mcmc_based_searches.py 91.7 KB
Newer Older
Gregory Ashton's avatar
Gregory Ashton committed
1
""" Searches using MCMC-based methods """
2
from __future__ import division, absolute_import, print_function
Gregory Ashton's avatar
Gregory Ashton committed
3

4
import sys
Gregory Ashton's avatar
Gregory Ashton committed
5
import os
6
import copy
Gregory Ashton's avatar
Gregory Ashton committed
7
import logging
8
from collections import OrderedDict
9
import subprocess
10
11
12
13
14
15
16
17

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

18
import pyfstat.core as core
19
20
from pyfstat.core import tqdm, args, read_par
import pyfstat.optimal_setup_functions as optimal_setup_functions
21
import pyfstat.helper_functions as helper_functions
22
23


24
class MCMCSearch(core.BaseSearchClass):
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
    """ MCMC search using ComputeFstat

    Parameters
    ----------
    label, outdir: str
        A label and directory to read/write data from/to
    sftfilepattern: str
        Pattern to match SFTs using wildcards (*?) and ranges [0-9];
        mutiple patterns can be given separated by colons.
    theta_prior: dict
        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.
    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.
    tref, minStartTime, maxStartTime: int
        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].
    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).
    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.
    binary: Bool
        If true, search over binary parameters
    detectors: str
        Two character reference to the data to use, specify None for no
        contraint.
    minCoverFreq, maxCoverFreq: float
        Minimum and maximum instantaneous frequency which will be covered
        over the SFT time span as passed to CreateFstatInput

    Attributes
    ----------
    symbol_dictionary: dict
        Key, val pairs of the parameters (i.e. `F0`, `F1`), to Latex math
        symbols for plots
    unit_dictionary: dict
        Key, val pairs of the parameters (i.e. `F0`, `F1`), and the
        units (i.e. `Hz`)
    transform_dictionary: dict
        Key, val pairs of the parameters (i.e. `F0`, `F1`), where the key is
        itself a dictionary which can item `multiplier`, `subtractor`, or
        `unit` by which to transform by and update the units.

    """
83
84

    symbol_dictionary = dict(
85
        F0='$f$', F1='$\dot{f}$', F2='$\ddot{f}$', Alpha=r'$\alpha$',
86
87
        Delta='$\delta$', asini='asini', period='P', ecc='ecc', tp='tp',
        argp='argp')
88
    unit_dictionary = dict(
89
90
        F0='Hz', F1='Hz/s', F2='Hz/s$^2$', Alpha=r'rad', Delta='rad',
        asini='', period='s', ecc='', tp='', argp='')
91
    transform_dictionary = {}
92

Gregory Ashton's avatar
Gregory Ashton committed
93
    @helper_functions.initializer
Gregory Ashton's avatar
Gregory Ashton committed
94
    def __init__(self, label, outdir, theta_prior, tref, minStartTime,
95
                 maxStartTime, sftfilepattern=None, nsteps=[100, 100],
96
                 nwalkers=100, ntemps=1, log10temperature_min=-5,
97
                 theta_initial=None, scatter_val=1e-10, rhohatmax=1000,
98
                 binary=False, BSGL=False, minCoverFreq=None, SSBprec=None,
99
100
                 maxCoverFreq=None, detectors=None,
                 injectSources=None, assumeSqrtSX=None):
101

Gregory Ashton's avatar
Gregory Ashton committed
102
103
        if os.path.isdir(outdir) is False:
            os.mkdir(outdir)
104
        self._add_log_file()
105
        logging.info('Set-up MCMC search for model {}'.format(self.label))
106
107
        if sftfilepattern:
            logging.info('Using data {}'.format(self.sftfilepattern))
108
        else:
109
            logging.info('No sftfilepattern given')
110
111
        if injectSources:
            logging.info('Inject sources: {}'.format(injectSources))
112
        self.pickle_path = '{}/{}_saved_data.p'.format(self.outdir, self.label)
113
        self._unpack_input_theta()
114
        self.ndim = len(self.theta_keys)
115
116
117
118
        if self.log10temperature_min:
            self.betas = np.logspace(0, self.log10temperature_min, self.ntemps)
        else:
            self.betas = None
119

