correlated_fields.py 16 KB
Newer Older
Martin Reinecke's avatar
Martin Reinecke committed
1
2
3
4
5
6
7
8
9
10
11
12
13
# This program is free software: you can redistribute it and/or modify
# it under the terms of the GNU General Public License as published by
# the Free Software Foundation, either version 3 of the License, or
# (at your option) any later version.
#
# This program is distributed in the hope that it will be useful,
# but WITHOUT ANY WARRANTY; without even the implied warranty of
# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
# GNU General Public License for more details.
#
# You should have received a copy of the GNU General Public License
# along with this program.  If not, see <http://www.gnu.org/licenses/>.
#
14
# Copyright(C) 2013-2019 Max-Planck-Society
Philipp Arras's avatar
Philipp Arras committed
15
# Authors: Philipp Frank, Philipp Arras
Martin Reinecke's avatar
Martin Reinecke committed
16
#
17
# NIFTy is being developed at the Max-Planck-Institut fuer Astrophysik.
Philipp Arras's avatar
Philipp Arras committed
18

Philipp Arras's avatar
Philipp Arras committed
19
import numpy as np
20

Philipp Arras's avatar
Philipp Arras committed
21
from ..domain_tuple import DomainTuple
Philipp Arras's avatar
Philipp Arras committed
22
23
from ..domains.power_space import PowerSpace
from ..domains.unstructured_domain import UnstructuredDomain
24
from ..field import Field
Philipp Arras's avatar
Philipp Arras committed
25
from ..operators.adder import Adder
26
from ..operators.contraction_operator import ContractionOperator
Philipp Arras's avatar
Philipp Arras committed
27
from ..operators.distributors import PowerDistributor
Philipp Arras's avatar
Philipp Arras committed
28
from ..operators.endomorphic_operator import EndomorphicOperator
Martin Reinecke's avatar
Martin Reinecke committed
29
from ..operators.harmonic_operators import HarmonicTransformOperator
Philipp Arras's avatar
Philipp Arras committed
30
from ..operators.linear_operator import LinearOperator
Philipp Arras's avatar
Philipp Arras committed
31
32
from ..operators.operator import Operator
from ..operators.simple_linear_operators import VdotOperator, ducktape
Philipp Arras's avatar
Philipp Arras committed
33
from ..operators.value_inserter import ValueInserter
34
from ..probing import StatCalculator
Philipp Arras's avatar
Philipp Arras committed
35
from ..sugar import from_global_data, full, makeDomain
36

Philipp Arras's avatar
Philipp Arras committed
37

38
39
def _lognormal_moments(mean, sig):
    mean, sig = float(mean), float(sig)
Philipp Arras's avatar
Philipp Arras committed
40
41
42
    assert sig > 0
    logsig = np.sqrt(np.log((sig/mean)**2 + 1))
    logmean = np.log(mean) - logsig**2/2
43
44
    return logmean, logsig

Philipp Arras's avatar
Philipp Arras committed
45

Philipp Arras's avatar
Philipp Arras committed
46
47
48
49
50
def _normal(mean, sig, key):
    return Adder(Field.scalar(mean)) @ (
        sig*ducktape(DomainTuple.scalar_domain(), None, key))


Philipp Arras's avatar
Philipp Arras committed
51
def _log_k_lengths(pspace):
Philipp Arras's avatar
Philipp Arras committed
52
    """Log(k_lengths) without zeromode"""
Philipp Arras's avatar
Philipp Arras committed
53
54
55
    return np.log(pspace.k_lengths[1:])


Philipp Arras's avatar
Philipp Arras committed
56
def _relative_log_k_lengths(power_space):
Philipp Arras's avatar
Philipp Arras committed
57
58
    """Log-distance to first bin
    logkl.shape==power_space.shape, logkl[0]=logkl[1]=0"""
Philipp Arras's avatar
Philipp Arras committed
59
60
61
62
63
64
    power_space = DomainTuple.make(power_space)
    assert isinstance(power_space[0], PowerSpace)
    assert len(power_space) == 1
    logkl = _log_k_lengths(power_space[0])
    assert logkl.shape[0] == power_space[0].shape[0] - 1
    logkl -= logkl[0]
Philipp Arras's avatar
Philipp Arras committed
65
    return np.insert(logkl, 0, 0)
Philipp Arras's avatar
Philipp Arras committed
66
67


