mcmc_based_searches.py 93.5 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
import subprocess
9
10
11
12
13
14
15
16

import numpy as np
import matplotlib
import matplotlib.pyplot as plt
import emcee
import corner
import dill as pickle

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


24
class MCMCSearch(core.BaseSearchClass):
Gregory Ashton's avatar
Gregory Ashton committed
25
    """ MCMC search using ComputeFstat"""
26
27

    symbol_dictionary = dict(
28
        F0='$f$', F1='$\dot{f}$', F2='$\ddot{f}$', Alpha=r'$\alpha$',
29
30
        Delta='$\delta$', asini='asini', period='P', ecc='ecc', tp='tp',
        argp='argp')
31
    unit_dictionary = dict(
32
33
        F0='Hz', F1='Hz/s', F2='Hz/s$^2$', Alpha=r'rad', Delta='rad',
        asini='', period='s', ecc='', tp='', argp='')
34
35
36
    rescale_dictionary = {}


Gregory Ashton's avatar
Gregory Ashton committed
37
    @helper_functions.initializer
Gregory Ashton's avatar
Gregory Ashton committed
38
    def __init__(self, label, outdir, theta_prior, tref, minStartTime,
39
                 maxStartTime, sftfilepattern=None, nsteps=[100, 100],
40
                 nwalkers=100, ntemps=1, log10temperature_min=-5,
41
                 theta_initial=None, scatter_val=1e-10, rhohatmax=1000,
42
                 binary=False, BSGL=False, minCoverFreq=None, SSBprec=None,
43
                 maxCoverFreq=None, detectors=None, earth_ephem=None,
44
                 sun_ephem=None, injectSources=None, assumeSqrtSX=None):
45
46
47
48
        """
        Parameters
        label, outdir: str
            A label and directory to read/write data from/to
49
        sftfilepattern: str
50
51
            Pattern to match SFTs using wildcards (*?) and ranges [0-9];
            mutiple patterns can be given separated by colons.
52
        theta_prior: dict
53
54
55
56
            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.
57
58
59
60
        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.
61
        tref, minStartTime, maxStartTime: int
62
63
64
65
66
67
            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].
68
69
70
71
72
73
        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).
74
75
76
77
        rhohatmax: float
            Upper bound for the SNR scale parameter (required to normalise the
            Bayes factor) - this needs to be carefully set when using the
            evidence.
78
79
        binary: Bool
            If true, search over binary parameters
80
        detectors: str
81
82
            Two character reference to the data to use, specify None for no
            contraint.
83
84
85
86
87
88
89
90
91
92
        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
93
94
        if os.path.isdir(outdir) is False:
            os.mkdir(outdir)
95
        self._add_log_file()
96
        logging.info('Set-up MCMC search for model {}'.format(self.label))
97
98
        if sftfilepattern:
            logging.info('Using data {}'.format(self.sftfilepattern))
99
        else:
100
            logging.info('No sftfilepattern given')
101
102
        if injectSources:
            logging.info('Inject sources: {}'.format(injectSources))
103
        self.pickle_path = '{}/{}_saved_data.p'.format(self.outdir, self.label)
104
        self._unpack_input_theta()
105
        self.ndim = len(self.theta_keys)
106
107
108
109
        if self.log10temperature_min:
            self.betas = np.logspace(0, self.log10temperature_min, self.ntemps)
        else:
            self.betas = None
110

111
112
113
114
115
116
117
118
        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")

119
120
        self.lnlikelihoodcoef = np.log(70./self.rhohatmax**4)

121
        self._log_input()
122

123
    def _log_input(self):
124
        logging.info('theta_prior = {}'.format(self.theta_prior))
125
        logging.info('nwalkers={}'.format(self.nwalkers))
126
127
128
129
        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(
130
            self.log10temperature_min))
131

132
    def _initiate_search_object(self):
133
        logging.info('Setting up search object')
134
        self.search = core.ComputeFstat(
135
            tref=self.tref, sftfilepattern=self.sftfilepattern,
136
137
            minCoverFreq=self.minCoverFreq, maxCoverFreq=self.maxCoverFreq,
            earth_ephem=self.earth_ephem, sun_ephem=self.sun_ephem,
138
            detectors=self.detectors, BSGL=self.BSGL, transient=False,
139
            minStartTime=self.minStartTime, maxStartTime=self.maxStartTime,
140
            binary=self.binary, injectSources=self.injectSources,
141
            assumeSqrtSX=self.assumeSqrtSX, SSBprec=self.SSBprec)
142
143

    def logp(self, theta_vals, theta_prior, theta_keys, search):
144
        H = [self._generic_lnprior(**theta_prior[key])(p) for p, key in
145
146
147
148
149
150
             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]
151
152
        FS = search.compute_fullycoherent_det_stat_single_point(
            *self.fixed_theta)
153
        return FS + self.lnlikelihoodcoef
154

