PyCBC 2.9.0__tar.gz → 2.10.0__tar.gz

{PyCBC-2.9.0 → pycbc-2.10.0}/PKG-INFO

@@ -1,28 +1,64 @@
-Metadata-Version: 2.1
+Metadata-Version: 2.4
 Name: PyCBC
-Version: 2.9.0
+Version: 2.10.0
 Summary: Core library to analyze gravitational-wave data, find signals, and study their parameters.
 Home-page: http://pycbc.org/
-Download-URL: https://github.com/gwastro/pycbc/tarball/v2.9.0
+Download-URL: https://github.com/gwastro/pycbc/tarball/v2.10.0
 Author: The PyCBC team
 Author-email: alex.nitz@gmail.org
 Keywords: ligo,physics,gravity,signal processing,gravitational waves
 Classifier: Programming Language :: Python
-Classifier: Programming Language :: Python :: 3.9
-Classifier: Programming Language :: Python :: 3.10
 Classifier: Programming Language :: Python :: 3.11
 Classifier: Programming Language :: Python :: 3.12
+Classifier: Programming Language :: Python :: 3.13
 Classifier: Intended Audience :: Science/Research
 Classifier: Natural Language :: English
 Classifier: Topic :: Scientific/Engineering
 Classifier: Topic :: Scientific/Engineering :: Astronomy
 Classifier: Topic :: Scientific/Engineering :: Physics
 Classifier: License :: OSI Approved :: GNU General Public License v3 (GPLv3)
-Requires-Python: >=3.9
+Requires-Python: >=3.11
 Description-Content-Type: text/markdown
+License-File: LICENSE
+Requires-Dist: numpy>=1.16.0
+Requires-Dist: cython>=0.29
+Requires-Dist: numpy!=1.19.0,!=2.2.2,>=1.16.0
+Requires-Dist: scipy>=0.16.0
+Requires-Dist: astropy!=4.0.5,!=4.2.1,>=2.0.3
+Requires-Dist: matplotlib>=1.5.1
+Requires-Dist: mpld3>=0.3
+Requires-Dist: pillow
+Requires-Dist: h5py!=3.7.0,>=3.0.0
+Requires-Dist: jinja2
+Requires-Dist: Mako>=1.0.1
+Requires-Dist: beautifulsoup4>=4.6.0
+Requires-Dist: tqdm
+Requires-Dist: setuptools
+Requires-Dist: gwdatafind
+Requires-Dist: pegasus-wms.api>=5.1.1
+Requires-Dist: igwn-ligolw
+Requires-Dist: igwn-segments
+Requires-Dist: lalsuite!=7.2
+Requires-Dist: lscsoft-glue>=1.59.3
+Requires-Dist: pykerr
 Provides-Extra: cuda
+Requires-Dist: pycuda>=2015.1; extra == "cuda"
+Requires-Dist: scikit-cuda; extra == "cuda"
 Provides-Extra: igwn
-License-File: LICENSE
+Requires-Dist: ciecplib>=0.7.0; extra == "igwn"
+Dynamic: author
+Dynamic: author-email
+Dynamic: classifier
+Dynamic: description
+Dynamic: description-content-type
+Dynamic: download-url
+Dynamic: home-page
+Dynamic: keywords
+Dynamic: license-file
+Dynamic: provides-extra
+Dynamic: requires-dist
+Dynamic: requires-python
+Dynamic: summary
 
 ![GW150914](https://raw.githubusercontent.com/gwastro/pycbc-logo/master/pycbc_logo_name.png)
 