Philipp Arras's avatar
Philipp Arras committed
68
69
70
71
72
73
74
75
def _log_vol(power_space):
    power_space = DomainTuple.make(power_space)
    assert isinstance(power_space[0], PowerSpace)
    assert len(power_space) == 1
    logk_lengths = _log_k_lengths(power_space[0])
    return logk_lengths[1:] - logk_lengths[:-1]


Philipp Arras's avatar
Philipp Arras committed
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
class _LognormalMomentMatching(Operator):
    def __init__(self, mean, sig, key):
        key = str(key)
        logmean, logsig = _lognormal_moments(mean, sig)
        self._mean = mean
        self._sig = sig
        op = _normal(logmean, logsig, key).exp()
        self._domain, self._target = op.domain, op.target
        self.apply = op.apply

    @property
    def mean(self):
        return self._mean

    @property
    def std(self):
        return self._sig


Philipp Frank's avatar
Philipp Frank committed
95
class _SlopeRemover(EndomorphicOperator):
Philipp Arras's avatar
Philipp Arras committed
96
    def __init__(self, domain):
Philipp Frank's avatar
Philipp Frank committed
97
        self._domain = makeDomain(domain)
Philipp Arras's avatar
Philipp Arras committed
98
99
        assert len(self._domain) == 1
        assert isinstance(self._domain[0], PowerSpace)
Philipp Arras's avatar
Philipp Arras committed
100
        logkl = _relative_log_k_lengths(self._domain)
101
        self._sc = logkl/float(logkl[-1])
Philipp Frank's avatar
Philipp Frank committed
102
        self._capability = self.TIMES | self.ADJOINT_TIMES
Philipp Arras's avatar
Philipp Arras committed
103

104
105
    def apply(self, x, mode):
        self._check_input(x, mode)
Philipp Frank's avatar
Philipp Frank committed
106
107
        x = x.to_global_data()
        if mode == self.TIMES:
108
            res = x - x[-1]*self._sc
Philipp Frank's avatar
Philipp Frank committed
109
        else:
110
            res = np.zeros(x.shape, dtype=x.dtype)
Philipp Frank's avatar
Philipp Frank committed
111
112
            res += x
            res[-1] -= (x*self._sc).sum()
113
        return from_global_data(self._tgt(mode), res)
Philipp Frank's avatar
Philipp Frank committed
114

115

Philipp Arras's avatar
Philipp Arras committed
116
117
118
class _TwoLogIntegrations(LinearOperator):
    def __init__(self, target):
        self._target = makeDomain(target)
Philipp Arras's avatar
Philipp Arras committed
119
120
        assert len(self._target) == 1
        assert isinstance(self._target[0], PowerSpace)
Philipp Arras's avatar
Philipp Arras committed
121
122
123
124
125
        self._domain = makeDomain(
            UnstructuredDomain((2, self.target.shape[0] - 2)))
        self._capability = self.TIMES | self.ADJOINT_TIMES
        if not isinstance(self._target[0], PowerSpace):
            raise TypeError
Philipp Arras's avatar
Philipp Arras committed
126
        self._log_vol = _log_vol(self._target[0])
Philipp Arras's avatar
Philipp Arras committed
127
128
129
130
131
132

