
109 | #n=53; plotdiff(x, target, self.call_mpmath(x, n), 'mp %d bits'%n, diff=diff) |
---|
110 | #n=83; plotdiff(x, target, self.call_mpmath(x, n), 'mp %d bits'%n, diff=diff) |
---|
111 | pass |
---|
112 | plotdiff(x, target, self.call_numpy(x, precision), 'numpy '+precision, diff=diff) |
---|
113 | plotdiff(x, target, self.call_ocl(x, precision, 0), 'OpenCL '+precision, diff=diff) |
---|
114 | pylab.xlabel(self.xaxis) |
---|
115 | if diff: |
---|
116 | pylab.ylabel("relative error") |
---|
117 | else: |
---|
118 | pylab.ylabel(self.name) |
---|
119 | pylab.semilogx(x, target, '-', label="true value") |
---|
120 | if linear: |
---|
121 | pylab.xscale('linear') |
---|
122 | |
---|
123 | def plotdiff(x, target, actual, label, diff=True): |
---|
124 | """ |
---|
125 | Plot the computed value. |
---|
126 | |
---|
127 | Use relative error if SHOW_DIFF, otherwise just plot the value directly. |
---|
128 | """ |
---|
129 | if diff: |
---|
130 | err = np.array([abs((t-a)/t) for t, a in zip(target, actual)], 'd') |
---|
131 | #err = np.clip(err, 0, 1) |
---|
132 | pylab.loglog(x, err, '-', label=label) |
---|
133 | else: |
---|
134 | limits = np.min(target), np.max(target) |
---|
135 | pylab.semilogx(x, np.clip(actual, *limits), '-', label=label) |
---|
136 | |
---|
137 | def make_ocl(function, name, source=[]): |
---|
138 | class Kernel(object): |
---|
139 | pass |
---|
140 | Kernel.__file__ = name+".py" |
---|
141 | Kernel.name = name |
---|
142 | Kernel.parameters = [] |
---|
143 | Kernel.source = source |
---|
144 | Kernel.Iq = function |
---|
145 | model_info = modelinfo.make_model_info(Kernel) |
---|
146 | return model_info |
---|
147 | |
---|
148 | |
---|
149 | # =============== FUNCTION DEFINITIONS ================ |
---|
150 | |
---|
151 | FUNCTIONS = {} |
---|
152 | def add_function(name, mp_function, np_function, ocl_function, |
---|
153 | shortname=None, xaxis="x", limits=(-inf, inf)): |
---|
154 | if shortname is None: |
---|
155 | shortname = name.replace('(x)', '').replace(' ', '') |
---|
156 | FUNCTIONS[shortname] = Comparator(name, mp_function, np_function, ocl_function, xaxis, limits) |
---|
157 | |
---|
158 | add_function( |
---|
159 | name="J0(x)", |
---|
160 | mp_function=mp.j0, |
---|
161 | np_function=scipy.special.j0, |
---|
162 | ocl_function=make_ocl("return sas_J0(q);", "sas_J0", ["lib/polevl.c", "lib/sas_J0.c"]), |
---|
163 | ) |
---|
164 | add_function( |
---|
165 | name="J1(x)", |
---|
166 | mp_function=mp.j1, |
---|
167 | np_function=scipy.special.j1, |
---|
168 | ocl_function=make_ocl("return sas_J1(q);", "sas_J1", ["lib/polevl.c", "lib/sas_J1.c"]), |
---|
169 | ) |
---|
170 | add_function( |
---|
171 | name="JN(-3, x)", |
---|
172 | mp_function=lambda x: mp.besselj(-3, x), |
---|
173 | np_function=lambda x: scipy.special.jn(-3, x), |
---|
174 | ocl_function=make_ocl("return sas_JN(-3, q);", "sas_JN", |
---|
175 | ["lib/polevl.c", "lib/sas_J0.c", "lib/sas_J1.c", "lib/sas_JN.c"]), |
---|
176 | shortname="J-3", |
---|
177 | ) |
---|