{PyCBC-2.9.0 → pycbc-2.10.0}/PyCBC.egg-info/PKG-INFO

@@ -1,28 +1,64 @@
-Metadata-Version: 2.1
+Metadata-Version: 2.4
 Name: PyCBC
-Version: 2.9.0
+Version: 2.10.0
 Summary: Core library to analyze gravitational-wave data, find signals, and study their parameters.
 Home-page: http://pycbc.org/
-Download-URL: https://github.com/gwastro/pycbc/tarball/v2.9.0
+Download-URL: https://github.com/gwastro/pycbc/tarball/v2.10.0
 Author: The PyCBC team
 Author-email: alex.nitz@gmail.org
 Keywords: ligo,physics,gravity,signal processing,gravitational waves
 Classifier: Programming Language :: Python
-Classifier: Programming Language :: Python :: 3.9
-Classifier: Programming Language :: Python :: 3.10
 Classifier: Programming Language :: Python :: 3.11
 Classifier: Programming Language :: Python :: 3.12
+Classifier: Programming Language :: Python :: 3.13
 Classifier: Intended Audience :: Science/Research
 Classifier: Natural Language :: English
 Classifier: Topic :: Scientific/Engineering
 Classifier: Topic :: Scientific/Engineering :: Astronomy
 Classifier: Topic :: Scientific/Engineering :: Physics
 Classifier: License :: OSI Approved :: GNU General Public License v3 (GPLv3)
-Requires-Python: >=3.9
+Requires-Python: >=3.11
 Description-Content-Type: text/markdown
+License-File: LICENSE
+Requires-Dist: numpy>=1.16.0
+Requires-Dist: cython>=0.29
+Requires-Dist: numpy!=1.19.0,!=2.2.2,>=1.16.0
+Requires-Dist: scipy>=0.16.0
+Requires-Dist: astropy!=4.0.5,!=4.2.1,>=2.0.3
+Requires-Dist: matplotlib>=1.5.1
+Requires-Dist: mpld3>=0.3
+Requires-Dist: pillow
+Requires-Dist: h5py!=3.7.0,>=3.0.0
+Requires-Dist: jinja2
+Requires-Dist: Mako>=1.0.1
+Requires-Dist: beautifulsoup4>=4.6.0
+Requires-Dist: tqdm
+Requires-Dist: setuptools
+Requires-Dist: gwdatafind
+Requires-Dist: pegasus-wms.api>=5.1.1
+Requires-Dist: igwn-ligolw
+Requires-Dist: igwn-segments
+Requires-Dist: lalsuite!=7.2
+Requires-Dist: lscsoft-glue>=1.59.3
+Requires-Dist: pykerr
 Provides-Extra: cuda
+Requires-Dist: pycuda>=2015.1; extra == "cuda"
+Requires-Dist: scikit-cuda; extra == "cuda"
 Provides-Extra: igwn
-License-File: LICENSE
+Requires-Dist: ciecplib>=0.7.0; extra == "igwn"
+Dynamic: author
+Dynamic: author-email
+Dynamic: classifier
+Dynamic: description
+Dynamic: description-content-type
+Dynamic: download-url
+Dynamic: home-page
+Dynamic: keywords
+Dynamic: license-file
+Dynamic: provides-extra
+Dynamic: requires-dist
+Dynamic: requires-python
+Dynamic: summary
 
 ![GW150914](https://raw.githubusercontent.com/gwastro/pycbc-logo/master/pycbc_logo_name.png)
 

