mcmc_based_searches.py 75.2 KB
Newer Older
Gregory Ashton's avatar
Gregory Ashton committed
1
2
""" Searches using MCMC-based methods """

3
import sys
Gregory Ashton's avatar
Gregory Ashton committed
4
import os
5
import copy
Gregory Ashton's avatar
Gregory Ashton committed
6
import logging
7
from collections import OrderedDict
8
9
10
11
12

import numpy as np
import matplotlib
import matplotlib.pyplot as plt
import emcee
13
import pymc3
14
15
16
import corner
import dill as pickle

17
18
from core import BaseSearchClass, ComputeFstat, SemiCoherentSearch
from optimal_setup_functions import get_V_estimate
Gregory Ashton's avatar
Gregory Ashton committed
19
20
21
from core import tqdm, args, earth_ephem, sun_ephem
from optimal_setup_functions import get_optimal_setup
import helper_functions
22
23


Gregory Ashton's avatar
Gregory Ashton committed
24
25
class MCMCSearch(BaseSearchClass):
    """ MCMC search using ComputeFstat"""
Gregory Ashton's avatar
Gregory Ashton committed
26
    @helper_functions.initializer
Gregory Ashton's avatar
Gregory Ashton committed
27
28
    def __init__(self, label, outdir, theta_prior, tref, minStartTime,
                 maxStartTime, sftfilepath=None, nsteps=[100, 100],
29
30
                 nwalkers=100, ntemps=1, log10temperature_min=-5,
                 theta_initial=None, scatter_val=1e-10,
31
                 binary=False, BSGL=False, minCoverFreq=None,
32
                 maxCoverFreq=None, detectors=None, earth_ephem=None,
33
                 sun_ephem=None, injectSources=None, assumeSqrtSX=None):
34
35
36
37
        """
        Parameters
        label, outdir: str
            A label and directory to read/write data from/to
38
39
        sftfilepath: str
            File patern to match SFTs
40
        theta_prior: dict
41
42
43
44
            Dictionary of priors and fixed values for the search parameters.
            For each parameters (key of the dict), if it is to be held fixed
            the value should be the constant float, if it is be searched, the
            value should be a dictionary of the prior.
45
46
47
48
        theta_initial: dict, array, (None)
            Either a dictionary of distribution about which to distribute the
            initial walkers about, an array (from which the walkers will be
            scattered by scatter_val, or  None in which case the prior is used.
49
        tref, minStartTime, maxStartTime: int
50
51
52
53
54
55
            GPS seconds of the reference time, start time and end time
        nsteps: list (m,)
            List specifying the number of steps to take, the last two entries
            give the nburn and nprod of the 'production' run, all entries
            before are for iterative initialisation steps (usually just one)
            e.g. [1000, 1000, 500].
56
57
58
59
60
61
62
63
        nwalkers, ntemps: int,
            The number of walkers and temperates to use in the parallel
            tempered PTSampler.
        log10temperature_min float < 0
            The  log_10(tmin) value, the set of betas passed to PTSampler are
            generated from np.logspace(0, log10temperature_min, ntemps).
        binary: Bool
            If true, search over binary parameters
64
        detectors: str
65
66
            Two character reference to the data to use, specify None for no
            contraint.
67
68
69
70
71
72
73
74
75
76
        minCoverFreq, maxCoverFreq: float
            Minimum and maximum instantaneous frequency which will be covered
            over the SFT time span as passed to CreateFstatInput
        earth_ephem, sun_ephem: str
            Paths of the two files containing positions of Earth and Sun,
            respectively at evenly spaced times, as passed to CreateFstatInput
            If None defaults defined in BaseSearchClass will be used

        """

Gregory Ashton's avatar
Gregory Ashton committed
77
78
        if os.path.isdir(outdir) is False:
            os.mkdir(outdir)
79
        self.add_log_file()
Gregory Ashton's avatar
Gregory Ashton committed
80
81
        logging.info(
            'Set-up MCMC search for model {} on data {}'.format(
82
                self.label, self.sftfilepath))
83
84
85
        self.pickle_path = '{}/{}_saved_data.p'.format(self.outdir, self.label)
        self.unpack_input_theta()
        self.ndim = len(self.theta_keys)
86
87
88
89
        if self.log10temperature_min:
            self.betas = np.logspace(0, self.log10temperature_min, self.ntemps)
        else:
            self.betas = None
90

91
92
93
94
95
96
97
98
        if earth_ephem is None:
            self.earth_ephem = self.earth_ephem_default
        if sun_ephem is None:
            self.sun_ephem = self.sun_ephem_default

        if args.clean and os.path.isfile(self.pickle_path):
            os.rename(self.pickle_path, self.pickle_path+".old")

99
100
101
        self.log_input()

    def log_input(self):
102
        logging.info('theta_prior = {}'.format(self.theta_prior))
103
        logging.info('nwalkers={}'.format(self.nwalkers))
104
105
106
107
        logging.info('scatter_val = {}'.format(self.scatter_val))
        logging.info('nsteps = {}'.format(self.nsteps))
        logging.info('ntemps = {}'.format(self.ntemps))
        logging.info('log10temperature_min = {}'.format(
108
            self.log10temperature_min))
109
110
111

    def inititate_search_object(self):
        logging.info('Setting up search object')
Gregory Ashton's avatar
Gregory Ashton committed
112
        self.search = ComputeFstat(
113
114
115
            tref=self.tref, sftfilepath=self.sftfilepath,
            minCoverFreq=self.minCoverFreq, maxCoverFreq=self.maxCoverFreq,
            earth_ephem=self.earth_ephem, sun_ephem=self.sun_ephem,
116
            detectors=self.detectors, BSGL=self.BSGL, transient=False,
117
            minStartTime=self.minStartTime, maxStartTime=self.maxStartTime,
118
119
            binary=self.binary, injectSources=self.injectSources,
            assumeSqrtSX=self.assumeSqrtSX)
120
121

    def logp(self, theta_vals, theta_prior, theta_keys, search):
Gregory Ashton's avatar
Gregory Ashton committed
122
        H = [self.generic_lnprior(**theta_prior[key])(p) for p, key in
123
124
125
126
127
128
             zip(theta_vals, theta_keys)]
        return np.sum(H)

    def logl(self, theta, search):
        for j, theta_i in enumerate(self.theta_idxs):
            self.fixed_theta[theta_i] = theta[j]
129
130
        FS = search.compute_fullycoherent_det_stat_single_point(
            *self.fixed_theta)
131
132
133
        return FS

    def unpack_input_theta(self):
134
        full_theta_keys = ['F0', 'F1', 'F2', 'Alpha', 'Delta']
135
136
137
        if self.binary:
            full_theta_keys += [
                'asini', 'period', 'ecc', 'tp', 'argp']
138
139
        full_theta_keys_copy = copy.copy(full_theta_keys)