155
    def _unpack_input_theta(self):
156
        full_theta_keys = ['F0', 'F1', 'F2', 'Alpha', 'Delta']
157
158
159
        if self.binary:
            full_theta_keys += [
                'asini', 'period', 'ecc', 'tp', 'argp']
160
161
        full_theta_keys_copy = copy.copy(full_theta_keys)

162
163
        full_theta_symbols = ['$f$', '$\dot{f}$', '$\ddot{f}$', r'$\alpha$',
                              r'$\delta$']
164
165
        if self.binary:
            full_theta_symbols += [
166
                'asini', 'period', 'ecc', 'tp', 'argp']
167

168
169
        self.theta_keys = []
        fixed_theta_dict = {}
170
        for key, val in self.theta_prior.iteritems():
171
172
            if type(val) is dict:
                fixed_theta_dict[key] = 0
Gregory Ashton's avatar
Gregory Ashton committed
173
                self.theta_keys.append(key)
174
175
176
177
178
179
            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
180
            full_theta_keys_copy.pop(full_theta_keys_copy.index(key))
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195

        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]

196
    def _check_initial_points(self, p0):
197
198
199
200
201
202
203
204
205
206
207
208
209
        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))

210
                p0 = self._generate_new_p0_to_fix_initial_points(
211
212
                    p0, nt, initial_priors)

213
    def _generate_new_p0_to_fix_initial_points(self, p0, nt, initial_priors):
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
        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
233

234
    def _OLD_run_sampler_with_progress_bar(self, sampler, ns, p0):
235
236
        for result in tqdm(sampler.sample(p0, iterations=ns), total=ns):
            pass
Gregory Ashton's avatar
Gregory Ashton committed
237
238
        return sampler

239
240
    def setup_convergence_testing(
            self, convergence_period=10, convergence_length=10,
241
            convergence_burnin_fraction=0.25, convergence_threshold_number=10,
242
            convergence_threshold=1.2, convergence_prod_threshold=2,
243
            convergence_plot_upper_lim=2, convergence_early_stopping=True):
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
        """
        If called, convergence testing is used during the MCMC simulation

        This uses the Gelmanr-Rubin statistic based on the ratio of between and
        within walkers variance. The original statistic was developed for
        multiple (independent) MCMC simulations, in this context we simply use
        the walkers

        Parameters
        ----------
        convergence_period: int
            period (in number of steps) at which to test convergence
        convergence_length: int
            number of steps to use in testing convergence - this should be
            large enough to measure the variance, but if it is too long
            this will result in incorect early convergence tests
        convergence_burnin_fraction: float [0, 1]
            the fraction of the burn-in period after which to start testing
        convergence_threshold_number: int
            the number of consecutive times where the test passes after which
            to break the burn-in and go to production
        convergence_threshold: float
            the threshold to use in diagnosing convergence. Gelman & Rubin
            recomend a value of 1.2, 1.1 for strict convergence
        convergence_prod_threshold: float
            the threshold to test the production values with
270
271
        convergence_plot_upper_lim: float
            the upper limit to use in the diagnostic plot
272
273
        convergence_early_stopping: bool
            if true, stop the burnin early if convergence is reached
274
        """
275
276
277
278
279
280
281

        if convergence_length > convergence_period:
            raise ValueError('convergence_length must be < convergence_period')
        logging.info('Setting up convergence testing')
        self.convergence_length = convergence_length
        self.convergence_period = convergence_period
        self.convergence_burnin_fraction = convergence_burnin_fraction
282
        self.convergence_prod_threshold = convergence_prod_threshold
283
284
285
286
287
        self.convergence_diagnostic = []
        self.convergence_diagnosticx = []
        self.convergence_threshold_number = convergence_threshold_number
        self.convergence_threshold = convergence_threshold
        self.convergence_number = 0
288
        self.convergence_plot_upper_lim = convergence_plot_upper_lim
289
        self.convergence_early_stopping = convergence_early_stopping
290