{PyCBC-2.9.0 → pycbc-2.10.0}/PyCBC.egg-info/SOURCES.txt

@@ -104,6 +104,7 @@ bin/hwinj/pycbc_generate_hwinj_from_xml
 bin/hwinj/pycbc_insert_frame_hwinj
 bin/hwinj/pycbc_plot_hwinj
 bin/inference/pycbc_inference
+bin/inference/pycbc_inference_create_calibration_config
 bin/inference/pycbc_inference_create_fits
 bin/inference/pycbc_inference_extract_samples
 bin/inference/pycbc_inference_model_stats
@@ -125,6 +126,7 @@ bin/inference/pycbc_inference_plot_samples
 bin/inference/pycbc_inference_plot_skymap
 bin/inference/pycbc_inference_plot_thermodynamic_integrand
 bin/inference/pycbc_inference_pp_table_summary
+bin/inference/pycbc_inference_savage_dickey
 bin/inference/pycbc_inference_start_from_samples
 bin/inference/pycbc_inference_table_summary
 bin/inference/pycbc_validate_test_posterior
@@ -271,6 +273,8 @@ examples/inference/analytic-normal2d/run.sh
 examples/inference/bbh-injection/make_injection.sh
 examples/inference/bbh-injection/run.sh
 examples/inference/bbh-injection/run_test.sh
+examples/inference/calibration_configuration_files/download_calibration_uncertainty_files_o1_o2_o3.sh
+examples/inference/calibration_configuration_files/generate_calibration_config.sh
 examples/inference/gw150914/plot.sh
 examples/inference/gw150914/run.sh
 examples/inference/gw150914/run_test.sh

{PyCBC-2.9.0 → pycbc-2.10.0}/PyCBC.egg-info/requires.txt

@@ -14,7 +14,7 @@ tqdm
 setuptools
 gwdatafind
 pegasus-wms.api>=5.1.1
-igwn-ligolw<2.1.0
+igwn-ligolw
 igwn-segments
 lalsuite!=7.2
 lscsoft-glue>=1.59.3

{PyCBC-2.9.0 → pycbc-2.10.0}/bin/hwinj/pycbc_generate_hwinj

@@ -40,7 +40,7 @@ from pycbc.filter import sigmasq
 from pycbc.types import TimeSeries
 from pycbc.types.optparse import convert_to_process_params_dict
 from pycbc.io.ligolw import create_process_table
-from pycbc.waveform import get_td_waveform, td_approximants, taper_timeseries
+from pycbc.waveform import get_td_waveform, td_approximants
 
 def _empty_row(obj):
     """Create an empty sim_inspiral or sngl_inspiral row where the columns have
@@ -427,7 +427,7 @@ for ifo in opts.instruments:
 
     # taper waveform
     logging.info('Tapering strain for %s', ifo)
-    strain = taper_timeseries(strain, tapermethod=sim.taper)
+    strain = strain.taper_timeseries(location=sim.taper)
 