120
121
122
        if args.clean and os.path.isfile(self.pickle_path):
            os.rename(self.pickle_path, self.pickle_path+".old")

123
        self._set_likelihoodcoef()
124
        self._log_input()
125
126
127

    def _set_likelihoodcoef(self):
        self.likelihoodcoef = np.log(70./self.rhohatmax**4)
128

129
    def _log_input(self):
130
        logging.info('theta_prior = {}'.format(self.theta_prior))
131
        logging.info('nwalkers={}'.format(self.nwalkers))
132
133
134
135
        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(
136
            self.log10temperature_min))
137

138
    def _initiate_search_object(self):
139
        logging.info('Setting up search object')
140
        self.search = core.ComputeFstat(
141
            tref=self.tref, sftfilepattern=self.sftfilepattern,
142
            minCoverFreq=self.minCoverFreq, maxCoverFreq=self.maxCoverFreq,
143
            detectors=self.detectors, BSGL=self.BSGL, transient=False,
144
            minStartTime=self.minStartTime, maxStartTime=self.maxStartTime,
145
            binary=self.binary, injectSources=self.injectSources,
146
            assumeSqrtSX=self.assumeSqrtSX, SSBprec=self.SSBprec)
147
148

    def logp(self, theta_vals, theta_prior, theta_keys, search):
149
        H = [self._generic_lnprior(**theta_prior[key])(p) for p, key in
150
151
152
153
154
155
             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]
156
157
        FS = search.compute_fullycoherent_det_stat_single_point(
            *self.fixed_theta)
158
        return FS + self.likelihoodcoef
159

160
    def _unpack_input_theta(self):
161
        full_theta_keys = ['F0', 'F1', 'F2', 'Alpha', 'Delta']
162
163
164
        if self.binary:
            full_theta_keys += [
                'asini', 'period', 'ecc', 'tp', 'argp']
165
166
        full_theta_keys_copy = copy.copy(full_theta_keys)

167
168
        full_theta_symbols = ['$f$', '$\dot{f}$', '$\ddot{f}$', r'$\alpha$',
                              r'$\delta$']
169
170
        if self.binary:
            full_theta_symbols += [
171
                'asini', 'period', 'ecc', 'tp', 'argp']
172

173
174
        self.theta_keys = []
        fixed_theta_dict = {}
175
        for key, val in self.theta_prior.iteritems():
176
177
            if type(val) is dict:
                fixed_theta_dict[key] = 0
Gregory Ashton's avatar
Gregory Ashton committed
178
                self.theta_keys.append(key)
179
180
181
182
183
184
            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
185
            full_theta_keys_copy.pop(full_theta_keys_copy.index(key))
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200

        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]

201
    def _check_initial_points(self, p0):
202
203
204
205
206
207
208
209
210
211
212
213
214
        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))

215
                p0 = self._generate_new_p0_to_fix_initial_points(
216
217
                    p0, nt, initial_priors)

218
    def _generate_new_p0_to_fix_initial_points(self, p0, nt, initial_priors):
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
        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
238

239
240
    def setup_burnin_convergence_testing(
            self, n=10, test_type='autocorr', windowed=False, **kwargs):
241
        """ Set up convergence testing during the MCMC simulation
242
243
244

        Parameters
        ----------
245
246
247
248
249
250
251
252
253
        n: int
            Number of steps after which to test convergence
        test_type: str ['autocorr', 'GR']
            If 'autocorr' use the exponential autocorrelation time (kwargs
            passed to `get_autocorr_convergence`). If 'GR' use the Gelman-Rubin
            statistic (kwargs passed to `get_GR_convergence`)
        windowed: bool
            If True, only calculate the convergence test in a window of length
            `n`
254
255
256
257
        **kwargs:
            Passed to either `_test_autocorr_convergence()` or
            `_test_GR_convergence()` depending on `test_type`.