291
    def _get_convergence_statistic(self, i, sampler):
292
        s = sampler.chain[0, :, i-self.convergence_length+1:i+1, :]
293
294
295
        N = float(self.convergence_length)
        M = float(self.nwalkers)
        W = np.mean(np.var(s, axis=1), axis=0)
296
297
        per_walker_mean = np.mean(s, axis=1)
        mean = np.mean(per_walker_mean, axis=0)
298
299
        B = N / (M-1.) * np.sum((per_walker_mean-mean)**2, axis=0)
        Vhat = (N-1)/N * W + (M+1)/(M*N) * B
300
        c = np.sqrt(Vhat/W)
301
        self.convergence_diagnostic.append(c)
302
        self.convergence_diagnosticx.append(i - self.convergence_length/2)
303
304
        return c

305
    def _burnin_convergence_test(self, i, sampler, nburn):
306
307
        if i < self.convergence_burnin_fraction*nburn:
            return False
308
        if np.mod(i+1, self.convergence_period) != 0:
309
            return False
310
        c = self._get_convergence_statistic(i, sampler)
311
312
        if np.all(c < self.convergence_threshold):
            self.convergence_number += 1
313
314
        else:
            self.convergence_number = 0
315
316
        if self.convergence_early_stopping:
            return self.convergence_number > self.convergence_threshold_number
317

318
    def _prod_convergence_test(self, i, sampler, nburn):
319
320
321
        testA = i > nburn + self.convergence_length
        testB = np.mod(i+1, self.convergence_period) == 0
        if testA and testB:
322
            self._get_convergence_statistic(i, sampler)
323

324
    def _check_production_convergence(self, k):
325
326
327
328
329
330
331
332
        bools = np.any(
            np.array(self.convergence_diagnostic)[k:, :]
            > self.convergence_prod_threshold, axis=1)
        if np.any(bools):
            logging.warning(
                '{} convergence tests in the production run of {} failed'
                .format(np.sum(bools), len(bools)))

333
    def _run_sampler(self, sampler, p0, nprod=0, nburn=0):
334
        if hasattr(self, 'convergence_period'):
335
336
337
338
            logging.info('Running {} burn-in steps with convergence testing'
                         .format(nburn))
            iterator = tqdm(sampler.sample(p0, iterations=nburn), total=nburn)
            for i, output in enumerate(iterator):
339
                if self._burnin_convergence_test(i, sampler, nburn):
340
341
342
343
344
345
346
347
348
349
350
                    logging.info(
                        'Converged at {} before max number {} of steps reached'
                        .format(i, nburn))
                    self.convergence_idx = i
                    break
            iterator.close()
            logging.info('Running {} production steps'.format(nprod))
            j = nburn
            k = len(self.convergence_diagnostic)
            for result in tqdm(sampler.sample(output[0], iterations=nprod),
                               total=nprod):
351
                self._prod_convergence_test(j, sampler, nburn)
352
                j += 1
353
            self._check_production_convergence(k)
354
355
356
357
358
359
            return sampler
        else:
            for result in tqdm(sampler.sample(p0, iterations=nburn+nprod),
                               total=nburn+nprod):
                pass
            return sampler
360

361
    def run(self, proposal_scale_factor=2, create_plots=True, **kwargs):
362
        """ Run the MCMC simulatation """
363

364
        self.old_data_is_okay_to_use = self._check_old_data_is_okay_to_use()
365
366
367
        if self.old_data_is_okay_to_use is True:
            logging.warning('Using saved data from {}'.format(
                self.pickle_path))
368
            d = self.get_saved_data_dictionary()
369
370
371
            self.samples = d['samples']
            self.lnprobs = d['lnprobs']
            self.lnlikes = d['lnlikes']
372
            self.all_lnlikelihood = d['all_lnlikelihood']
373
374
            return

375
        self._initiate_search_object()
376
377
378
379

        sampler = emcee.PTSampler(
            self.ntemps, self.nwalkers, self.ndim, self.logl, self.logp,
            logpargs=(self.theta_prior, self.theta_keys, self.search),
380
            loglargs=(self.search,), betas=self.betas, a=proposal_scale_factor)
381

382
383
384
        p0 = self._generate_initial_p0()
        p0 = self._apply_corrections_to_p0(p0)
        self._check_initial_points(p0)
385
386
387
388

        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
389
                j, ninit_steps, n))
390
            sampler = self._run_sampler(sampler, p0, nburn=n)
391
392
            logging.info("Mean acceptance fraction: {}"
                         .format(np.mean(sampler.acceptance_fraction, axis=1)))
393
394
395
            if self.ntemps > 1:
                logging.info("Tswap acceptance fraction: {}"
                             .format(sampler.tswap_acceptance_fraction))
396
            if create_plots:
397
                fig, axes = self._plot_walkers(sampler,
398
399
400
401
                                              symbols=self.theta_symbols,
                                              **kwargs)
                fig.tight_layout()
                fig.savefig('{}/{}_init_{}_walkers.png'.format(
402
                    self.outdir, self.label, j), dpi=400)
403