140
141
        full_theta_symbols = ['$f$', '$\dot{f}$', '$\ddot{f}$', r'$\alpha$',
                              r'$\delta$']
142
143
144
145
        if self.binary:
            full_theta_symbols += [
                'asini', 'period', 'period', 'ecc', 'tp', 'argp']

146
147
        self.theta_keys = []
        fixed_theta_dict = {}
148
        for key, val in self.theta_prior.iteritems():
149
150
            if type(val) is dict:
                fixed_theta_dict[key] = 0
Gregory Ashton's avatar
Gregory Ashton committed
151
                self.theta_keys.append(key)
152
153
154
155
156
157
            elif type(val) in [float, int, np.float64]:
                fixed_theta_dict[key] = val
            else:
                raise ValueError(
                    'Type {} of {} in theta not recognised'.format(
                        type(val), key))
Gregory Ashton's avatar
Gregory Ashton committed
158
            full_theta_keys_copy.pop(full_theta_keys_copy.index(key))
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174

        if len(full_theta_keys_copy) > 0:
            raise ValueError(('Input dictionary `theta` is missing the'
                              'following keys: {}').format(
                                  full_theta_keys_copy))

        self.fixed_theta = [fixed_theta_dict[key] for key in full_theta_keys]
        self.theta_idxs = [full_theta_keys.index(k) for k in self.theta_keys]
        self.theta_symbols = [full_theta_symbols[i] for i in self.theta_idxs]

        idxs = np.argsort(self.theta_idxs)
        self.theta_idxs = [self.theta_idxs[i] for i in idxs]
        self.theta_symbols = [self.theta_symbols[i] for i in idxs]
        self.theta_keys = [self.theta_keys[i] for i in idxs]

    def check_initial_points(self, p0):
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
        for nt in range(self.ntemps):
            logging.info('Checking temperature {} chains'.format(nt))
            initial_priors = np.array([
                self.logp(p, self.theta_prior, self.theta_keys, self.search)
                for p in p0[nt]])
            number_of_initial_out_of_bounds = sum(initial_priors == -np.inf)

            if number_of_initial_out_of_bounds > 0:
                logging.warning(
                    'Of {} initial values, {} are -np.inf due to the prior'
                    .format(len(initial_priors),
                            number_of_initial_out_of_bounds))

                p0 = self.generate_new_p0_to_fix_initial_points(
                    p0, nt, initial_priors)

    def generate_new_p0_to_fix_initial_points(self, p0, nt, initial_priors):
        logging.info('Attempting to correct intial values')
        idxs = np.arange(self.nwalkers)[initial_priors == -np.inf]
        count = 0
        while sum(initial_priors == -np.inf) > 0 and count < 100:
            for j in idxs:
                p0[nt][j] = (p0[nt][np.random.randint(0, self.nwalkers)]*(
                             1+np.random.normal(0, 1e-10, self.ndim)))
            initial_priors = np.array([
                self.logp(p, self.theta_prior, self.theta_keys,
                          self.search)
                for p in p0[nt]])
            count += 1

        if sum(initial_priors == -np.inf) > 0:
            logging.info('Failed to fix initial priors')
        else:
            logging.info('Suceeded to fix initial priors')

        return p0
211

212
    def OLD_run_sampler_with_progress_bar(self, sampler, ns, p0):
213
214
        for result in tqdm(sampler.sample(p0, iterations=ns), total=ns):
            pass
Gregory Ashton's avatar
Gregory Ashton committed
215
216
        return sampler

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
    #def run_sampler(self, sampler, ns, p0):
    #    convergence_period = 200
    #    convergence_diagnostic = []
    #    convergence_diagnosticx = []
    #    for i, result in enumerate(tqdm(
    #            sampler.sample(p0, iterations=ns), total=ns)):
    #        if np.mod(i+1, convergence_period) == 0:
    #            s = sampler.chain[0, :, i-convergence_period+1:i+1, :]
    #            score_per_parameter = []
    #            for j in range(self.ndim):
    #                scores = []
    #                for k in range(self.nwalkers):
    #                    out = pymc3.geweke(
    #                        s[k, :, j].reshape((convergence_period)),
    #                        intervals=2, first=0.4, last=0.4)
    #                    scores.append(out[0][1])
    #                score_per_parameter.append(np.median(scores))
    #            convergence_diagnostic.append(score_per_parameter)
    #            convergence_diagnosticx.append(i - convergence_period/2)
    #    self.convergence_diagnostic = np.array(np.abs(convergence_diagnostic))
    #    self.convergence_diagnosticx = convergence_diagnosticx
    #    return sampler