258
        """
259
        logging.info('Setting up convergence testing')
260
261
262
263
        self.convergence_n = n
        self.convergence_windowed = windowed
        self.convergence_test_type = test_type
        self.convergence_kwargs = kwargs
264
265
        self.convergence_diagnostic = []
        self.convergence_diagnosticx = []
266
        if test_type in ['autocorr']:
267
            self._get_convergence_test = self._test_autocorr_convergence
268
        elif test_type in ['GR']:
269
            self._get_convergence_test = self._test_GR_convergence
270
271
272
        else:
            raise ValueError('test_type {} not understood'.format(test_type))

273
    def _test_autocorr_convergence(self, i, sampler, test=True, n_cut=5):
274
275
276
277
278
279
280
281
282
283
284
        try:
            acors = np.zeros((self.ntemps, self.ndim))
            for temp in range(self.ntemps):
                if self.convergence_windowed:
                    j = i-self.convergence_n
                else:
                    j = 0
                x = np.mean(sampler.chain[temp, :, j:i, :], axis=0)
                acors[temp, :] = emcee.autocorr.exponential_time(x)
            c = np.max(acors, axis=0)
        except emcee.autocorr.AutocorrError:
Gregory Ashton's avatar
Gregory Ashton committed
285
286
287
288
            logging.info('Failed to calculate exponential autocorrelation')
            c = np.zeros(self.ndim) + np.nan
        except AttributeError:
            logging.info('Unable to calculate exponential autocorrelation')
289
290
291
292
293
294
295
296
            c = np.zeros(self.ndim) + np.nan

        self.convergence_diagnosticx.append(i - self.convergence_n/2.)
        self.convergence_diagnostic.append(list(c))

        if test:
            return i > n_cut * np.max(c)

297
    def _test_GR_convergence(self, i, sampler, test=True, R=1.1):
298
299
300
301
302
        if self.convergence_windowed:
            s = sampler.chain[0, :, i-self.convergence_n+1:i+1, :]
        else:
            s = sampler.chain[0, :, :i+1, :]
        N = float(self.convergence_n)
303
304
        M = float(self.nwalkers)
        W = np.mean(np.var(s, axis=1), axis=0)
305
306
        per_walker_mean = np.mean(s, axis=1)
        mean = np.mean(per_walker_mean, axis=0)
307
308
        B = N / (M-1.) * np.sum((per_walker_mean-mean)**2, axis=0)
        Vhat = (N-1)/N * W + (M+1)/(M*N) * B
309
        c = np.sqrt(Vhat/W)
310
        self.convergence_diagnostic.append(c)
311
        self.convergence_diagnosticx.append(i - self.convergence_n/2.)
312

313
314
315
        if test and np.max(c) < R:
            return True
        else:
316
            return False
317
318
319
320

    def _test_convergence(self, i, sampler, **kwargs):
        if np.mod(i+1, self.convergence_n) == 0:
            return self._get_convergence_test(i, sampler, **kwargs)
321
        else:
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
            return False

    def _run_sampler_with_conv_test(self, sampler, p0, nprod=0, nburn=0):
        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):
            if self._test_convergence(i, sampler, test=True,
                                      **self.convergence_kwargs):
                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
        iterator = tqdm(sampler.sample(output[0], iterations=nprod),
                        total=nprod)
        for result in iterator:
            self._test_convergence(j, sampler, test=False,
                                   **self.convergence_kwargs)
            j += 1
        return sampler
346

347
    def _run_sampler(self, sampler, p0, nprod=0, nburn=0):
348
349
        if hasattr(self, 'convergence_n'):
            self._run_sampler_with_conv_test(sampler, p0, nprod, nburn)
350
351
352
353
        else:
            for result in tqdm(sampler.sample(p0, iterations=nburn+nprod),
                               total=nburn+nprod):
                pass
354

355
356
        self.mean_acceptance_fraction = np.mean(
            sampler.acceptance_fraction, axis=1)
357
        logging.info("Mean acceptance fraction: {}"
358
                     .format(self.mean_acceptance_fraction))
359
        if self.ntemps > 1:
360
            self.tswap_acceptance_fraction = sampler.tswap_acceptance_fraction
361
362
363
            logging.info("Tswap acceptance fraction: {}"
                         .format(sampler.tswap_acceptance_fraction))
        try:
364
            self.autocorr_time = sampler.get_autocorr_time(c=4)
365
            logging.info("Autocorrelation length: {}".format(
366
                self.autocorr_time))
367
        except emcee.autocorr.AutocorrError as e:
368
            self.autocorr_time = np.nan
369
370
371
372
373
            logging.warning(
                'Autocorrelation calculation failed with message {}'.format(e))

        return sampler

374
    def _estimate_run_time(self):
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
        """ Print the estimated run time

        Uses timing coefficients based on a Lenovo T460p Intel(R)
        Core(TM) i5-6300HQ CPU @ 2.30GHz.