404
405
406
            p0 = self._get_new_p0(sampler)
            p0 = self._apply_corrections_to_p0(p0)
            self._check_initial_points(p0)
407
408
            sampler.reset()

Gregory Ashton's avatar
Gregory Ashton committed
409
410
411
412
        if len(self.nsteps) > 1:
            nburn = self.nsteps[-2]
        else:
            nburn = 0
413
414
415
        nprod = self.nsteps[-1]
        logging.info('Running final burn and prod with {} steps'.format(
            nburn+nprod))
416
        sampler = self._run_sampler(sampler, p0, nburn=nburn, nprod=nprod)
417
418
        logging.info("Mean acceptance fraction: {}"
                     .format(np.mean(sampler.acceptance_fraction, axis=1)))
419
420
421
        if self.ntemps > 1:
            logging.info("Tswap acceptance fraction: {}"
                         .format(sampler.tswap_acceptance_fraction))
422

423
        if create_plots:
424
            fig, axes = self._plot_walkers(sampler, symbols=self.theta_symbols,
425
                                          nprod=nprod, **kwargs)
426
427
428
            fig.tight_layout()
            fig.savefig('{}/{}_walkers.png'.format(self.outdir, self.label),
                        dpi=200)
429
430
431
432

        samples = sampler.chain[0, :, nburn:, :].reshape((-1, self.ndim))
        lnprobs = sampler.lnprobability[0, :, nburn:].reshape((-1))
        lnlikes = sampler.lnlikelihood[0, :, nburn:].reshape((-1))
433
        all_lnlikelihood = sampler.lnlikelihood[:, :, nburn:]
434
435
436
        self.samples = samples
        self.lnprobs = lnprobs
        self.lnlikes = lnlikes
437
438
        self.all_lnlikelihood = all_lnlikelihood
        self._save_data(sampler, samples, lnprobs, lnlikes, all_lnlikelihood)
439

440
    def _get_rescale_multiplier_for_key(self, key):
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
        """ Get the rescale multiplier from the rescale_dictionary

        Can either be a float, a string (in which case it is interpretted as
        a attribute of the MCMCSearch class, e.g. minStartTime, or non-existent
        in which case 0 is returned
        """
        if key not in self.rescale_dictionary:
            return 1

        if 'multiplier' in self.rescale_dictionary[key]:
            val = self.rescale_dictionary[key]['multiplier']
            if type(val) == str:
                if hasattr(self, val):
                    multiplier = getattr(
                        self, self.rescale_dictionary[key]['multiplier'])
                else:
                    raise ValueError(
                        "multiplier {} not a class attribute".format(val))
            else:
                multiplier = val
        else:
            multiplier = 1
        return multiplier

465
    def _get_rescale_subtractor_for_key(self, key):
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
        """ Get the rescale subtractor from the rescale_dictionary

        Can either be a float, a string (in which case it is interpretted as
        a attribute of the MCMCSearch class, e.g. minStartTime, or non-existent
        in which case 0 is returned
        """
        if key not in self.rescale_dictionary:
            return 0

        if 'subtractor' in self.rescale_dictionary[key]:
            val = self.rescale_dictionary[key]['subtractor']
            if type(val) == str:
                if hasattr(self, val):
                    subtractor = getattr(
                        self, self.rescale_dictionary[key]['subtractor'])
                else:
                    raise ValueError(
                        "subtractor {} not a class attribute".format(val))
            else:
                subtractor = val
        else:
            subtractor = 0
        return subtractor

490
    def _scale_samples(self, samples, theta_keys):
491
        """ Scale the samples using the rescale_dictionary """
492
493
494
495
        for key in theta_keys:
            if key in self.rescale_dictionary:
                idx = theta_keys.index(key)
                s = samples[:, idx]
496
                subtractor = self._get_rescale_subtractor_for_key(key)
497
                s = s - subtractor
498
                multiplier = self._get_rescale_multiplier_for_key(key)
499
                s *= multiplier
500
501
                samples[:, idx] = s

502
503
        return samples

504
    def _get_labels(self):
505
        """ Combine the units, symbols and rescaling to give labels """
506