    #def run_sampler(self, sampler, ns, p0):
    #    convergence_period = 200
    #    convergence_diagnostic = []
    #    convergence_diagnosticx = []
    #    for i, result in enumerate(tqdm(
    #            sampler.sample(p0, iterations=ns), total=ns)):
    #        if np.mod(i+1, convergence_period) == 0:
    #            s = sampler.chain[0, :, i-convergence_period+1:i+1, :]
    #            mean_per_chain = np.mean(s, axis=1)
    #            std_per_chain = np.std(s, axis=1)
    #            mean = np.mean(mean_per_chain, axis=0)
    #            B = convergence_period * np.sum((mean_per_chain - mean)**2, axis=0) / (self.nwalkers - 1)
    #            W = np.sum(std_per_chain**2, axis=0) / self.nwalkers
    #            print B, W
    #            convergence_diagnostic.append(W/B)
    #            convergence_diagnosticx.append(i - convergence_period/2)
    #    self.convergence_diagnostic = np.array(np.abs(convergence_diagnostic))
    #    self.convergence_diagnosticx = convergence_diagnosticx
    #    return sampler

    def run_sampler(self, sampler, ns, p0):
        convergence_period = 200
        convergence_diagnostic = []
        convergence_diagnosticx = []
        for i, result in enumerate(tqdm(
                sampler.sample(p0, iterations=ns), total=ns)):
            if np.mod(i+1, convergence_period) == 0:
                s = sampler.chain[0, :, i-convergence_period+1:i+1, :]
                Z = (s - np.mean(s, axis=(0, 1)))/np.std(s, axis=(0, 1))
                convergence_diagnostic.append(np.mean(Z, axis=(0, 1)))
                convergence_diagnosticx.append(i - convergence_period/2)
        self.convergence_diagnostic = np.array(np.abs(convergence_diagnostic))
        self.convergence_diagnosticx = convergence_diagnosticx
        return sampler

275
    def run(self, proposal_scale_factor=2, create_plots=True, **kwargs):
276

Gregory Ashton's avatar
Gregory Ashton committed
277
        self.old_data_is_okay_to_use = self.check_old_data_is_okay_to_use()
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
        if self.old_data_is_okay_to_use is True:
            logging.warning('Using saved data from {}'.format(
                self.pickle_path))
            d = self.get_saved_data()
            self.sampler = d['sampler']
            self.samples = d['samples']
            self.lnprobs = d['lnprobs']
            self.lnlikes = d['lnlikes']
            return

        self.inititate_search_object()

        sampler = emcee.PTSampler(
            self.ntemps, self.nwalkers, self.ndim, self.logl, self.logp,
            logpargs=(self.theta_prior, self.theta_keys, self.search),
293
            loglargs=(self.search,), betas=self.betas, a=proposal_scale_factor)
294

Gregory Ashton's avatar
Gregory Ashton committed
295
296
        p0 = self.generate_initial_p0()
        p0 = self.apply_corrections_to_p0(p0)
297
298
299
300
301
        self.check_initial_points(p0)

        ninit_steps = len(self.nsteps) - 2
        for j, n in enumerate(self.nsteps[:-2]):
            logging.info('Running {}/{} initialisation with {} steps'.format(
Gregory Ashton's avatar
Gregory Ashton committed
302
                j, ninit_steps, n))
303
            sampler = self.run_sampler(sampler, n, p0)
304
305
            logging.info("Mean acceptance fraction: {}"
                         .format(np.mean(sampler.acceptance_fraction, axis=1)))
306
307
308
            if self.ntemps > 1:
                logging.info("Tswap acceptance fraction: {}"
                             .format(sampler.tswap_acceptance_fraction))
309
310
311
312
313
314
315
            if create_plots:
                fig, axes = self.plot_walkers(sampler,
                                              symbols=self.theta_symbols,
                                              **kwargs)
                fig.tight_layout()
                fig.savefig('{}/{}_init_{}_walkers.png'.format(
                    self.outdir, self.label, j), dpi=200)
316

317
            p0 = self.get_new_p0(sampler)
Gregory Ashton's avatar
Gregory Ashton committed
318
            p0 = self.apply_corrections_to_p0(p0)
319
320
321
            self.check_initial_points(p0)
            sampler.reset()

Gregory Ashton's avatar
Gregory Ashton committed
322
323
324
325
        if len(self.nsteps) > 1:
            nburn = self.nsteps[-2]
        else:
            nburn = 0
326
327
328
        nprod = self.nsteps[-1]
        logging.info('Running final burn and prod with {} steps'.format(
            nburn+nprod))
329
        sampler = self.run_sampler(sampler, nburn+nprod, p0)
330
331
        logging.info("Mean acceptance fraction: {}"
                     .format(np.mean(sampler.acceptance_fraction, axis=1)))
332
333
334
        if self.ntemps > 1:
            logging.info("Tswap acceptance fraction: {}"
                         .format(sampler.tswap_acceptance_fraction))
335

336
337
338
339
340
341
        if create_plots:
            fig, axes = self.plot_walkers(sampler, symbols=self.theta_symbols,
                                          burnin_idx=nburn, **kwargs)
            fig.tight_layout()
            fig.savefig('{}/{}_walkers.png'.format(self.outdir, self.label),
                        dpi=200)
342
343
344
345
346
347
348
349
350
351

        samples = sampler.chain[0, :, nburn:, :].reshape((-1, self.ndim))
        lnprobs = sampler.lnprobability[0, :, nburn:].reshape((-1))
        lnlikes = sampler.lnlikelihood[0, :, nburn:].reshape((-1))
        self.sampler = sampler
        self.samples = samples
        self.lnprobs = lnprobs
        self.lnlikes = lnlikes
        self.save_data(sampler, samples, lnprobs, lnlikes)

352
    def plot_corner(self, figsize=(7, 7),  tglitch_ratio=False,
353
354
355
                    add_prior=False, nstds=None, label_offset=0.4,
                    dpi=300, rc_context={}, **kwargs):

Gregory Ashton's avatar
Gregory Ashton committed
356
357
358
359
360
361
362
363
364
365
        if self.ndim < 2:
            with plt.rc_context(rc_context):
                fig, ax = plt.subplots(figsize=figsize)
                ax.hist(self.samples, bins=50, histtype='stepfilled')
                ax.set_xlabel(self.theta_symbols[0])

            fig.savefig('{}/{}_corner.png'.format(
                self.outdir, self.label), dpi=dpi)
            return