     # FFT strain
     logging.info('FFT strain for %s', ifo)

pycbc-2.10.0/bin/inference/pycbc_inference_create_calibration_config

@@ -0,0 +1,253 @@
+#!/usr/bin/env python
+
+# Copyright (C) 2023 Sumit K.
+#
+# This program is free software; you can redistribute it and/or modify it
+# under the terms of the GNU General Public License as published by the
+# Free Software Foundation; either version 3 of the License, or (at your
+# option) any later version.
+#
+# This program is distributed in the hope that it will be useful, but
+# WITHOUT ANY WARRANTY; without even the implied warranty of
+# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU General
+# Public License for more details.
+#
+# You should have received a copy of the GNU General Public License along
+# with this program; if not, write to the Free Software Foundation, Inc.,
+# 51 Franklin Street, Fifth Floor, Boston, MA  02110-1301, USA.
+
+"""Generate the calibration configuration file for a given event/GPS time.
+Also plots the calibration envelop for each detector for sanity checks.
+"""
+
+import os, argparse
+import pycbc
+import numpy
+import matplotlib.pyplot as plt
+
+from pycbc.types import MultiDetOptionAction
+from pycbc.strain.recalibrate import get_calibration_files_O1_O2_O3, read_calibration_envelop_file
+
+
+parser = argparse.ArgumentParser()
+
+# command line usage
+#parser = argparse.ArgumentParser(
+#            usage="pycbc_inference_create_calibration_config [--options]",
+#            description="Generate the calibration configuration files "
+#                        "to be used woth pycbc_inference.")
+
+pycbc.add_common_pycbc_options(parser)
+
+parser.add_argument('--calibration-files-path',
+                    help='Calibration envelope file path. It can be pre-'
+                          'downloaded or downloaded on the fly',
+                    type=str)
+
+
+parser.add_argument('--calibration-files', nargs='+', action=MultiDetOptionAction,
+                    metavar = 'DETECTOR:COLUMN', type=str,
+                    help='Calibration file for each detector',)
+
+parser.add_argument('--ifos', nargs='+',
+                        help='Specify ifos for the analysis')
+
+parser.add_argument('--minimum-frequency', nargs='+', action=MultiDetOptionAction,
+                        metavar='DETECTOR:COLUMN', type=float,
+                        help='For each detector, provide minimum frequency for'
+                        'the analysis. The minimum frequency should be higher than'
+                        'the minimum frequency in calibration envelop file so that'
+                        'we do not extrapolate.')
+
+
+parser.add_argument('--maximum-frequency', nargs='+', action=MultiDetOptionAction,
+                        metavar='DETECTOR:COLUMN', type=float,
+                        help='For each detector, provide maximum frequency for'
+                        'the analysis. The minimum frequency should be lower than'
+                        'the maximum frequency in calibration envelop file so that'
+                        'we do not extrapolate.')
+
+
+parser.add_argument('--gps-time',
+                    help='GPS time at which calibration configuration files'
+                          'are needed',
+                    required=True, type=float)
+
+parser.add_argument('--output-dir',                                 
+                    help="Path for the output calibration config files",                    
+                    default='.', type=str)
+
+parser.add_argument('--plots-dir',
+                    help="Path for the 'sanity check' plots",
+                    default='.', type=str)
+
+parser.add_argument('--n-nodes',
+                    help='Number of frequency nodes to be used',
+                    default=10, type=int)
+
+parser.add_argument('--correction-type', nargs='+', action=MultiDetOptionAction, 
+                    metavar='DETECTOR:COLUMN',
+                    help="Provide the correction type for each detector",
+                    required=True)
+
+parser.add_argument('--plot-sanity-checks', help='Make some plots for sanity checks',
+                    default=True, type=bool)
+
+parser.add_argument('--tag', help='Provide your tag for naming calibration file',
+                    default=None)
+
+
+# parse the command line
+opts = parser.parse_args()
+
+
+if opts.calibration_files_path and opts.calibration_files:
+    raise ValueError("Only one of the arguments should be provided among:"