507
508
509
510
511
512
513
514
515
516
517
518
519
520
521
522
523
        labels = []
        for key in self.theta_keys:
            label = None
            s = self.symbol_dictionary[key]
            s.replace('_{glitch}', r'_\textrm{glitch}')
            u = self.unit_dictionary[key]
            if key in self.rescale_dictionary:
                if 'symbol' in self.rescale_dictionary[key]:
                    s = self.rescale_dictionary[key]['symbol']
                if 'label' in self.rescale_dictionary[key]:
                    label = self.rescale_dictionary[key]['label']
                if 'unit' in self.rescale_dictionary[key]:
                    u = self.rescale_dictionary[key]['unit']
            if label is None:
                label = '{} \n [{}]'.format(s, u)
            labels.append(label)
        return labels
524

525
526
    def plot_corner(self, figsize=(7, 7), add_prior=False, nstds=None,
                    label_offset=0.4, dpi=300, rc_context={},
527
                    tglitch_ratio=False, fig_and_axes=None, save_fig=True,
528
                    **kwargs):
529
530
531
532
533
534
535
536
537
        """ Generate a corner plot of the posterior

        Using the `corner` package (https://pypi.python.org/pypi/corner/),
        generate estimates of the posterior from the production samples.

        Parameters
        ----------
        figsize: tuple (7, 7)
            Figure size in inches (passed to plt.subplots)
538
539
540
        add_prior: bool, str
            If true, plot the prior as a red line. If 'full' then for uniform
            priors plot the full extent of the prior.
541
542
543
544
545
546
547
548
549
550
551
552
553
554
        nstds: float
            The number of standard deviations to plot centered on the mean
        label_offset: float
            Offset the labels from the plot: useful to precent overlapping the
            tick labels with the axis labels
        dpi: int
            Passed to plt.savefig
        rc_context: dict
            Dictionary of rc values to set while generating the figure (see
            matplotlib rc for more details)
        tglitch_ratio: bool
            If true, and tglitch is a parameter, plot posteriors as the
            fractional time at which the glitch occurs instead of the actual
            time
555
556
557
558
559
        fig_and_axes: tuple
            fig and axes to plot on, the axes must be of the right shape,
            namely (ndim, ndim)
        save_fig: bool
            If true, save the figure, else return the fig, axes
560

561
        Note: kwargs are passed on to corner.corner
562
563

        """
564

565
566
567
568
        if 'truths' in kwargs and len(kwargs['truths']) != self.ndim:
            logging.warning('len(Truths) != ndim, Truths will be ignored')
            kwargs['truths'] = None

Gregory Ashton's avatar
Gregory Ashton committed
569
570
        if self.ndim < 2:
            with plt.rc_context(rc_context):
571
572
573
574
                if fig_and_axes is None:
                    fig, ax = plt.subplots(figsize=figsize)
                else:
                    fig, ax = fig_and_axes
Gregory Ashton's avatar
Gregory Ashton committed
575
576
577
578
579
580
581
                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

582
        with plt.rc_context(rc_context):
583
584
585
586
587
            if fig_and_axes is None:
                fig, axes = plt.subplots(self.ndim, self.ndim,
                                         figsize=figsize)
            else:
                fig, axes = fig_and_axes
588
589

            samples_plt = copy.copy(self.samples)
590
            labels = self._get_labels()
591

592
            samples_plt = self._scale_samples(samples_plt, self.theta_keys)
593
594
595
596
597

            if tglitch_ratio:
                for j, k in enumerate(self.theta_keys):
                    if k == 'tglitch':
                        s = samples_plt[:, j]
598
599
600
                        samples_plt[:, j] = (
                            s - self.minStartTime)/(
                                self.maxStartTime - self.minStartTime)
601
                        labels[j] = r'$R_{\textrm{glitch}}$'
602
603
604
605
606
607
608

            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))
609
610
            elif 'range' in kwargs:
                _range = kwargs.pop('range')
611
612
613
            else:
                _range = None

614
615
616
617
            hist_kwargs = kwargs.pop('hist_kwargs', dict())
            if 'normed' not in hist_kwargs:
                hist_kwargs['normed'] = True

618
            fig_triangle = corner.corner(samples_plt,
619
                                         labels=labels,
620
621
622
623
624
625
626
627
628
                                         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,
629
                                         hist_kwargs=hist_kwargs,
630
631
632
633
634
635
636
637
638
639
640
641
642
643
644
645
                                         **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:
646
                self._add_prior_to_corner(axes, self.samples, add_prior)
647

648
649
650
651
652
            if save_fig:
                fig_triangle.savefig('{}/{}_corner.png'.format(
                    self.outdir, self.label), dpi=dpi)
            else:
                return fig, axes
653