366
367
368
369
370
371
        with plt.rc_context(rc_context):
            fig, axes = plt.subplots(self.ndim, self.ndim,
                                     figsize=figsize)

            samples_plt = copy.copy(self.samples)
            theta_symbols_plt = copy.copy(self.theta_symbols)
372
373
            theta_symbols_plt = [s.replace('_{glitch}', r'_\textrm{glitch}')
                                 for s in theta_symbols_plt]
374
375
376
377
378

            if tglitch_ratio:
                for j, k in enumerate(self.theta_keys):
                    if k == 'tglitch':
                        s = samples_plt[:, j]
379
380
381
                        samples_plt[:, j] = (
                            s - self.minStartTime)/(
                                self.maxStartTime - self.minStartTime)
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
                        theta_symbols_plt[j] = r'$R_{\textrm{glitch}}$'

            if type(nstds) is int and 'range' not in kwargs:
                _range = []
                for j, s in enumerate(samples_plt.T):
                    median = np.median(s)
                    std = np.std(s)
                    _range.append((median - nstds*std, median + nstds*std))
            else:
                _range = None

            fig_triangle = corner.corner(samples_plt,
                                         labels=theta_symbols_plt,
                                         fig=fig,
                                         bins=50,
                                         max_n_ticks=4,
                                         plot_contours=True,
                                         plot_datapoints=True,
                                         label_kwargs={'fontsize': 8},
                                         data_kwargs={'alpha': 0.1,
                                                      'ms': 0.5},
                                         range=_range,
                                         **kwargs)

            axes_list = fig_triangle.get_axes()
            axes = np.array(axes_list).reshape(self.ndim, self.ndim)
            plt.draw()
            for ax in axes[:, 0]:
                ax.yaxis.set_label_coords(-label_offset, 0.5)
            for ax in axes[-1, :]:
                ax.xaxis.set_label_coords(0.5, -label_offset)
            for ax in axes_list:
                ax.set_rasterized(True)
                ax.set_rasterization_zorder(-10)
            plt.tight_layout(h_pad=0.0, w_pad=0.0)
            fig.subplots_adjust(hspace=0.05, wspace=0.05)

            if add_prior:
                self.add_prior_to_corner(axes, samples_plt)

            fig_triangle.savefig('{}/{}_corner.png'.format(
                self.outdir, self.label), dpi=dpi)
424
425
426
427
428
429

    def add_prior_to_corner(self, axes, samples):
        for i, key in enumerate(self.theta_keys):
            ax = axes[i][i]
            xlim = ax.get_xlim()
            s = samples[:, i]
Gregory Ashton's avatar
Gregory Ashton committed
430
            prior = self.generic_lnprior(**self.theta_prior[key])
431
432
433
434
435
436
            x = np.linspace(s.min(), s.max(), 100)
            ax2 = ax.twinx()
            ax2.get_yaxis().set_visible(False)
            ax2.plot(x, [prior(xi) for xi in x], '-r')
            ax.set_xlim(xlim)

437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
    def plot_prior_posterior(self, normal_stds=2):
        """ Plot the posterior in the context of the prior """
        fig, axes = plt.subplots(nrows=self.ndim, figsize=(8, 4*self.ndim))
        N = 1000
        from scipy.stats import gaussian_kde

        for i, (ax, key) in enumerate(zip(axes, self.theta_keys)):
            prior_dict = self.theta_prior[key]
            prior_func = self.generic_lnprior(**prior_dict)
            if prior_dict['type'] == 'unif':
                x = np.linspace(prior_dict['lower'], prior_dict['upper'], N)
                prior = prior_func(x)
                prior[0] = 0
                prior[-1] = 0
            elif prior_dict['type'] == 'norm':
                lower = prior_dict['loc'] - normal_stds * prior_dict['scale']
                upper = prior_dict['loc'] + normal_stds * prior_dict['scale']
                x = np.linspace(lower, upper, N)
                prior = prior_func(x)
456
457
458
459
460
            elif prior_dict['type'] == 'halfnorm':
                lower = prior_dict['loc']
                upper = prior_dict['loc'] + normal_stds * prior_dict['scale']
                x = np.linspace(lower, upper, N)
                prior = [prior_func(xi) for xi in x]
Gregory Ashton's avatar
Gregory Ashton committed
461
462
463
464
465
            elif prior_dict['type'] == 'neghalfnorm':
                upper = prior_dict['loc']
                lower = prior_dict['loc'] - normal_stds * prior_dict['scale']
                x = np.linspace(lower, upper, N)
                prior = [prior_func(xi) for xi in x]
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
            else:
                raise ValueError('Not implemented for prior type {}'.format(
                    prior_dict['type']))
            priorln = ax.plot(x, prior, 'r', label='prior')
            ax.set_xlabel(self.theta_symbols[i])

            s = self.samples[:, i]
            while len(s) > 10**4:
                # random downsample to avoid slow calculation of kde
                s = np.random.choice(s, size=int(len(s)/2.))
            kde = gaussian_kde(s)
            ax2 = ax.twinx()
            postln = ax2.plot(x, kde.pdf(x), 'k', label='posterior')
            ax2.set_yticklabels([])
            ax.set_yticklabels([])

        lns = priorln + postln
        labs = [l.get_label() for l in lns]
        axes[0].legend(lns, labs, loc=1, framealpha=0.8)

        fig.savefig('{}/{}_prior_posterior.png'.format(
            self.outdir, self.label))

489
    def plot_cumulative_max(self, **kwargs):
Gregory Ashton's avatar
Gregory Ashton committed
490
491
492
493
        d, maxtwoF = self.get_max_twoF()
        for key, val in self.theta_prior.iteritems():
            if key not in d:
                d[key] = val
494
495
496
497
498
499

        if hasattr(self, 'search') is False:
            self.inititate_search_object()
        if self.binary is False:
            self.search.plot_twoF_cumulative(
                self.label, self.outdir, F0=d['F0'], F1=d['F1'], F2=d['F2'],
500
                Alpha=d['Alpha'], Delta=d['Delta'],
501
                tstart=self.minStartTime, tend=self.maxStartTime,
502
                **kwargs)
503
504
505
506
507
        else:
            self.search.plot_twoF_cumulative(
                self.label, self.outdir, F0=d['F0'], F1=d['F1'], F2=d['F2'],
                Alpha=d['Alpha'], Delta=d['Delta'], asini=d['asini'],
                period=d['period'], ecc=d['ecc'], argp=d['argp'], tp=d['argp'],
508
                tstart=self.minStartTime, tend=self.maxStartTime, **kwargs)
Gregory Ashton's avatar
Gregory Ashton committed
509