        """
        # Todo: add option to time on a machine, and move coefficients to
        # ~/.pyfstat.conf
        if (type(self.theta_prior['Alpha']) == dict or
                type(self.theta_prior['Delta']) == dict):
            tau0S = 7.3e-5
            tau0LD = 4.2e-7
        else:
            tau0S = 5.0e-5
            tau0LD = 6.2e-8
390
        Nsfts = (self.maxStartTime - self.minStartTime) / 1800.
391
392
393
        numb_evals = np.sum(self.nsteps)*self.nwalkers*self.ntemps
        a = tau0S * numb_evals
        b = tau0LD * Nsfts * numb_evals
394
395
396
        logging.info('Estimated run-time = {} s = {:1.0f}:{:1.0f} m'.format(
            a+b, *divmod(a+b, 60)))

397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
    def run(self, proposal_scale_factor=2, create_plots=True, c=5, **kwargs):
        """ Run the MCMC simulatation

        Parameters
        ----------
        proposal_scale_factor: float
            The proposal scale factor used by the sampler, see Goodman & Weare
            (2010). If the acceptance fraction is too low, you can raise it by
            decreasing the a parameter; and if it is too high, you can reduce
            it by increasing the a parameter [Foreman-Mackay (2013)].
        create_plots: bool
            If true, save trace plots of the walkers
        c: int
            The minimum number of autocorrelation times needed to trust the
            result when estimating the autocorrelation time (see
            emcee.autocorr.integrated_time for further details. Default is 5
        **kwargs:
            Passed to _plot_walkers to control the figures

416
417
418
419
420
        Returns
        -------
        sampler: emcee.ptsampler.PTSampler
            The emcee ptsampler object

421
        """
422

423
        self.old_data_is_okay_to_use = self._check_old_data_is_okay_to_use()
424
425
426
        if self.old_data_is_okay_to_use is True:
            logging.warning('Using saved data from {}'.format(
                self.pickle_path))
427
            d = self.get_saved_data_dictionary()
428
429
430
            self.samples = d['samples']
            self.lnprobs = d['lnprobs']
            self.lnlikes = d['lnlikes']
431
            self.all_lnlikelihood = d['all_lnlikelihood']
432
433
            return

434
        self._initiate_search_object()
435
        self._estimate_run_time()
436
437
438
439

        sampler = emcee.PTSampler(
            self.ntemps, self.nwalkers, self.ndim, self.logl, self.logp,
            logpargs=(self.theta_prior, self.theta_keys, self.search),
440
            loglargs=(self.search,), betas=self.betas, a=proposal_scale_factor)
441

442
443
444
        p0 = self._generate_initial_p0()
        p0 = self._apply_corrections_to_p0(p0)
        self._check_initial_points(p0)
445
446
447
448

        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
449
                j, ninit_steps, n))
450
            sampler = self._run_sampler(sampler, p0, nburn=n)
451
            if create_plots:
452
                fig, axes = self._plot_walkers(sampler,
453
454
                                               symbols=self.theta_symbols,
                                               **kwargs)
455
456
                fig.tight_layout()
                fig.savefig('{}/{}_init_{}_walkers.png'.format(
Gregory Ashton's avatar
Gregory Ashton committed
457
                    self.outdir, self.label, j))
458

459
460
461
            p0 = self._get_new_p0(sampler)
            p0 = self._apply_corrections_to_p0(p0)
            self._check_initial_points(p0)
462
463
            sampler.reset()