654
    def _add_prior_to_corner(self, axes, samples, add_prior):
655
656
657
        for i, key in enumerate(self.theta_keys):
            ax = axes[i][i]
            s = samples[:, i]
658
659
660
661
662
663
664
665
666
667
            lnprior = self._generic_lnprior(**self.theta_prior[key])
            if add_prior == 'full' and self.theta_prior[key]['type'] == 'unif':
                lower = self.theta_prior[key]['lower']
                upper = self.theta_prior[key]['upper']
                r = upper-lower
                xlim = [lower-0.05*r, upper+0.05*r]
                x = np.linspace(xlim[0], xlim[1], 1000)
            else:
                xlim = ax.get_xlim()
                x = np.linspace(s.min(), s.max(), 1000)
668
669
            multiplier = self._get_rescale_multiplier_for_key(key)
            subtractor = self._get_rescale_subtractor_for_key(key)
670
671
672
673
674
675
676
677
            ax.plot((x-subtractor)*multiplier,
                    [np.exp(lnprior(xi)) for xi in x], '-C3',
                    label='prior')

            for j in range(i, self.ndim):
                axes[j][i].set_xlim(xlim[0], xlim[1])
            for k in range(0, i):
                axes[i][k].set_ylim(xlim[0], xlim[1])
678

679
680
681
682
683
684
685
686
    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]
687
            prior_func = self._generic_lnprior(**prior_dict)
688
689
690
691
692
            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
Gregory Ashton's avatar
Gregory Ashton committed
693
694
695
696
697
            elif prior_dict['type'] == 'log10unif':
                upper = prior_dict['log10upper']
                lower = prior_dict['log10lower']
                x = np.linspace(lower, upper, N)
                prior = [prior_func(xi) for xi in x]
698
699
700
701
702
            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)
703
704
705
706
707
            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
708
709
710
711
712
            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]
713
714
715
            else:
                raise ValueError('Not implemented for prior type {}'.format(
                    prior_dict['type']))
716
            priorln = ax.plot(x, prior, 'C3', label='prior')
717
718
719
720
721
722
723
724
725
726
727
728
729
730
731
732
733
734
735
            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))

736
    def plot_cumulative_max(self, **kwargs):
737
738
739
740
        """ Plot the cumulative twoF for the maximum posterior estimate

        See the pyfstat.core.plot_twoF_cumulative function for further details
        """
Gregory Ashton's avatar
Gregory Ashton committed
741
742
743
744
        d, maxtwoF = self.get_max_twoF()
        for key, val in self.theta_prior.iteritems():
            if key not in d:
                d[key] = val
745
746

        if hasattr(self, 'search') is False:
747
            self._initiate_search_object()
748
749
750
        if self.binary is False:
            self.search.plot_twoF_cumulative(
                self.label, self.outdir, F0=d['F0'], F1=d['F1'], F2=d['F2'],
751
                Alpha=d['Alpha'], Delta=d['Delta'],
752
                tstart=self.minStartTime, tend=self.maxStartTime,
753
                **kwargs)
754
755
756
757
758
        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'],
759
                tstart=self.minStartTime, tend=self.maxStartTime, **kwargs)
Gregory Ashton's avatar
Gregory Ashton committed
760

761
    def _generic_lnprior(self, **kwargs):
762
763
764
765
766
767
768
769
770
        """ Return a lambda function of the pdf

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

        """

Gregory Ashton's avatar
Gregory Ashton committed
771
        def log_of_unif(x, a, b):
772
773
774
775
776
777
778
779
780
781
782
783
784
            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

Gregory Ashton's avatar
Gregory Ashton committed
785
786
787
788
789
790
791
792
793
794
795
796
797
798
799
800
        def log_of_log10unif(x, log10lower, log10upper):
            log10x = np.log10(x)
            above = log10x < log10upper
            below = log10x > log10lower
            if type(above) is not np.ndarray:
                if above and below:
                    return -np.log(x*np.log(10)*(log10upper-log10lower))
                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(x*np.log(10)*(log10upper-log10lower))
                return p

        def log_of_halfnorm(x, loc, scale):
801
            if x < loc:
802
803
804
805
806
807
808
809
810
811
812
813
814
815
                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':
Gregory Ashton's avatar
Gregory Ashton committed
816
817
818
819
            return lambda x: log_of_unif(x, kwargs['lower'], kwargs['upper'])
        if kwargs['type'] == 'log10unif':
            return lambda x: log_of_log10unif(
                x, kwargs['log10lower'], kwargs['log10upper'])
820
        elif kwargs['type'] == 'halfnorm':
Gregory Ashton's avatar
Gregory Ashton committed
821
            return lambda x: log_of_halfnorm(x, kwargs['loc'], kwargs['scale'])
822
        elif kwargs['type'] == 'neghalfnorm':
Gregory Ashton's avatar
Gregory Ashton committed
823
824
            return lambda x: log_of_halfnorm(
                -x, kwargs['loc'], kwargs['scale'])
825
826
827
828
829
830
831
        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")