Gregory Ashton's avatar
Gregory Ashton committed
510
    def generic_lnprior(self, **kwargs):
511
512
513
514
515
516
517
518
519
520
521
522
523
524
525
526
527
528
529
530
531
532
533
534
        """ Return a lambda function of the pdf

        Parameters
        ----------
        kwargs: dict
            A dictionary containing 'type' of pdf and shape parameters

        """

        def logunif(x, a, b):
            above = x < b
            below = x > a
            if type(above) is not np.ndarray:
                if above and below:
                    return -np.log(b-a)
                else:
                    return -np.inf
            else:
                idxs = np.array([all(tup) for tup in zip(above, below)])
                p = np.zeros(len(x)) - np.inf
                p[idxs] = -np.log(b-a)
                return p

        def halfnorm(x, loc, scale):
535
            if x < loc:
536
537
538
539
540
541
542
543
544
545
546
547
548
549
550
551
552
                return -np.inf
            else:
                return -0.5*((x-loc)**2/scale**2+np.log(0.5*np.pi*scale**2))

        def cauchy(x, x0, gamma):
            return 1.0/(np.pi*gamma*(1+((x-x0)/gamma)**2))

        def exp(x, x0, gamma):
            if x > x0:
                return np.log(gamma) - gamma*(x - x0)
            else:
                return -np.inf

        if kwargs['type'] == 'unif':
            return lambda x: logunif(x, kwargs['lower'], kwargs['upper'])
        elif kwargs['type'] == 'halfnorm':
            return lambda x: halfnorm(x, kwargs['loc'], kwargs['scale'])
553
554
        elif kwargs['type'] == 'neghalfnorm':
            return lambda x: halfnorm(-x, kwargs['loc'], kwargs['scale'])
555
556
557
558
559
560
561
        elif kwargs['type'] == 'norm':
            return lambda x: -0.5*((x - kwargs['loc'])**2/kwargs['scale']**2
                                   + np.log(2*np.pi*kwargs['scale']**2))
        else:
            logging.info("kwargs:", kwargs)
            raise ValueError("Print unrecognise distribution")

Gregory Ashton's avatar
Gregory Ashton committed
562
    def generate_rv(self, **kwargs):
563
564
565
566
567
568
569
570
        dist_type = kwargs.pop('type')
        if dist_type == "unif":
            return np.random.uniform(low=kwargs['lower'], high=kwargs['upper'])
        if dist_type == "norm":
            return np.random.normal(loc=kwargs['loc'], scale=kwargs['scale'])
        if dist_type == "halfnorm":
            return np.abs(np.random.normal(loc=kwargs['loc'],
                                           scale=kwargs['scale']))
571
572
573
        if dist_type == "neghalfnorm":
            return -1 * np.abs(np.random.normal(loc=kwargs['loc'],
                                                scale=kwargs['scale']))
574
575
576
577
578
579
        if dist_type == "lognorm":
            return np.random.lognormal(
                mean=kwargs['loc'], sigma=kwargs['scale'])
        else:
            raise ValueError("dist_type {} unknown".format(dist_type))

Gregory Ashton's avatar
Gregory Ashton committed
580
    def plot_walkers(self, sampler, symbols=None, alpha=0.4, color="k", temp=0,
Gregory Ashton's avatar
Gregory Ashton committed
581
                     lw=0.1, burnin_idx=None, add_det_stat_burnin=False,
582
                     fig=None, axes=None, xoffset=0, plot_det_stat=True,
583
                     context='classic', subtractions=None, labelpad=0.05):
584
585
        """ Plot all the chains from a sampler """

586
587
588
        if np.ndim(axes) > 1:
            axes = axes.flatten()

589
590
591
592
593
594
595
596
597
598
599
600
601
        shape = sampler.chain.shape
        if len(shape) == 3:
            nwalkers, nsteps, ndim = shape
            chain = sampler.chain[:, :, :]
        if len(shape) == 4:
            ntemps, nwalkers, nsteps, ndim = shape
            if temp < ntemps:
                logging.info("Plotting temperature {} chains".format(temp))
            else:
                raise ValueError(("Requested temperature {} outside of"
                                  "available range").format(temp))
            chain = sampler.chain[temp, :, :, :]

602
603
        if subtractions is None:
            subtractions = [0 for i in range(ndim)]
604
605
606
        else:
            if len(subtractions) != self.ndim:
                raise ValueError('subtractions must be of length ndim')
607

608
        with plt.style.context((context)):
Gregory Ashton's avatar
Gregory Ashton committed
609
            plt.rcParams['text.usetex'] = True
Gregory Ashton's avatar
Gregory Ashton committed
610
            if fig is None and axes is None:
611
                fig = plt.figure(figsize=(4, 3.0*ndim))
Gregory Ashton's avatar
Gregory Ashton committed
612
                ax = fig.add_subplot(ndim+1, 1, 1)
Gregory Ashton's avatar
Gregory Ashton committed
613
                axes = [ax] + [fig.add_subplot(ndim+1, 1, i)
Gregory Ashton's avatar
Gregory Ashton committed
614
                               for i in range(2, ndim+1)]
615

Gregory Ashton's avatar
Gregory Ashton committed
616
            idxs = np.arange(chain.shape[1])
617
618
            if ndim > 1:
                for i in range(ndim):
619
                    axes[i].ticklabel_format(useOffset=False, axis='y')
Gregory Ashton's avatar
Gregory Ashton committed
620
621
                    cs = chain[:, :, i].T
                    if burnin_idx:
Gregory Ashton's avatar
Gregory Ashton committed
622
                        axes[i].plot(xoffset+idxs[:burnin_idx],
623
624
                                     cs[:burnin_idx]-subtractions[i],
                                     color="r", alpha=alpha,
Gregory Ashton's avatar
Gregory Ashton committed
625
                                     lw=lw)
626
627
                    axes[i].plot(xoffset+idxs[burnin_idx:],
                                 cs[burnin_idx:]-subtractions[i],
Gregory Ashton's avatar
Gregory Ashton committed
628
                                 color="k", alpha=alpha, lw=lw)
629
                    if symbols:
630
                        if subtractions[i] == 0:
631
                            axes[i].set_ylabel(symbols[i], labelpad=labelpad)
632
633
                        else:
                            axes[i].set_ylabel(
634
635
                                symbols[i]+'$-$'+symbols[i]+'$_0$',
                                labelpad=labelpad)
636