178 | add_function( |
---|
179 | name="JN(3, x)", |
---|
180 | mp_function=lambda x: mp.besselj(3, x), |
---|
181 | np_function=lambda x: scipy.special.jn(3, x), |
---|
182 | ocl_function=make_ocl("return sas_JN(3, q);", "sas_JN", |
---|
183 | ["lib/polevl.c", "lib/sas_J0.c", "lib/sas_J1.c", "lib/sas_JN.c"]), |
---|
184 | shortname="J3", |
---|
185 | ) |
---|
186 | add_function( |
---|
187 | name="JN(2, x)", |
---|
188 | mp_function=lambda x: mp.besselj(2, x), |
---|
189 | np_function=lambda x: scipy.special.jn(2, x), |
---|
190 | ocl_function=make_ocl("return sas_JN(2, q);", "sas_JN", |
---|
191 | ["lib/polevl.c", "lib/sas_J0.c", "lib/sas_J1.c", "lib/sas_JN.c"]), |
---|
192 | shortname="J2", |
---|
193 | ) |
---|
194 | add_function( |
---|
195 | name="2 J1(x)/x", |
---|
196 | mp_function=lambda x: 2*mp.j1(x)/x, |
---|
197 | np_function=lambda x: 2*scipy.special.j1(x)/x, |
---|
198 | ocl_function=make_ocl("return sas_2J1x_x(q);", "sas_2J1x_x", ["lib/polevl.c", "lib/sas_J1.c"]), |
---|
199 | ) |
---|
200 | add_function( |
---|
201 | name="J1(x)", |
---|
202 | mp_function=mp.j1, |
---|
203 | np_function=scipy.special.j1, |
---|
204 | ocl_function=make_ocl("return sas_J1(q);", "sas_J1", ["lib/polevl.c", "lib/sas_J1.c"]), |
---|
205 | ) |
---|
206 | add_function( |
---|
207 | name="Si(x)", |
---|
208 | mp_function=mp.si, |
---|
209 | np_function=lambda x: scipy.special.sici(x)[0], |
---|
210 | ocl_function=make_ocl("return sas_Si(q);", "sas_Si", ["lib/sas_Si.c"]), |
---|
211 | ) |
---|
212 | #import fnlib |
---|
213 | #add_function( |
---|
214 | # name="fnlibJ1", |
---|
215 | # mp_function=mp.j1, |
---|
216 | # np_function=fnlib.J1, |
---|
217 | # ocl_function=make_ocl("return sas_J1(q);", "sas_J1", ["lib/polevl.c", "lib/sas_J1.c"]), |
---|
218 | #) |
---|
219 | add_function( |
---|
220 | name="sin(x)", |
---|
221 | mp_function=mp.sin, |
---|
222 | np_function=np.sin, |
---|
223 | #ocl_function=make_ocl("double sn, cn; SINCOS(q,sn,cn); return sn;", "sas_sin"), |
---|
224 | ocl_function=make_ocl("return sin(q);", "sas_sin"), |
---|
225 | ) |
---|
226 | add_function( |
---|
227 | name="sin(x)/x", |
---|
228 | mp_function=lambda x: mp.sin(x)/x if x != 0 else 1, |
---|
229 | ## scipy sinc function is inaccurate and has an implied pi*x term |
---|
230 | #np_function=lambda x: scipy.special.sinc(x/pi), |
---|
231 | ## numpy sin(x)/x needs to check for x=0 |
---|
232 | np_function=lambda x: np.sin(x)/x, |
---|
233 | ocl_function=make_ocl("return sas_sinx_x(q);", "sas_sinc"), |
---|
234 | ) |
---|
235 | add_function( |
---|
236 | name="cos(x)", |
---|
237 | mp_function=mp.cos, |
---|
238 | np_function=np.cos, |
---|
239 | #ocl_function=make_ocl("double sn, cn; SINCOS(q,sn,cn); return cn;", "sas_cos"), |
---|
240 | ocl_function=make_ocl("return cos(q);", "sas_cos"), |
---|
241 | ) |
---|
242 | add_function( |
---|
243 | name="gamma(x)", |
---|
244 | mp_function=mp.gamma, |
---|
245 | np_function=scipy.special.gamma, |
---|
246 | ocl_function=make_ocl("return sas_gamma(q);", "sas_gamma", ["lib/sas_gamma.c"]), |