832
    def _generate_rv(self, **kwargs):
833
834
835
        dist_type = kwargs.pop('type')
        if dist_type == "unif":
            return np.random.uniform(low=kwargs['lower'], high=kwargs['upper'])
Gregory Ashton's avatar
Gregory Ashton committed
836
837
838
        if dist_type == "log10unif":
            return 10**(np.random.uniform(low=kwargs['log10lower'],
                                          high=kwargs['log10upper']))
839
840
841
842
843
        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']))
844
845
846
        if dist_type == "neghalfnorm":
            return -1 * np.abs(np.random.normal(loc=kwargs['loc'],
                                                scale=kwargs['scale']))
847
848
849
850
851
852
        if dist_type == "lognorm":
            return np.random.lognormal(
                mean=kwargs['loc'], sigma=kwargs['scale'])
        else:
            raise ValueError("dist_type {} unknown".format(dist_type))

853
    def _plot_walkers(self, sampler, symbols=None, alpha=0.8, color="k",
854
855
                      temp=0, lw=0.1, nprod=0, add_det_stat_burnin=False,
                      fig=None, axes=None, xoffset=0, plot_det_stat=False,
856
                      context='ggplot', subtractions=None, labelpad=0.05):
857
858
        """ Plot all the chains from a sampler """

859
860
861
862
863
        if context not in plt.style.available:
            raise ValueError((
                'The requested context {} is not available; please select a'
                ' context from `plt.style.available`').format(context))

864
865
866
        if np.ndim(axes) > 1:
            axes = axes.flatten()

867
868
869
870
871
872
873
874
875
876
877
878
879
        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, :, :, :]

880
881
        if subtractions is None:
            subtractions = [0 for i in range(ndim)]
882
883
884
        else:
            if len(subtractions) != self.ndim:
                raise ValueError('subtractions must be of length ndim')
885

886
887
888
889
        if plot_det_stat:
            extra_subplots = 1
        else:
            extra_subplots = 0
890
        with plt.style.context((context)):
Gregory Ashton's avatar
Gregory Ashton committed
891
            plt.rcParams['text.usetex'] = True
Gregory Ashton's avatar
Gregory Ashton committed
892
            if fig is None and axes is None:
893
                fig = plt.figure(figsize=(4, 3.0*ndim))
894
895
                ax = fig.add_subplot(ndim+extra_subplots, 1, 1)
                axes = [ax] + [fig.add_subplot(ndim+extra_subplots, 1, i)
Gregory Ashton's avatar
Gregory Ashton committed
896
                               for i in range(2, ndim+1)]
897

Gregory Ashton's avatar
Gregory Ashton committed
898
            idxs = np.arange(chain.shape[1])
899
900
901
902
903
            burnin_idx = chain.shape[1] - nprod
            if hasattr(self, 'convergence_idx'):
                convergence_idx = self.convergence_idx
            else:
                convergence_idx = burnin_idx
904
905
            if ndim > 1:
                for i in range(ndim):
906
                    axes[i].ticklabel_format(useOffset=False, axis='y')
Gregory Ashton's avatar
Gregory Ashton committed
907
                    cs = chain[:, :, i].T
908
                    if burnin_idx > 0:
909
910
                        axes[i].plot(xoffset+idxs[:convergence_idx+1],
                                     cs[:convergence_idx+1]-subtractions[i],
911
                                     color="C3", alpha=alpha,
Gregory Ashton's avatar
Gregory Ashton committed
912
                                     lw=lw)
913
                        axes[i].axvline(xoffset+convergence_idx,
914
                                        color='k', ls='--', lw=0.25)
915
916
                    axes[i].plot(xoffset+idxs[burnin_idx:],
                                 cs[burnin_idx:]-subtractions[i],
Gregory Ashton's avatar
Gregory Ashton committed
917
                                 color="k", alpha=alpha, lw=lw)
918
                    if symbols:
919
                        if subtractions[i] == 0:
920
                            axes[i].set_ylabel(symbols[i], labelpad=labelpad)
921
922
                        else:
                            axes[i].set_ylabel(
923
924
                                symbols[i]+'$-$'+symbols[i]+'$_0$',
                                labelpad=labelpad)
925