Gregory Ashton's avatar
Gregory Ashton committed
464
465
466
467
        if len(self.nsteps) > 1:
            nburn = self.nsteps[-2]
        else:
            nburn = 0
468
469
470
        nprod = self.nsteps[-1]
        logging.info('Running final burn and prod with {} steps'.format(
            nburn+nprod))
471
        sampler = self._run_sampler(sampler, p0, nburn=nburn, nprod=nprod)
472
        if create_plots:
473
            fig, axes = self._plot_walkers(sampler, symbols=self.theta_symbols,
474
                                           nprod=nprod, **kwargs)
475
476
            fig.tight_layout()
            fig.savefig('{}/{}_walkers.png'.format(self.outdir, self.label),
Gregory Ashton's avatar
Gregory Ashton committed
477
                        )
478
479
480
481

        samples = sampler.chain[0, :, nburn:, :].reshape((-1, self.ndim))
        lnprobs = sampler.lnprobability[0, :, nburn:].reshape((-1))
        lnlikes = sampler.lnlikelihood[0, :, nburn:].reshape((-1))
482
        all_lnlikelihood = sampler.lnlikelihood[:, :, nburn:]
483
484
485
        self.samples = samples
        self.lnprobs = lnprobs
        self.lnlikes = lnlikes
486
487
        self.all_lnlikelihood = all_lnlikelihood
        self._save_data(sampler, samples, lnprobs, lnlikes, all_lnlikelihood)
Gregory Ashton's avatar
Gregory Ashton committed
488
        return sampler
489

490
    def _get_rescale_multiplier_for_key(self, key):
491
        """ Get the rescale multiplier from the transform_dictionary
492
493
494
495
496

        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
        """
497
        if key not in self.transform_dictionary:
498
499
            return 1

500
501
        if 'multiplier' in self.transform_dictionary[key]:
            val = self.transform_dictionary[key]['multiplier']
502
503
504
            if type(val) == str:
                if hasattr(self, val):
                    multiplier = getattr(
505
                        self, self.transform_dictionary[key]['multiplier'])
506
507
508
509
510
511
512
513
514
                else:
                    raise ValueError(
                        "multiplier {} not a class attribute".format(val))
            else:
                multiplier = val
        else:
            multiplier = 1
        return multiplier

515
    def _get_rescale_subtractor_for_key(self, key):
516
        """ Get the rescale subtractor from the transform_dictionary
517
518
519
520
521

        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
        """
522
        if key not in self.transform_dictionary:
523
524
            return 0

525
526
        if 'subtractor' in self.transform_dictionary[key]:
            val = self.transform_dictionary[key]['subtractor']
527
528
529
            if type(val) == str:
                if hasattr(self, val):
                    subtractor = getattr(
530
                        self, self.transform_dictionary[key]['subtractor'])
531
532
533
534
535
536
537
538
539
                else:
                    raise ValueError(
                        "subtractor {} not a class attribute".format(val))
            else:
                subtractor = val
        else:
            subtractor = 0
        return subtractor

540
    def _scale_samples(self, samples, theta_keys):
541
        """ Scale the samples using the transform_dictionary """
542
        for key in theta_keys:
543
            if key in self.transform_dictionary:
544
545
                idx = theta_keys.index(key)
                s = samples[:, idx]
546
                subtractor = self._get_rescale_subtractor_for_key(key)
547
                s = s - subtractor
548
                multiplier = self._get_rescale_multiplier_for_key(key)
549
                s *= multiplier
550
551
                samples[:, idx] = s

552
553
        return samples

554
    def _get_labels(self):
555
        """ Combine the units, symbols and rescaling to give labels """
556

557
558
559
560
561
562
        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]
563
564
565
566
567
568
569
            if key in self.transform_dictionary:
                if 'symbol' in self.transform_dictionary[key]:
                    s = self.transform_dictionary[key]['symbol']
                if 'label' in self.transform_dictionary[key]:
                    label = self.transform_dictionary[key]['label']
                if 'unit' in self.transform_dictionary[key]:
                    u = self.transform_dictionary[key]['unit']
570
571
572
573
            if label is None:
                label = '{} \n [{}]'.format(s, u)
            labels.append(label)
        return labels
574

575
576
    def plot_corner(self, figsize=(7, 7), add_prior=False, nstds=None,
                    label_offset=0.4, dpi=300, rc_context={},
577
                    tglitch_ratio=False, fig_and_axes=None, save_fig=True,
578
                    **kwargs):
579
580
581
582
583
584
585
586
587
        """ 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)
588
589
590
        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.
591
592
593
594
595
596
597
598
599
600
601
602
603
604
        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
605
606
607
608
609
        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
610
611
        **kwargs:
            Passed to corner.corner
612

613
614
615
616
        Returns
        -------
        fig, axes:
            The matplotlib figure and axes, only returned if save_fig = False
617
618