---|
247 | limits=(-3.1,10), |
---|
248 | ) |
---|
249 | add_function( |
---|
250 | name="erf(x)", |
---|
251 | mp_function=mp.erf, |
---|
252 | np_function=scipy.special.erf, |
---|
253 | ocl_function=make_ocl("return sas_erf(q);", "sas_erf", ["lib/polevl.c", "lib/sas_erf.c"]), |
---|
254 | limits=(-5.,5.), |
---|
255 | ) |
---|
256 | add_function( |
---|
257 | name="erfc(x)", |
---|
258 | mp_function=mp.erfc, |
---|
259 | np_function=scipy.special.erfc, |
---|
260 | ocl_function=make_ocl("return sas_erfc(q);", "sas_erfc", ["lib/polevl.c", "lib/sas_erf.c"]), |
---|
261 | limits=(-5.,5.), |
---|
262 | ) |
---|
263 | add_function( |
---|
264 | name="arctan(x)", |
---|
265 | mp_function=mp.atan, |
---|
266 | np_function=np.arctan, |
---|
267 | ocl_function=make_ocl("return atan(q);", "sas_arctan"), |
---|
268 | ) |
---|
269 | add_function( |
---|
270 | name="3 j1(x)/x", |
---|
271 | mp_function=lambda x: 3*(mp.sin(x)/x - mp.cos(x))/(x*x), |
---|
272 | # Note: no taylor expansion near 0 |
---|
273 | np_function=lambda x: 3*(np.sin(x)/x - np.cos(x))/(x*x), |
---|
274 | ocl_function=make_ocl("return sas_3j1x_x(q);", "sas_j1c", ["lib/sas_3j1x_x.c"]), |
---|
275 | ) |
---|
276 | add_function( |
---|
277 | name="fmod_2pi", |
---|
278 | mp_function=lambda x: mp.fmod(x, 2*mp.pi), |
---|
279 | np_function=lambda x: np.fmod(x, 2*np.pi), |
---|
280 | ocl_function=make_ocl("return fmod(q, 2*M_PI);", "sas_fmod"), |
---|
281 | ) |
---|
282 | |
---|
283 | RADIUS=3000 |
---|
284 | LENGTH=30 |
---|
285 | THETA=45 |
---|
286 | def mp_cyl(x): |
---|
287 | f = mp.mpf |
---|
288 | theta = f(THETA)*mp.pi/f(180) |
---|
289 | qr = x * f(RADIUS)*mp.sin(theta) |
---|
290 | qh = x * f(LENGTH)/f(2)*mp.cos(theta) |
---|
291 | return (f(2)*mp.j1(qr)/qr * mp.sin(qh)/qh)**f(2) |
---|
292 | def np_cyl(x): |
---|
293 | f = np.float64 if x.dtype == np.float64 else np.float32 |
---|
294 | theta = f(THETA)*f(np.pi)/f(180) |
---|
295 | qr = x * f(RADIUS)*np.sin(theta) |
---|
296 | qh = x * f(LENGTH)/f(2)*np.cos(theta) |
---|
297 | return (f(2)*scipy.special.j1(qr)/qr*np.sin(qh)/qh)**f(2) |
---|
298 | ocl_cyl = """\ |
---|
299 | double THETA = %(THETA).15e*M_PI_180; |
---|
300 | double qr = q*%(RADIUS).15e*sin(THETA); |
---|
301 | double qh = q*0.5*%(LENGTH).15e*cos(THETA); |
---|
302 | return square(sas_2J1x_x(qr)*sas_sinx_x(qh)); |
---|
303 | """%{"LENGTH":LENGTH, "RADIUS": RADIUS, "THETA": THETA} |
---|
304 | add_function( |
---|
305 | name="cylinder(r=%g, l=%g, theta=%g)"%(RADIUS, LENGTH, THETA), |
---|
306 | mp_function=mp_cyl, |
---|
307 | np_function=np_cyl, |
---|
308 | ocl_function=make_ocl(ocl_cyl, "ocl_cyl", ["lib/polevl.c", "lib/sas_J1.c"]), |
---|
309 | shortname="cylinder", |
---|
310 | xaxis="$q/A^{-1}$", |
---|
311 | ) |
---|
312 | |
---|
313 | lanczos_gamma = """\ |
---|
314 | const double coeff[] = { |
---|
315 | 76.18009172947146, -86.50532032941677, |
---|
316 | 24.01409824083091, -1.231739572450155, |
---|
317 | 0.1208650973866179e-2,-0.5395239384953e-5 |