637
638
639
640
                    if hasattr(self, 'convergence_diagnostic'):
                        ax = axes[i].twinx()
                        ax.plot(self.convergence_diagnosticx,
                                self.convergence_diagnostic[:, i], '-b')
641
            else:
Gregory Ashton's avatar
Gregory Ashton committed
642
                axes[0].ticklabel_format(useOffset=False, axis='y')
Gregory Ashton's avatar
Gregory Ashton committed
643
                cs = chain[:, :, temp].T
Gregory Ashton's avatar
Gregory Ashton committed
644
645
646
647
648
649
                if burnin_idx:
                    axes[0].plot(idxs[:burnin_idx], cs[:burnin_idx],
                                 color="r", alpha=alpha, lw=lw)
                axes[0].plot(idxs[burnin_idx:], cs[burnin_idx:], color="k",
                             alpha=alpha, lw=lw)
                if symbols:
650
                    axes[0].set_ylabel(symbols[0], labelpad=labelpad)
651

652
            if plot_det_stat:
653
654
655
                if len(axes) == ndim:
                    axes.append(fig.add_subplot(ndim+1, 1, ndim+1))

656
657
658
                lnl = sampler.lnlikelihood[temp, :, :]
                if burnin_idx and add_det_stat_burnin:
                    burn_in_vals = lnl[:, :burnin_idx].flatten()
659
660
661
662
663
664
665
                    try:
                        axes[-1].hist(burn_in_vals[~np.isnan(burn_in_vals)],
                                      bins=50, histtype='step', color='r')
                    except ValueError:
                        logging.info('Det. Stat. hist failed, most likely all '
                                     'values where the same')
                        pass
666
667
668
                else:
                    burn_in_vals = []
                prod_vals = lnl[:, burnin_idx:].flatten()
669
670
671
672
673
674
675
                try:
                    axes[-1].hist(prod_vals[~np.isnan(prod_vals)], bins=50,
                                  histtype='step', color='k')
                except ValueError:
                    logging.info('Det. Stat. hist failed, most likely all '
                                 'values where the same')
                    pass
676
677
678
679
680
681
682
683
684
685
686
687
                if self.BSGL:
                    axes[-1].set_xlabel(r'$\mathcal{B}_\mathrm{S/GL}$')
                else:
                    axes[-1].set_xlabel(r'$\widetilde{2\mathcal{F}}$')
                axes[-1].set_ylabel(r'$\textrm{Counts}$')
                combined_vals = np.append(burn_in_vals, prod_vals)
                if len(combined_vals) > 0:
                    minv = np.min(combined_vals)
                    maxv = np.max(combined_vals)
                    Range = abs(maxv-minv)
                    axes[-1].set_xlim(minv-0.1*Range, maxv+0.1*Range)

688
                xfmt = matplotlib.ticker.ScalarFormatter()
689
                xfmt.set_powerlimits((-4, 4))
690
691
                axes[-1].xaxis.set_major_formatter(xfmt)

692
            axes[-2].set_xlabel(r'$\textrm{Number of steps}$', labelpad=0.2)
693
694
        return fig, axes

Gregory Ashton's avatar
Gregory Ashton committed
695
696
697
698
699
    def apply_corrections_to_p0(self, p0):
        """ Apply any correction to the initial p0 values """
        return p0

    def generate_scattered_p0(self, p):
700
        """ Generate a set of p0s scattered about p """
Gregory Ashton's avatar
Gregory Ashton committed
701
        p0 = [[p + self.scatter_val * p * np.random.randn(self.ndim)
702
703
704
705
               for i in xrange(self.nwalkers)]
              for j in xrange(self.ntemps)]
        return p0

Gregory Ashton's avatar
Gregory Ashton committed
706
    def generate_initial_p0(self):
707
708
709
        """ Generate a set of init vals for the walkers """

        if type(self.theta_initial) == dict:
710
            logging.info('Generate initial values from initial dictionary')
711
            if hasattr(self, 'nglitch') and self.nglitch > 1:
712
                raise ValueError('Initial dict not implemented for nglitch>1')
Gregory Ashton's avatar
Gregory Ashton committed
713
            p0 = [[[self.generate_rv(**self.theta_initial[key])
714
715
716
                    for key in self.theta_keys]
                   for i in range(self.nwalkers)]
                  for j in range(self.ntemps)]
717
718
719
720
721
722
        elif type(self.theta_initial) == list:
            logging.info('Generate initial values from list of theta_initial')
            p0 = [[[self.generate_rv(**val)
                    for val in self.theta_initial]
                   for i in range(self.nwalkers)]
                  for j in range(self.ntemps)]
723
        elif self.theta_initial is None:
724
            logging.info('Generate initial values from prior dictionary')
Gregory Ashton's avatar
Gregory Ashton committed
725
            p0 = [[[self.generate_rv(**self.theta_prior[key])
726
727
728
729
                    for key in self.theta_keys]
                   for i in range(self.nwalkers)]
                  for j in range(self.ntemps)]
        elif len(self.theta_initial) == self.ndim:
Gregory Ashton's avatar
Gregory Ashton committed
730
            p0 = self.generate_scattered_p0(self.theta_initial)
731
732
733
734
735
        else:
            raise ValueError('theta_initial not understood')

        return p0

736
    def get_new_p0(self, sampler):
737
738
739
740
741
742
        """ Returns new initial positions for walkers are burn0 stage

        This returns new positions for all walkers by scattering points about
        the maximum posterior with scale `scatter_val`.