    def apply(self, x, mode):
        self._check_input(x, mode)
        if mode == self.TIMES:
            x = x.to_global_data()
            res = np.empty(self._target.shape)
Philipp Arras's avatar
Philipp Arras committed
133
            res[0] = res[1] = 0
Philipp Arras's avatar
Philipp Arras committed
134
            res[2:] = np.cumsum(x[1])
Philipp Arras's avatar
Philipp Arras committed
135
            res[2:] = (res[2:] + res[1:-1])/2*self._log_vol + x[0]
Philipp Arras's avatar
Philipp Arras committed
136
137
138
139
140
141
142
            res[2:] = np.cumsum(res[2:])
            return from_global_data(self._target, res)
        else:
            x = x.to_global_data_rw()
            res = np.zeros(self._domain.shape)
            x[2:] = np.cumsum(x[2:][::-1])[::-1]
            res[0] += x[2:]
Philipp Arras's avatar
Philipp Arras committed
143
            x[2:] *= self._log_vol/2.
144
145
            x[1:-1] += x[2:]
            res[1] += np.cumsum(x[2:][::-1])[::-1]
Philipp Arras's avatar
Philipp Arras committed
146
147
148
149
150
151
            return from_global_data(self._domain, res)


class _Normalization(Operator):
    def __init__(self, domain):
        self._domain = self._target = makeDomain(domain)
Philipp Arras's avatar
Philipp Arras committed
152
        assert len(self._domain) == 1
Philipp Arras's avatar
Philipp Arras committed
153
        assert isinstance(self._domain[0], PowerSpace)
Philipp Arras's avatar
Philipp Arras committed
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
        hspace = self._domain[0].harmonic_partner
        pd = PowerDistributor(hspace, power_space=self._domain[0])
        cst = pd.adjoint(full(pd.target, 1.)).to_global_data_rw()
        cst[0] = 0
        self._cst = from_global_data(self._domain, cst)
        self._specsum = _SpecialSum(self._domain)

    def apply(self, x):
        self._check_input(x)
        amp = x.exp()
        spec = (2*x).exp()
        # FIXME This normalizes also the zeromode which is supposed to be left
        # untouched by this operator
        return self._specsum(self._cst*spec)**(-0.5)*amp


class _SpecialSum(EndomorphicOperator):
    def __init__(self, domain):
        self._domain = makeDomain(domain)
Philipp Arras's avatar
Philipp Arras committed
173
        assert len(self._domain) == 1
Philipp Arras's avatar
Philipp Arras committed
174
175
176
177
178
179
180
        self._capability = self.TIMES | self.ADJOINT_TIMES

    def apply(self, x, mode):
        self._check_input(x, mode)
        return full(self._tgt(mode), x.sum())


181
182
183
class _Amplitude(Operator):
    def __init__(self, target, fluctuations, flexibility, asperity,
                 loglogavgslope, key):
Philipp Arras's avatar
Philipp Arras committed
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
        """
        fluctuations > 0
        flexibility > 0
        asperity > 0
        loglogavgslope probably negative
        """
        assert isinstance(fluctuations, Operator)
        assert isinstance(flexibility, Operator)
        assert isinstance(asperity, Operator)
        assert isinstance(loglogavgslope, Operator)
        target = makeDomain(target)
        assert len(target) == 1
        assert isinstance(target[0], PowerSpace)

        twolog = _TwoLogIntegrations(target)
Philipp Arras's avatar
Philipp Arras committed
199
200
201
202
203
204
205
        dom = twolog.domain
        shp = dom.shape
        totvol = target[0].harmonic_partner.get_default_codomain().total_volume

        # Prepare constant fields
        foo = np.zeros(shp)
        foo[0] = foo[1] = np.sqrt(_log_vol(target))
Philipp Arras's avatar
Philipp Arras committed
206
        vflex = from_global_data(dom, foo)
Philipp Arras's avatar
Philipp Arras committed
207
208
209

        foo = np.zeros(shp, dtype=np.float64)
        foo[0] += 1
Philipp Arras's avatar
Philipp Arras committed
210
        vasp = from_global_data(dom, foo)
Philipp Arras's avatar
Philipp Arras committed
211
212
213
214
215

        foo = np.ones(shp)
        foo[0] = _log_vol(target)**2/12.
        shift = from_global_data(dom, foo)