---|
318 | }; |
---|
319 | const double x = q; |
---|
320 | double tmp = x + 5.5; |
---|
321 | tmp -= (x + 0.5)*log(tmp); |
---|
322 | double ser = 1.000000000190015; |
---|
323 | for (int k=0; k < 6; k++) ser += coeff[k]/(x + k+1); |
---|
324 | return -tmp + log(2.5066282746310005*ser/x); |
---|
325 | """ |
---|
326 | add_function( |
---|
327 | name="log gamma(x)", |
---|
328 | mp_function=mp.loggamma, |
---|
329 | np_function=scipy.special.gammaln, |
---|
330 | ocl_function=make_ocl(lanczos_gamma, "lgamma"), |
---|
331 | ) |
---|
332 | |
---|
333 | # Alternate versions of 3 j1(x)/x, for posterity |
---|
334 | def taylor_3j1x_x(x): |
---|
335 | """ |
---|
336 | Calculation using taylor series. |
---|
337 | """ |
---|
338 | # Generate coefficients using the precision of the target value. |
---|
339 | n = 5 |
---|
340 | cinv = [3991680, -45360, 840, -30, 3] |
---|
341 | three = x.dtype.type(3) |
---|
342 | p = three/np.array(cinv, x.dtype) |
---|
343 | return np.polyval(p[-n:], x*x) |
---|
344 | add_function( |
---|
345 | name="3 j1(x)/x: taylor", |
---|
346 | mp_function=lambda x: 3*(mp.sin(x)/x - mp.cos(x))/(x*x), |
---|
347 | np_function=taylor_3j1x_x, |
---|
348 | ocl_function=make_ocl("return sas_3j1x_x(q);", "sas_j1c", ["lib/sas_3j1x_x.c"]), |
---|
349 | ) |
---|
350 | def trig_3j1x_x(x): |
---|
351 | r""" |
---|
352 | Direct calculation using linear combination of sin/cos. |
---|
353 | |
---|
354 | Use the following trig identity: |
---|
355 | |
---|
356 | .. math:: |
---|
357 | |
---|
358 | a \sin(x) + b \cos(x) = c \sin(x + \phi) |
---|
359 | |
---|
360 | where $c = \surd(a^2+b^2)$ and $\phi = \tan^{-1}(b/a) to calculate the |
---|
361 | numerator $\sin(x) - x\cos(x)$. |
---|
362 | """ |
---|
363 | one = x.dtype.type(1) |
---|
364 | three = x.dtype.type(3) |
---|
365 | c = np.sqrt(one + x*x) |
---|
366 | phi = np.arctan2(-x, one) |
---|
367 | return three*(c*np.sin(x+phi))/(x*x*x) |
---|
368 | add_function( |
---|
369 | name="3 j1(x)/x: trig", |
---|
370 | mp_function=lambda x: 3*(mp.sin(x)/x - mp.cos(x))/(x*x), |
---|
371 | np_function=trig_3j1x_x, |
---|
372 | ocl_function=make_ocl("return sas_3j1x_x(q);", "sas_j1c", ["lib/sas_3j1x_x.c"]), |
---|
373 | ) |
---|
374 | def np_2J1x_x(x): |
---|
375 | """ |
---|
376 | numpy implementation of 2J1(x)/x using single precision algorithm |
---|
377 | """ |
---|
378 | # pylint: disable=bad-continuation |
---|
379 | f = x.dtype.type |
---|
380 | ax = abs(x) |
---|
381 | if ax < f(8.0): |
---|
382 | y = x*x |
---|
383 | ans1 = f(2)*(f(72362614232.0) |
---|
384 | + y*(f(-7895059235.0) |
---|
385 | + y*(f(242396853.1) |
---|
386 | + y*(f(-2972611.439) |
---|
387 | + y*(f(15704.48260) |
---|
388 | + y*(f(-30.16036606))))))) |
---|
389 | ans2 = (f(144725228442.0) |
---|
390 | + y*(f(2300535178.0) |
---|
391 | + y*(f(18583304.74) |
---|
392 | + y*(f(99447.43394) |
---|
393 | + y*(f(376.9991397) |
---|
394 | + y))))) |
---|
395 | return ans1/ans2 |
---|
396 | else: |
---|
397 | y = f(64.0)/(ax*ax) |