        """
619

620
621
622
623
        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
624
625
        if self.ndim < 2:
            with plt.rc_context(rc_context):
626
627
628
629
                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
630
631
632
633
634
635
636
                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

637
        with plt.rc_context(rc_context):
638
639
640
641
642
            if fig_and_axes is None:
                fig, axes = plt.subplots(self.ndim, self.ndim,
                                         figsize=figsize)
            else:
                fig, axes = fig_and_axes
643
644

            samples_plt = copy.copy(self.samples)
645
            labels = self._get_labels()
646

647
            samples_plt = self._scale_samples(samples_plt, self.theta_keys)
648
649
650
651
652

            if tglitch_ratio:
                for j, k in enumerate(self.theta_keys):
                    if k == 'tglitch':
                        s = samples_plt[:, j]
653
654
655
                        samples_plt[:, j] = (
                            s - self.minStartTime)/(
                                self.maxStartTime - self.minStartTime)
656
                        labels[j] = r'$R_{\textrm{glitch}}$'
657
658
659
660
661
662
663

            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))
664
665
            elif 'range' in kwargs:
                _range = kwargs.pop('range')
666
667
668
            else:
                _range = None

669
670
671
672
            hist_kwargs = kwargs.pop('hist_kwargs', dict())
            if 'normed' not in hist_kwargs:
                hist_kwargs['normed'] = True

673
            fig_triangle = corner.corner(samples_plt,
674
                                         labels=labels,
675
676
677
678
679
680
681
682
683
                                         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,
684
                                         hist_kwargs=hist_kwargs,
685
686
687
688
689
690
691
692
693
694
695
696
697
698
699
700
                                         **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:
701
                self._add_prior_to_corner(axes, self.samples, add_prior)
702

703
704
705
706
707
            if save_fig:
                fig_triangle.savefig('{}/{}_corner.png'.format(
                    self.outdir, self.label), dpi=dpi)
            else:
                return fig, axes
708

709
    def _add_prior_to_corner(self, axes, samples, add_prior):
710
711
712
        for i, key in enumerate(self.theta_keys):
            ax = axes[i][i]
            s = samples[:, i]
713
714
715
716
717
718
719
720
721
722
            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)
723
724
            multiplier = self._get_rescale_multiplier_for_key(key)
            subtractor = self._get_rescale_subtractor_for_key(key)
725
726
727
728
729
730
731
732
            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])
733

734
735
736
737
738
739
740
741
    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]
742
            prior_func = self._generic_lnprior(**prior_dict)
743
744
745
746
747
            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
748
749
750
751
752
            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]
753
754
755
756
757
            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)
758
759
760
761
762
            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
763
764
765
766
767
            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]
768
769
770
            else:
                raise ValueError('Not implemented for prior type {}'.format(
                    prior_dict['type']))
771
            priorln = ax.plot(x, prior, 'C3', label='prior')
772
773
774
775
776
777
778
779
780
781
782
783
784
785
786
787
788
789
790
            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))

791
    def plot_cumulative_max(self, **kwargs):
792
793
794
795
        """ 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
796
797
798
799
        d, maxtwoF = self.get_max_twoF()
        for key, val in self.theta_prior.iteritems():
            if key not in d:
                d[key] = val
800
801

        if hasattr(self, 'search') is False:
802
            self._initiate_search_object()
803
804
805
        if self.binary is False:
            self.search.plot_twoF_cumulative(
                self.label, self.outdir, F0=d['F0'], F1=d['F1'], F2=d['F2'],
806
                Alpha=d['Alpha'], Delta=d['Delta'],
807
                tstart=self.minStartTime, tend=self.maxStartTime,
808
                **kwargs)
809
810
811
812
813
        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'],
814
                tstart=self.minStartTime, tend=self.maxStartTime, **kwargs)
Gregory Ashton's avatar
Gregory Ashton committed
815

816
    def _generic_lnprior(self, **kwargs):
817
818
819
820
        """ Return a lambda function of the pdf

        Parameters
        ----------
821
        **kwargs:
822
823
824
825
            A dictionary containing 'type' of pdf and shape parameters

        """