Philipp Arras's avatar
Philipp Arras committed
216
        vslope = from_global_data(target, _relative_log_k_lengths(target))
Philipp Arras's avatar
Philipp Arras committed
217

Philipp Frank's avatar
fixup    
Philipp Frank committed
218
        foo, bar = [np.zeros(target.shape) for _ in range(2)]
Philipp Arras's avatar
Philipp Arras committed
219
220
        bar[1:] = foo[0] = totvol
        vol0, vol1 = [from_global_data(target, aa) for aa in (foo, bar)]
Philipp Arras's avatar
Philipp Arras committed
221
222
        # End prepare constant fields

Philipp Arras's avatar
Philipp Arras committed
223
224
225
        slope = VdotOperator(vslope).adjoint @ loglogavgslope
        sig_flex = VdotOperator(vflex).adjoint @ flexibility
        sig_asp = VdotOperator(vasp).adjoint @ asperity
Philipp Arras's avatar
Philipp Arras committed
226
227
228
        sig_fluc = VdotOperator(vol1).adjoint @ fluctuations

        xi = ducktape(dom, None, key)
Philipp Arras's avatar
Philipp Arras committed
229
        sigma = sig_flex*(Adder(shift) @ sig_asp).sqrt()
Philipp Arras's avatar
Philipp Arras committed
230
231
232
233
        smooth = _SlopeRemover(target) @ twolog @ (sigma*xi)
        op = _Normalization(target) @ (slope + smooth)
        op = Adder(vol0) @ (sig_fluc*op)

Philipp Arras's avatar
Philipp Arras committed
234
        self.apply = op.apply
Philipp Arras's avatar
Philipp Arras committed
235
        self._fluc = fluctuations
Philipp Arras's avatar
Philipp Arras committed
236
        self._domain, self._target = op.domain, op.target
Philipp Arras's avatar
Philipp Arras committed
237

Philipp Arras's avatar
Philipp Arras committed
238
239
240
241
    @property
    def fluctuation_amplitude(self):
        return self._fluc

242
243
244
245

class CorrelatedFieldMaker:
    def __init__(self):
        self._a = []
246
        self._azm = None
247
        self._position_spaces = []
248
249

    def add_fluctuations(self,
250
                         position_space,
251
252
253
254
255
256
257
258
                         fluctuations_mean,
                         fluctuations_stddev,
                         flexibility_mean,
                         flexibility_stddev,
                         asperity_mean,
                         asperity_stddev,
                         loglogavgslope_mean,
                         loglogavgslope_stddev,
259
                         prefix='',
Philipp Arras's avatar
Philipp Arras committed
260
                         index=None):
Philipp Arras's avatar
Philipp Arras committed
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
        fluctuations_mean = float(fluctuations_mean)
        fluctuations_stddev = float(fluctuations_stddev)
        flexibility_mean = float(flexibility_mean)
        flexibility_stddev = float(flexibility_stddev)
        asperity_mean = float(asperity_mean)
        asperity_stddev = float(asperity_stddev)
        loglogavgslope_mean = float(loglogavgslope_mean)
        loglogavgslope_stddev = float(loglogavgslope_stddev)
        prefix = str(prefix)
        assert fluctuations_stddev > 0
        assert fluctuations_mean > 0
        assert flexibility_stddev > 0
        assert flexibility_mean > 0
        assert asperity_stddev > 0
        assert asperity_mean > 0
        assert loglogavgslope_stddev > 0