---|
398 | xx = ax - f(2.356194491) |
---|
399 | ans1 = (f(1.0) |
---|
400 | + y*(f(0.183105e-2) |
---|
401 | + y*(f(-0.3516396496e-4) |
---|
402 | + y*(f(0.2457520174e-5) |
---|
403 | + y*f(-0.240337019e-6))))) |
---|
404 | ans2 = (f(0.04687499995) |
---|
405 | + y*(f(-0.2002690873e-3) |
---|
406 | + y*(f(0.8449199096e-5) |
---|
407 | + y*(f(-0.88228987e-6) |
---|
408 | + y*f(0.105787412e-6))))) |
---|
409 | sn, cn = np.sin(xx), np.cos(xx) |
---|
410 | ans = np.sqrt(f(0.636619772)/ax) * (cn*ans1 - (f(8.0)/ax)*sn*ans2) * f(2)/x |
---|
411 | return -ans if (x < f(0.0)) else ans |
---|
412 | add_function( |
---|
413 | name="2 J1(x)/x:alt", |
---|
414 | mp_function=lambda x: 2*mp.j1(x)/x, |
---|
415 | np_function=lambda x: np.asarray([np_2J1x_x(v) for v in x], x.dtype), |
---|
416 | ocl_function=make_ocl("return sas_2J1x_x(q);", "sas_2J1x_x", ["lib/polevl.c", "lib/sas_J1.c"]), |
---|
417 | ) |
---|
418 | |
---|
419 | ALL_FUNCTIONS = set(FUNCTIONS.keys()) |
---|
420 | ALL_FUNCTIONS.discard("loggamma") # OCL version not ready yet |
---|
421 | ALL_FUNCTIONS.discard("3j1/x:taylor") |
---|
422 | ALL_FUNCTIONS.discard("3j1/x:trig") |
---|
423 | ALL_FUNCTIONS.discard("2J1/x:alt") |
---|
424 | |
---|
425 | # =============== MAIN PROGRAM ================ |
---|
426 | |
---|
427 | def usage(): |
---|
428 | names = ", ".join(sorted(ALL_FUNCTIONS)) |
---|
429 | print("""\ |
---|
430 | usage: precision.py [-f] [--log|--linear|--zoom|--neg] name... |
---|
431 | where |
---|
432 | -f indicates that the function value should be plotted rather than error, |
---|
433 | --log indicates log stepping in [10^-3, 10^5] |
---|
434 | --linear indicates linear stepping in [1, 1000] |
---|
435 | --zoom indicates linear stepping in [1000, 1010] |
---|
436 | --neg indicates linear stepping in [-100.1, 100.1] |
---|
437 | and name is "all [first]" or one of: |
---|
438 | """+names) |
---|
439 | sys.exit(1) |
---|
440 | |
---|
441 | def main(): |
---|
442 | import sys |
---|
443 | diff = True |
---|
444 | xrange = "log" |
---|
445 | args = sys.argv[1:] |
---|
446 | if '-f' in args: |
---|
447 | args.remove('-f') |
---|
448 | diff = False |
---|
449 | for k in "log linear zoom neg".split(): |
---|
450 | if '--'+k in args: |
---|
451 | args.remove('--'+k) |
---|
452 | xrange = k |
---|
453 | if not args: |
---|
454 | usage() |
---|
455 | if args[0] == "all": |
---|
456 | cutoff = args[1] if len(args) > 1 else "" |
---|
457 | args = list(sorted(ALL_FUNCTIONS)) |
---|
458 | args = [k for k in args if k >= cutoff] |
---|
459 | if any(k not in FUNCTIONS for k in args): |
---|
460 | usage() |
---|
461 | multiple = len(args) > 1 |
---|
462 | pylab.interactive(multiple) |
---|
463 | for k in args: |
---|
464 | pylab.clf() |
---|
465 | comparator = FUNCTIONS[k] |
---|
466 | comparator.run(xrange=xrange, diff=diff) |
---|
467 | if multiple: |
---|
468 | raw_input() |
---|
469 | if not multiple: |
---|
470 | pylab.show() |
---|
471 | |
---|
472 | if __name__ == "__main__": |
---|
473 | main() |
---|