926
927
                    if hasattr(self, 'convergence_diagnostic'):
                        ax = axes[i].twinx()
928
929
                        axes[i].set_zorder(ax.get_zorder()+1)
                        axes[i].patch.set_visible(False)
930
931
                        c_x = np.array(self.convergence_diagnosticx)
                        c_y = np.array(self.convergence_diagnostic)
932
                        break_idx = np.argmin(np.abs(c_x - burnin_idx))
933
934
935
936
                        ax.plot(c_x[:break_idx], c_y[:break_idx, i], '-C0',
                                zorder=-10)
                        ax.plot(c_x[break_idx:], c_y[break_idx:, i], '-C0',
                                zorder=-10)
937
                        ax.set_ylabel('PSRF')
938
                        ax.ticklabel_format(useOffset=False)
939
                        ax.set_ylim(0.5, self.convergence_plot_upper_lim)
940
            else:
Gregory Ashton's avatar
Gregory Ashton committed
941
                axes[0].ticklabel_format(useOffset=False, axis='y')
Gregory Ashton's avatar
Gregory Ashton committed
942
                cs = chain[:, :, temp].T
Gregory Ashton's avatar
Gregory Ashton committed
943
944
                if burnin_idx:
                    axes[0].plot(idxs[:burnin_idx], cs[:burnin_idx],
945
                                 color="C3", alpha=alpha, lw=lw)
Gregory Ashton's avatar
Gregory Ashton committed
946
947
948
                axes[0].plot(idxs[burnin_idx:], cs[burnin_idx:], color="k",
                             alpha=alpha, lw=lw)
                if symbols:
949
                    axes[0].set_ylabel(symbols[0], labelpad=labelpad)
950

Gregory Ashton's avatar
Gregory Ashton committed
951
952
            axes[-1].set_xlabel(r'$\textrm{Number of steps}$', labelpad=0.2)

953
            if plot_det_stat:
954
955
956
                if len(axes) == ndim:
                    axes.append(fig.add_subplot(ndim+1, 1, ndim+1))

957
958
959
                lnl = sampler.lnlikelihood[temp, :, :]
                if burnin_idx and add_det_stat_burnin:
                    burn_in_vals = lnl[:, :burnin_idx].flatten()
960
961
                    try:
                        axes[-1].hist(burn_in_vals[~np.isnan(burn_in_vals)],
962
                                      bins=50, histtype='step', color='C3')
963
964
965
966
                    except ValueError:
                        logging.info('Det. Stat. hist failed, most likely all '
                                     'values where the same')
                        pass
967
968
969
                else:
                    burn_in_vals = []
                prod_vals = lnl[:, burnin_idx:].flatten()
970
971
972
973
974
975
976
                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
977
978
979
980
981
982
983
984
985
986
987
988
                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)

989
                xfmt = matplotlib.ticker.ScalarFormatter()
990
                xfmt.set_powerlimits((-4, 4))
991
992
                axes[-1].xaxis.set_major_formatter(xfmt)

993
994
        return fig, axes

995
    def _apply_corrections_to_p0(self, p0):
Gregory Ashton's avatar
Gregory Ashton committed
996
997
998
        """ Apply any correction to the initial p0 values """
        return p0

999
    def _generate_scattered_p0(self, p):
1000
        """ Generate a set of p0s scattered about p """
Gregory Ashton's avatar
Gregory Ashton committed
1001
        p0 = [[p + self.scatter_val * p * np.random.randn(self.ndim)
1002
1003
1004
1005
               for i in xrange(self.nwalkers)]
              for j in xrange(self.ntemps)]
        return p0

1006
    def _generate_initial_p0(self):
1007
1008
1009
        """ Generate a set of init vals for the walkers """

        if type(self.theta_initial) == dict:
1010
            logging.info('Generate initial values from initial dictionary')
1011
            if hasattr(self, 'nglitch') and self.nglitch > 1:
1012
                raise ValueError('Initial dict not implemented for nglitch>1')
1013
            p0 = [[[self._generate_rv(**self.theta_initial[key])
1014
1015
1016
                    for key in self.theta_keys]
                   for i in range(self.nwalkers)]
                  for j in range(self.ntemps)]
1017
1018
        elif type(self.theta_initial) == list:
            logging.info('Generate initial values from list of theta_initial')
1019
            p0 = [[[self._generate_rv(**val)
1020
1021
1022
                    for val in self.theta_initial]
                   for i in range(self.nwalkers)]
                  for j in range(self.ntemps)]
1023
        elif self.theta_initial is None:
1024
            logging.info('Generate initial values from prior dictionary')
1025
            p0 = [[[self._generate_rv(**self.theta_prior[key])
1026
1027
1028
1029
                    for key in self