        """
Gregory Ashton's avatar
Gregory Ashton committed
743
744
745
746
        temp_idx = 0
        pF = sampler.chain[temp_idx, :, :, :]
        lnl = sampler.lnlikelihood[temp_idx, :, :]
        lnp = sampler.lnprobability[temp_idx, :, :]
747
748

        # General warnings about the state of lnp
Gregory Ashton's avatar
Gregory Ashton committed
749
        if np.any(np.isnan(lnp)):
750
751
            logging.warning(
                "Of {} lnprobs {} are nan".format(
Gregory Ashton's avatar
Gregory Ashton committed
752
753
                    np.shape(lnp), np.sum(np.isnan(lnp))))
        if np.any(np.isposinf(lnp)):
754
755
            logging.warning(
                "Of {} lnprobs {} are +np.inf".format(
Gregory Ashton's avatar
Gregory Ashton committed
756
757
                    np.shape(lnp), np.sum(np.isposinf(lnp))))
        if np.any(np.isneginf(lnp)):
758
759
            logging.warning(
                "Of {} lnprobs {} are -np.inf".format(
Gregory Ashton's avatar
Gregory Ashton committed
760
                    np.shape(lnp), np.sum(np.isneginf(lnp))))
761

762
763
        lnp_finite = copy.copy(lnp)
        lnp_finite[np.isinf(lnp)] = np.nan
Gregory Ashton's avatar
Gregory Ashton committed
764
765
        idx = np.unravel_index(np.nanargmax(lnp_finite), lnp_finite.shape)
        p = pF[idx]
766
        p0 = self.generate_scattered_p0(p)
767

768
769
770
771
772
773
774
775
        self.search.BSGL = False
        twoF = self.logl(p, self.search)
        self.search.BSGL = self.BSGL

        logging.info(('Gen. new p0 from pos {} which had det. stat.={:2.1f},'
                      ' twoF={:2.1f} and lnp={:2.1f}')
                     .format(idx[1], lnl[idx], twoF, lnp_finite[idx]))

776
777
778
779
780
        return p0

    def get_save_data_dictionary(self):
        d = dict(nsteps=self.nsteps, nwalkers=self.nwalkers,
                 ntemps=self.ntemps, theta_keys=self.theta_keys,
Gregory Ashton's avatar
Gregory Ashton committed
781
                 theta_prior=self.theta_prior, scatter_val=self.scatter_val,
782
                 log10temperature_min=self.log10temperature_min,
783
                 BSGL=self.BSGL)
784
785
786
787
788
789
790
791
792
793
794
795
796
797
798
799
800
801
802
803
804
805
        return d

    def save_data(self, sampler, samples, lnprobs, lnlikes):
        d = self.get_save_data_dictionary()
        d['sampler'] = sampler
        d['samples'] = samples
        d['lnprobs'] = lnprobs
        d['lnlikes'] = lnlikes

        if os.path.isfile(self.pickle_path):
            logging.info('Saving backup of {} as {}.old'.format(
                self.pickle_path, self.pickle_path))
            os.rename(self.pickle_path, self.pickle_path+".old")
        with open(self.pickle_path, "wb") as File:
            pickle.dump(d, File)

    def get_saved_data(self):
        with open(self.pickle_path, "r") as File:
            d = pickle.load(File)
        return d

    def check_old_data_is_okay_to_use(self):
806
807
808
809
        if args.use_old_data:
            logging.info("Forcing use of old data")
            return True

810
811
812
813
        if os.path.isfile(self.pickle_path) is False:
            logging.info('No pickled data found')
            return False

Gregory Ashton's avatar
Gregory Ashton committed
814
815
816
817
818
819
        if self.sftfilepath is not None:
            oldest_sft = min([os.path.getmtime(f) for f in
                              self.get_list_of_matching_sfts()])
            if os.path.getmtime(self.pickle_path) < oldest_sft:
                logging.info('Pickled data outdates sft files')
                return False
820
821
822
823
824
825
826
827
828
829
830
831
832
833
834

        old_d = self.get_saved_data().copy()
        new_d = self.get_save_data_dictionary().copy()

        old_d.pop('samples')
        old_d.pop('sampler')
        old_d.pop('lnprobs')
        old_d.pop('lnlikes')

        mod_keys = []
        for key in new_d.keys():
            if key in old_d:
                if new_d[key] != old_d[key]:
                    mod_keys.append((key, old_d[key], new_d[key]))
            else:
835
                raise ValueError('Keys {} not in old dictionary'.format(key))
836
837
838
839
840
841
842
843
844

        if len(mod_keys) == 0:
            return True
        else:
            logging.warning("Saved data differs from requested")
            logging.info("Differences found in following keys:")
            for key in mod_keys:
                if len(key) == 3:
                    if np.isscalar(key[1]) or key[0] == 'nsteps':
845
                        logging.info("    {} : {} -> {}".format(*key))
846
                    else:
847
                        logging.info("    " + key[0])
848
849
850
851
852
                else:
                    logging.info(key)
            return False

    def get_max_twoF(self, threshold=0.05):
853
        """ Returns the max likelihood sample and the corresponding 2F value
854
855
856
857
858
859
860
861
862
863
864
865
866
867

        Note: the sample is returned as a dictionary along with an estimate of
        the standard deviation calculated from the std of all samples with a
        twoF within `threshold` (relative) to the max twoF

        """
        if any(np.isposinf(self.lnlikes)):
            logging.info('twoF values contain positive infinite values')
        if any(np.isneginf(self.lnlikes)):
            logging.info('twoF values contain negative infinite values')
        if any(np.isnan(self.lnlikes)):
            logging.info('twoF values contain nan')
        idxs = np.isfinite(self.lnlikes)
        jmax = np.nanargmax(self.lnlikes[idxs])
868
        maxlogl = self.lnlikes[jmax]
869
        d = OrderedDict()
870

871
872
873
874
875
876
877
878
879
880
        if self.BSGL:
            if hasattr(self, 'search') is False:
                self.inititate_search_object()
            p = self.samples[jmax]
            self.search.BSGL = False
            maxtwoF = self.logl(p, self.search)
            self.search.BSGL = self.BSGL
        else:
            maxtwoF = maxlogl