Philipp Arras's avatar
Philipp Arras committed
278
279
280
281
282
283
284
        fluct = _LognormalMomentMatching(fluctuations_mean,
                                         fluctuations_stddev,
                                         prefix + 'fluctuations')
        flex = _LognormalMomentMatching(flexibility_mean, flexibility_stddev,
                                        prefix + 'flexibility')
        asp = _LognormalMomentMatching(asperity_mean, asperity_stddev,
                                       prefix + 'asperity')
285
        avgsl = _normal(loglogavgslope_mean, loglogavgslope_stddev,
Philipp Arras's avatar
Philipp Arras committed
286
                        prefix + 'loglogavgslope')
287
288
        amp = _Amplitude(PowerSpace(position_space.get_default_codomain()),
                         fluct, flex, asp, avgsl, prefix + 'spectrum')
289
290
        if index is not None:
            self._a.insert(index, amp)
291
            self._position_spaces.insert(index, position_space)
292
293
        else:
            self._a.append(amp)
294
            self._position_spaces.append(position_space)
295
296
297

    def finalize_from_op(self, zeromode, prefix=''):
        assert isinstance(zeromode, Operator)
298
        self._azm = zeromode
299
300
        hspace = makeDomain([dd.get_default_codomain()
                             for dd in self._position_spaces])
301
302
303
304
305
306
307
        foo = np.ones(hspace.shape)
        zeroind = len(hspace.shape)*(0,)
        foo[zeroind] = 0
        azm = Adder(from_global_data(hspace, foo)) @ ValueInserter(
            hspace, zeroind) @ zeromode

        n_amplitudes = len(self._a)
308
309
        ht = HarmonicTransformOperator(hspace, self._position_spaces[0],
                                       space=0)
310
        for i in range(1, n_amplitudes):
311
312
313
            ht = (HarmonicTransformOperator(ht.target,
                                            self._position_spaces[i],
                                            space=i) @ ht)
314
315
316
317
318
319
320
321
322
323
324

        pd = PowerDistributor(hspace, self._a[0].target[0], 0)
        for i in range(1, n_amplitudes):
            foo = PowerDistributor(pd.domain, self._a[i].target[0], space=i)
            pd = pd @ foo

        spaces = tuple(range(n_amplitudes))
        a = ContractionOperator(pd.domain, spaces[1:]).adjoint @ self._a[0]
        for i in range(1, n_amplitudes):
            co = ContractionOperator(pd.domain, spaces[:i] + spaces[(i + 1):])
            a = a*(co.adjoint @ self._a[i])
Philipp Arras's avatar
Philipp Arras committed
325

326
        return ht(azm*(pd @ a)*ducktape(hspace, None, prefix + 'xi'))
Philipp Arras's avatar
Philipp Arras committed
327
328
329
330

    def finalize(self,
                 offset_amplitude_mean,
                 offset_amplitude_stddev,
331
                 prefix='',
Philipp Arras's avatar
Philipp Arras committed
332
333
334
335
336
337
338
339
340
                 offset=None):
        """
        offset vs zeromode: volume factor
        """
        offset_amplitude_stddev = float(offset_amplitude_stddev)
        offset_amplitude_mean = float(offset_amplitude_mean)
        assert offset_amplitude_stddev > 0
        assert offset_amplitude_mean > 0
        if offset is not None:
341
            raise NotImplementedError
Philipp Arras's avatar
Philipp Arras committed
342
            offset = float(offset)
Philipp Arras's avatar
Philipp Arras committed
343
344
345
        azm = _LognormalMomentMatching(offset_amplitude_mean,
                                       offset_amplitude_stddev,
                                       prefix + 'zeromode')
346
        return self.finalize_from_op(azm, prefix)
Philipp Arras's avatar
Philipp Arras committed
347
348
349

    @property
    def amplitudes(self):
350
        return self._a
351

352
353
354
355
356
357
358
    @property
    def amplitude_total_offset(self):
        return self._azm

    @property
    def total_fluctuation(self):
        if len(self._a) == 0:
Philipp Arras's avatar
Philipp Arras committed
359
            raise NotImplementedError
360
361
362
363
364
        if len(self._a) == 1:
            return self._a[0].fluctuation_amplitude
        q = 1.
        for a in self._a:
            fl = a.fluctuation_amplitude
Philipp Arras's avatar
Philipp Arras committed
365
366
            q = q*(Adder(full(fl.target, 1.)) @ fl**2)
        return (Adder(full(q.target, -1.)) @ q).sqrt()
367