Gregory Ashton's avatar
Gregory Ashton committed
826
        def log_of_unif(x, a, b):
827
828
829
830
831
832
833
834
835
836
837
838
839
            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
840
841
842
843
844
845
846
847
848
849
850
851
852
853
854
855
        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):
856
            if x < loc:
857
858
859
860
861
862
863
864
865
866
867
868
869
870
                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
871
872
873
874
            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'])
875
        elif kwargs['type'] == 'halfnorm':
Gregory Ashton's avatar
Gregory Ashton committed
876
            return lambda x: log_of_halfnorm(x, kwargs['loc'], kwargs['scale'])
877
        elif kwargs['type'] == 'neghalfnorm':
Gregory Ashton's avatar
Gregory Ashton committed
878
879
            return lambda x: log_of_halfnorm(
                -x, kwargs['loc'], kwargs['scale'])
880
881
882
883
884
885
886
        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")

887
    def _generate_rv(self, **kwargs):
888
889
890
        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
891
892
893
        if dist_type == "log10unif":
            return 10**(np.random.uniform(low=kwargs['log10lower'],
                                          high=kwargs['log10upper']))
894
895
896
897
898
        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']))
899
900
901
        if dist_type == "neghalfnorm":
            return -1 * np.abs(np.random.normal(loc=kwargs['loc'],
                                                scale=kwargs['scale']))
902
903
904
905
906
907
        if dist_type == "lognorm":
            return np.random.lognormal(
                mean=kwargs['loc'], sigma=kwargs['scale'])
        else:
            raise ValueError("dist_type {} unknown".format(dist_type))

908
    def _plot_walkers(self, sampler, symbols=None, alpha=0.8, color="k",
909
910
                      temp=0, lw=0.1, nprod=0, add_det_stat_burnin=False,
                      fig=None, axes=None, xoffset=0, plot_det_stat=False,
911
                      context='ggplot', subtractions=None, labelpad=0.05):
912
913
        """ Plot all the chains from a sampler """

914
915
916
917
918
        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))

919
920
921
        if np.ndim(axes) > 1:
            axes = axes.flatten()

922
923
924
925
926
927
928
929
930
931
932
933
934
        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, :, :, :]

935
936
        if subtractions is None:
            subtractions = [0 for i in range(ndim)]
937
938
939
        else:
            if len(subtractions) != self.ndim:
                raise ValueError('subtractions must be of length ndim')
940

941
942
943
944
        if plot_det_stat:
            extra_subplots = 1
        else:
            extra_subplots = 0
945
        with plt.style.context((context)):
Gregory Ashton's avatar
Gregory Ashton committed
946
            plt.rcParams['text.usetex'] = True
Gregory Ashton's avatar
Gregory Ashton committed
947
            if fig is None and axes is None:
948
                fig = plt.figure(figsize=(4, 3.0*ndim))
949
950
                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
951
                               for i in range(2, ndim+1)]
952

Gregory Ashton's avatar
Gregory Ashton committed
953
            idxs = np.arange(chain.shape[1])
954
955
956
957
958
            burnin_idx = chain.shape[1] - nprod
            if hasattr(self, 'convergence_idx'):
                convergence_idx = self.convergence_idx
            else:
                convergence_idx = burnin_idx
959
960
            if ndim > 1:
                for i in range(ndim):
961
                    axes[i].ticklabel_format(useOffset=False, axis='y')
Gregory Ashton's avatar
Gregory Ashton committed
962
                    cs = chain[:, :, i].T
963
                    if burnin_idx > 0:
964
965
                        axes[i].plot(xoffset+idxs[:convergence_idx+1],
                                     cs[:convergence_idx+1]-subtractions[i],
966
                                     color="C3", alpha=alpha,
Gregory Ashton's avatar
Gregory Ashton committed
967
                                     lw=lw)
968
                        axes[i].axvline(xoffset+convergence_idx,
969
                                        color='k', ls='--', lw=0.25)
970
971
                    axes[i].plot(xoffset+idxs[burnin_idx:],
                                 cs[burnin_idx:]-subtractions[i],
Gregory Ashton's avatar
Gregory Ashton committed
972
                                 color="k", alpha=alpha, lw=lw)
Gregory Ashton's avatar
Gregory Ashton committed
973
974

                    axes[i].set_xlim(0, xoffset+idxs[-1])
975
                    if symbols:
976
                        if subtractions[i] == 0:
977
                            axes[i].set_ylabel(symbols[i], labelpad=labelpad)
978
979
                        else:
                            axes[i].set_ylabel(
980
981
                                symbols[i]+'$-$'+symbols[i]+'$_0$',
                                labelpad=labelpad)
982