Gregory Ashton's avatar
Gregory Ashton committed
881
        repeats = []
882
        for i, k in enumerate(self.theta_keys):
Gregory Ashton's avatar
Gregory Ashton committed
883
884
885
886
887
888
889
890
891
892
            if k in d and k not in repeats:
                d[k+'_0'] = d[k]  # relabel the old key
                d.pop(k)
                repeats.append(k)
            if k in repeats:
                k = k + '_0'
                count = 1
                while k in d:
                    k = k.replace('_{}'.format(count-1), '_{}'.format(count))
                    count += 1
893
894
895
896
897
            d[k] = self.samples[jmax][i]
        return d, maxtwoF

    def get_median_stds(self):
        """ Returns a dict of the median and std of all production samples """
898
        d = OrderedDict()
Gregory Ashton's avatar
Gregory Ashton committed
899
        repeats = []
900
        for s, k in zip(self.samples.T, self.theta_keys):
Gregory Ashton's avatar
Gregory Ashton committed
901
902
903
904
905
906
907
908
909
910
911
912
913
            if k in d and k not in repeats:
                d[k+'_0'] = d[k]  # relabel the old key
                d[k+'_0_std'] = d[k+'_std']
                d.pop(k)
                d.pop(k+'_std')
                repeats.append(k)
            if k in repeats:
                k = k + '_0'
                count = 1
                while k in d:
                    k = k.replace('_{}'.format(count-1), '_{}'.format(count))
                    count += 1

914
915
916
917
            d[k] = np.median(s)
            d[k+'_std'] = np.std(s)
        return d

918
919
920
921
922
923
924
925
926
927
928
929
930
931
932
933
934
935
936
937
    def check_if_samples_are_railing(self, threshold=0.01):
        return_flag = False
        for s, k in zip(self.samples.T, self.theta_keys):
            prior = self.theta_prior[k]
            if prior['type'] == 'unif':
                prior_range = prior['upper'] - prior['lower']
                edges = []
                fracs = []
                for l in ['lower', 'upper']:
                    bools = np.abs(s - prior[l])/prior_range < threshold
                    if np.any(bools):
                        edges.append(l)
                        fracs.append(str(100*float(np.sum(bools))/len(bools)))
                if len(edges) > 0:
                    logging.warning(
                        '{}% of the {} posterior is railing on the {} edges'
                        .format('% & '.join(fracs), k, ' & '.join(edges)))
                    return_flag = True
        return return_flag

938
939
940
941
    def write_par(self, method='med'):
        """ Writes a .par of the best-fit params with an estimated std """
        logging.info('Writing {}/{}.par using the {} method'.format(
            self.outdir, self.label, method))
942
943
944
945

        median_std_d = self.get_median_stds()
        max_twoF_d, max_twoF = self.get_max_twoF()

Gregory Ashton's avatar
Gregory Ashton committed
946
        logging.info('Writing par file with max twoF = {}'.format(max_twoF))
947
948
949
        filename = '{}/{}.par'.format(self.outdir, self.label)
        with open(filename, 'w+') as f:
            f.write('MaxtwoF = {}\n'.format(max_twoF))
Gregory Ashton's avatar
Gregory Ashton committed
950
            f.write('tref = {}\n'.format(self.tref))
951
952
            if hasattr(self, 'theta0_index'):
                f.write('theta0_index = {}\n'.format(self.theta0_idx))
953
            if method == 'med':
954
955
                for key, val in median_std_d.iteritems():
                    f.write('{} = {:1.16e}\n'.format(key, val))
956
            if method == 'twoFmax':
957
958
959
960
                for key, val in max_twoF_d.iteritems():
                    f.write('{} = {:1.16e}\n'.format(key, val))

    def print_summary(self):
Gregory Ashton's avatar
Gregory Ashton committed
961
        max_twoFd, max_twoF = self.get_max_twoF()
962
        median_std_d = self.get_median_stds()
Gregory Ashton's avatar
Gregory Ashton committed
963
        logging.info('Summary:')
964
        if hasattr(self, 'theta0_idx'):
Gregory Ashton's avatar
Gregory Ashton committed
965
966
            logging.info('theta0 index: {}'.format(self.theta0_idx))
        logging.info('Max twoF: {} with parameters:'.format(max_twoF))
Gregory Ashton's avatar
Gregory Ashton committed
967
968
        for k in np.sort(max_twoFd.keys()):
            print('  {:10s} = {:1.9e}'.format(k, max_twoFd[k]))
Gregory Ashton's avatar
Gregory Ashton committed
969
        logging.info('Median +/- std for production values')
970
        for k in np.sort(median_std_d.keys()):
971
            if 'std' not in k:
Gregory Ashton's avatar
Gregory Ashton committed
972
                logging.info('  {:10s} = {:1.9e} +/- {:1.9e}'.format(
973
                    k, median_std_d[k], median_std_d[k+'_std']))
Gregory Ashton's avatar
Gregory Ashton committed
974
        logging.info('\n')
975

976
977
978
979
980
981
982
983
984
985
986
987
988
989
990
991
992
993
994
995
996
    def CF_twoFmax(self, theta, twoFmax, ntrials):
        Fmax = twoFmax/2.0
        return (np.exp(1j*theta*twoFmax)*ntrials/2.0
                * Fmax*np.exp(-Fmax)*(1-(1+Fmax)*np.exp(-Fmax))**(ntrials-1))

    def pdf_twoFhat(self, twoFhat, nglitch, ntrials, twoFmax=100, dtwoF=0.1):
        if np.ndim(ntrials) == 0:
            ntrials = np.zeros(nglitch+1) + ntrials
        twoFmax_int = np.arange(0, twoFmax, dtwoF)
        theta_int = np.arange(-1/dtwoF, 1./dtwoF, 1./twoFmax)
        CF_twoFmax_theta = np.array(
            [[np.trapz(self.CF_twoFmax(t, twoFmax_int, ntrial), twoFmax_int)
              for t in theta_int]
             for ntrial in ntrials])
        CF_twoFhat_theta = np.prod(CF_twoFmax_theta, axis=0)
        pdf = (1/(2*np.pi)) * np.array(
            [np.trapz(np.exp(-1j*theta_int*twoFhat_val)
             * CF_twoFhat_theta, theta_int) for twoFhat_val in twoFhat])
        return pdf.real

    def p_val_twoFhat(self, twoFhat, ntrials, twoFhatmax=500, Npoints=1000):
997
        """ Caluculate the p-value for the given twoFhat in Gaussian noise
998
999
1000

        Parameters
        ----------