Philipp Arras's avatar
Philipp Arras committed
368
    def slice_fluctuation(self, space):
369
        if len(self._a) == 0:
Philipp Arras's avatar
Philipp Arras committed
370
            raise NotImplementedError
371
372
373
374
375
376
377
        assert space < len(self._a)
        if len(self._a) == 1:
            return self._a[0].fluctuation_amplitude
        q = 1.
        for j in range(len(self._a)):
            fl = self._a[j].fluctuation_amplitude
            if j == space:
Philipp Arras's avatar
Philipp Arras committed
378
                q = q*fl**2
379
            else:
Philipp Arras's avatar
Philipp Arras committed
380
                q = q*(Adder(full(fl.target, 1.)) @ fl**2)
381
        return q.sqrt()
Philipp Arras's avatar
Philipp Arras committed
382
383

    def average_fluctuation(self, space):
384
        if len(self._a) == 0:
Philipp Arras's avatar
Philipp Arras committed
385
            raise NotImplementedError
386
387
388
389
390
        assert space < len(self._a)
        if len(self._a) == 1:
            return self._a[0].fluctuation_amplitude
        return self._a[space].fluctuation_amplitude

Philipp Arras's avatar
Philipp Arras committed
391
    def average_fluctuation_realized(self, samples, space):
392
393
394
395
        ldom = len(samples[0].domain)
        assert space < ldom
        if ldom == 1:
            return self.total_fluctuation_realized(samples)
Philipp Arras's avatar
Philipp Arras committed
396
        spaces = ()
397
398
399
400
401
402
        for i in range(ldom):
            if i != space:
                spaces += (i,)
        res = 0.
        for s in samples:
            r = s.mean(spaces)
Philipp Arras's avatar
Philipp Arras committed
403
            res = res + (r - r.mean())**2
404
405
        res = res/len(samples)
        return np.sqrt(res.mean())
Philipp Arras's avatar
Philipp Arras committed
406
407

    def slice_fluctuation_realized(self, samples, space):
408
409
410
411
412
413
414
415
416
417
418
        ldom = len(samples[0].domain)
        assert space < ldom
        if ldom == 1:
            return self.total_fluctuation_realized(samples)
        res1 = 0.
        res2 = 0.
        for s in samples:
            res1 = res1 + s**2
            res2 = res2 + s.mean(space)**2
        res1 = res1/len(samples)
        res2 = res2/len(samples)
Philipp Arras's avatar
Philipp Arras committed
419
        res = res1.mean() - res2.mean()
420
421
        return np.sqrt(res)

Philipp Arras's avatar
Philipp Arras committed
422
    def moment_slice_to_average(self, fluctuations_slice_mean, nsamples=1000):
423
424
425
426
427
        fluctuations_slice_mean = float(fluctuations_slice_mean)
        assert fluctuations_slice_mean > 0
        scm = 1.
        for a in self._a:
            m, std = a.fluctuation_amplitude.mean, a.fluctuation_amplitude.std
Philipp Arras's avatar
Philipp Arras committed
428
429
            mu, sig = _lognormal_moments(m, std)
            flm = np.exp(mu + sig*np.random.normal(size=nsamples))
430
431
            scm *= flm**2 + 1.
        scm = np.mean(np.sqrt(scm))
Philipp Arras's avatar
Philipp Arras committed
432
        return fluctuations_slice_mean/scm
Philipp Arras's avatar
Philipp Arras committed
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454

    @staticmethod
    def offset_amplitude_realized(samples):
        res = 0.
        for s in samples:
            res += s.mean()**2
        return np.sqrt(res/len(samples))

    @staticmethod
    def total_fluctuation_realized(samples):
        res = 0.
        for s in samples:
            res = res + (s - s.mean())**2
        res = res/len(samples)
        return np.sqrt(res.mean())

    @staticmethod
    def stats(op, samples):
        sc = StatCalculator()
        for s in samples:
            sc.add(op(s.extract(op.domain)))
        return sc.mean.to_global_data(), sc.var.sqrt().to_global_data()