983
984
                    if hasattr(self, 'convergence_diagnostic'):
                        ax = axes[i].twinx()
985
986
                        axes[i].set_zorder(ax.get_zorder()+1)
                        axes[i].patch.set_visible(False)
987
988
                        c_x = np.array(self.convergence_diagnosticx)
                        c_y = np.array(self.convergence_diagnostic)
989
                        break_idx = np.argmin(np.abs(c_x - burnin_idx))
990
991
992
993
                        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)
994
995
996
997
                        if self.convergence_test_type == 'autocorr':
                            ax.set_ylabel(r'$\tau_\mathrm{exp}$')
                        elif self.convergence_test_type == 'GR':
                            ax.set_ylabel('PSRF')
998
                        ax.ticklabel_format(useOffset=False)
999
            else:
Gregory Ashton's avatar
Gregory Ashton committed
1000
                axes[0].ticklabel_format(useOffset=False, axis='y')
Gregory Ashton's avatar
Gregory Ashton committed
1001
                cs = chain[:, :, temp].T
Gregory Ashton's avatar
Gregory Ashton committed
1002
1003
                if burnin_idx:
                    axes[0].plot(idxs[:burnin_idx], cs[:burnin_idx],
1004
                                 color="C3", alpha=alpha, lw=lw)
Gregory Ashton's avatar
Gregory Ashton committed
1005
1006
1007
                axes[0].plot(idxs[burnin_idx:], cs[burnin_idx:], color="k",
                             alpha=alpha, lw=lw)
                if symbols:
1008
                    axes[0].set_ylabel(symbols[0], labelpad=labelpad)
1009

Gregory Ashton's avatar
Gregory Ashton committed
1010
1011
            axes[-1].set_xlabel(r'$\textrm{Number of steps}$', labelpad=0.2)

1012
            if plot_det_stat:
1013
1014
1015
                if len(axes) == ndim:
                    axes.append(fig.add_subplot(ndim+1, 1, ndim+1))

1016
1017
1018
                lnl = sampler.lnlikelihood[temp, :, :]
                if burnin_idx and add_det_stat_burnin:
                    burn_in_vals = lnl[:, :burnin_idx].flatten()
1019
                    try:
1020
1021
1022
1023
                        twoF_burnin = (burn_in_vals[~np.isnan(burn_in_vals)]
                                       - self.likelihoodcoef)
                        axes[-1].hist(twoF_burnin, bins=50, histtype='step',
                                      color='C3')
1024
1025
1026
1027
                    except ValueError:
                        logging.info('Det. Stat. hist failed, most likely all '
                                     'values where the same')
                        pass
1028
                else:
1029
                    twoF_burnin = []
1030
                prod_vals = lnl[:, burnin_idx:].flatten()
1031
                try:
1032
1033
                    twoF = prod_vals[~np.isnan(prod_vals)]-self.likelihoodcoef
                    axes[-1].hist(twoF, bins=50, histtype='step', color='k')
1034
1035
1036
1037
                except ValueError:
                    logging.info('Det. Stat. hist failed, most likely all '
                                 'values where the same')
                    pass
1038
1039
1040
1041
1042
                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'$\textrm{Counts}$')
1043
                combined_vals = np.append(twoF_burnin, twoF)
1044
1045
1046
1047
1048
1049
                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)

1050
                xfmt = matplotlib.ticker.ScalarFormatter()
1051
                xfmt.set_powerlimits((-4, 4))
1052
1053
                axes[-1].xaxis.set_major_formatter(xfmt)

1054
1055
        return fig, axes

1056
    def _apply_corrections_to_p0(self, p0):
Gregory Ashton's avatar
Gregory Ashton committed
1057
1058
1059
        """ Apply any correction to the initial p0 values """
        return p0

1060
    def _generate_scattered_p0(self, p):
1061
        """ Generate a set of p0s scattered about p """
Gregory Ashton's avatar
Gregory Ashton committed
1062
        p0 = [[p + self.scatter_val * p * np.random.randn(self.ndim)