Note
Click here to download the full example code
MCMC search: Fully coherent F-statisticΒΆ
Directed MCMC search for an isolated CW signal using the fully coherent F-statistic.
9 import os
10
11 import numpy as np
12
13 import pyfstat
14 from pyfstat.utils import get_predict_fstat_parameters_from_dict
15
16 label = "PyFstatExampleFullyCoherentMCMCSearch"
17 outdir = os.path.join("PyFstat_example_data", label)
18 logger = pyfstat.set_up_logger(label=label, outdir=outdir)
19
20 # Properties of the GW data
21 data_parameters = {
22 "sqrtSX": 1e-23,
23 "tstart": 1000000000,
24 "duration": 100 * 86400,
25 "detectors": "H1",
26 }
27 tend = data_parameters["tstart"] + data_parameters["duration"]
28 mid_time = 0.5 * (data_parameters["tstart"] + tend)
29
30 # Properties of the signal
31 depth = 10
32 signal_parameters = {
33 "F0": 30.0,
34 "F1": -1e-10,
35 "F2": 0,
36 "Alpha": np.radians(83.6292),
37 "Delta": np.radians(22.0144),
38 "tref": mid_time,
39 "h0": data_parameters["sqrtSX"] / depth,
40 "cosi": 1.0,
41 }
42
43 data = pyfstat.Writer(
44 label=label, outdir=outdir, **data_parameters, **signal_parameters
45 )
46 data.make_data()
47
48 # The predicted twoF (expectation over noise realizations) can be accessed by
49 twoF = data.predict_fstat()
50 logger.info("Predicted twoF value: {}\n".format(twoF))
51
52 DeltaF0 = 1e-7
53 DeltaF1 = 1e-13
54 VF0 = (np.pi * data_parameters["duration"] * DeltaF0) ** 2 / 3.0
55 VF1 = (np.pi * data_parameters["duration"] ** 2 * DeltaF1) ** 2 * 4 / 45.0
56 logger.info("\nV={:1.2e}, VF0={:1.2e}, VF1={:1.2e}\n".format(VF0 * VF1, VF0, VF1))
57
58 theta_prior = {
59 "F0": {
60 "type": "unif",
61 "lower": signal_parameters["F0"] - DeltaF0 / 2.0,
62 "upper": signal_parameters["F0"] + DeltaF0 / 2.0,
63 },
64 "F1": {
65 "type": "unif",
66 "lower": signal_parameters["F1"] - DeltaF1 / 2.0,
67 "upper": signal_parameters["F1"] + DeltaF1 / 2.0,
68 },
69 }
70 for key in "F2", "Alpha", "Delta":
71 theta_prior[key] = signal_parameters[key]
72
73 ntemps = 2
74 log10beta_min = -0.5
75 nwalkers = 100
76 nsteps = [300, 300]
77
78 mcmc = pyfstat.MCMCSearch(
79 label=label,
80 outdir=outdir,
81 sftfilepattern=data.sftfilepath,
82 theta_prior=theta_prior,
83 tref=mid_time,
84 minStartTime=data_parameters["tstart"],
85 maxStartTime=tend,
86 nsteps=nsteps,
87 nwalkers=nwalkers,
88 ntemps=ntemps,
89 log10beta_min=log10beta_min,
90 )
91 mcmc.transform_dictionary = dict(
92 F0=dict(subtractor=signal_parameters["F0"], symbol="$f-f^\\mathrm{s}$"),
93 F1=dict(
94 subtractor=signal_parameters["F1"], symbol="$\\dot{f}-\\dot{f}^\\mathrm{s}$"
95 ),
96 )
97 mcmc.run(
98 walker_plot_args={"plot_det_stat": True, "injection_parameters": signal_parameters}
99 )
100 mcmc.print_summary()
101 mcmc.plot_corner(add_prior=True, truths=signal_parameters)
102 mcmc.plot_prior_posterior(injection_parameters=signal_parameters)
103
104 mcmc.generate_loudest()
105
106 # plot cumulative 2F, first building a dict as required for PredictFStat
107 d, maxtwoF = mcmc.get_max_twoF()
108 for key, val in mcmc.theta_prior.items():
109 if key not in d:
110 d[key] = val
111 d["h0"] = data.h0
112 d["cosi"] = data.cosi
113 d["psi"] = data.psi
114 PFS_input = get_predict_fstat_parameters_from_dict(d)
115 mcmc.plot_cumulative_max(PFS_input=PFS_input)
Total running time of the script: ( 0 minutes 0.000 seconds)