+                    "--calibration-files-path and --calibration-files")
+
+
+calibration_files_path = opts.calibration_files_path
+min_freq_list = opts.minimum_frequency
+max_freq_list = opts.maximum_frequency
+correction_type_list = opts.correction_type
+gps_time = opts.gps_time
+output_dir = opts.output_dir
+n_nodes = opts.n_nodes
+if opts.tag is None:
+    tag = int(opts.gps_time)
+else:
+    tag = opts.tag
+
+for ifo in opts.ifos:
+  logging.info("Minimum %s frequency: %.3f", ifs, min_freq_list[ifo])
+
+# Read calibration files
+if opts.calibration_files:
+    calibration_files_dict = {}
+    for ifo in opts.ifos:
+        calibration_files_dict[ifo] = opts.calibration_files[ifo]
+else:
+    calibration_files_dict = get_calibration_files_O1_O2_O3(opts.ifos, gps_time, calibration_files_path)
+
+logging.debug("Calibration files dictionary: %s", calibration_files_dict)
+
+input_dict = {}
+for ii in range(len(opts.ifos)):
+    ifo = str(opts.ifos[ii])
+    input_dict['min_freq_%s'%ifo]=min_freq_list[ifo]
+    input_dict['max_freq_%s'%ifo]=max_freq_list[ifo]
+    input_dict['correction_type_%s'%ifo]=correction_type_list[ifo]
+    input_dict['calibration_envelop_file_%s'%ifo] = calibration_files_dict[ifo.upper()]
+logging.debug("Input dictionary: %s", input_dict)
+# Upper and Lower indices of detectors seems unnecessary
+# but they are there to make sure the correct convention
+# is followed for recalibration module in inference config
+# file as well as in detector notation.
+prior_dict = {}
+for ifo in opts.ifos:
+    #ifo = ifo.lower()
+    min_freq = input_dict['min_freq_%s'%ifo]
+    max_freq = input_dict['max_freq_%s'%ifo]
+    correction_type = input_dict['correction_type_%s'%ifo]
+    calib_env_file = input_dict['calibration_envelop_file_%s'%ifo]
+    log_nodes, amplitude_median_nodes, amplitude_sigma_nodes, phase_median_nodes, phase_sigma_nodes = \
+               read_calibration_envelop_file(calib_env_file, correction_type, 
+                                             min_freq, max_freq, n_nodes)
+    prior_dict['log_nodes_%s'%ifo] = log_nodes
+    prior_dict['amplitude_median_nodes_%s'%ifo] = amplitude_median_nodes
+    prior_dict['amplitude_sigma_nodes_%s'%ifo] = amplitude_sigma_nodes
+    prior_dict['phase_median_nodes_%s'%ifo] = phase_median_nodes
+    prior_dict['phase_sigma_nodes_%s'%ifo] =phase_sigma_nodes
+    logging.info('Done for detector %s', ifo)
+
+
+calibration_filename = '%s/calibration-%s.ini'%(output_dir, tag)
+if os.path.isfile(calibration_filename) != True:
+    text_file = open(calibration_filename, "w")
+    text_file.write("[calibration] \n")
+    for ifo in opts.ifos:
+        min_freq = input_dict['min_freq_%s'%ifo]
+        max_freq = input_dict['max_freq_%s'%ifo]
+        text_file.write("%s_model = cubic_spline \n"%ifo.lower())
+        text_file.write("%s_minimum_frequency = %d \n"%(ifo.lower(),min_freq))
+        text_file.write("%s_maximum_frequency = %d \n"%(ifo.lower(),max_freq))
+        text_file.write("%s_n_points = %d \n"%(ifo.lower(),n_nodes))
+    text_file.write(" \n")
+    text_file.write("[variable_params] \n")
+    for ifo in opts.ifos:
+        for ii in range(n_nodes):
+            text_file.write("recalib_amplitude_%s_%d = \n"%(ifo.lower(),ii))
+            text_file.write("recalib_phase_%s_%d = \n"%(ifo.lower(),ii))
+    text_file.write(" \n")
+    for ifo in opts.ifos:
+        for ii in range(n_nodes):
+            amplitude_median_nodes = prior_dict['amplitude_median_nodes_%s'%ifo][ii]
+            amplitude_sigma_nodes = prior_dict['amplitude_sigma_nodes_%s'%ifo][ii]
+            text_file.write("[prior-recalib_amplitude_%s_%d] \n"%(ifo.lower(),ii))
+            text_file.write("name = gaussian \n")
+            text_file.write("recalib_amplitude_%s_%d_mean = %.3g \n"%(ifo.lower(),ii,amplitude_median_nodes))
+            text_file.write("recalib_amplitude_%s_%d_var = %.3g \n"%(ifo.lower(),ii,amplitude_sigma_nodes**2))
+            text_file.write(" \n")
+    for ifo in opts.ifos:
+        for ii in range(n_nodes):
+            phase_median_nodes = prior_dict['phase_median_nodes_%s'%ifo][ii]
+            phase_sigma_nodes = prior_dict['phase_sigma_nodes_%s'%ifo][ii]
+            text_file.write("[prior-recalib_phase_%s_%d] \n"%(ifo.lower(),ii))
+            text_file.write("name = gaussian \n")
+            text_file.write("recalib_phase_%s_%d_mean = %.3g \n"%(ifo.lower(),ii,phase_median_nodes))
+            text_file.write("recalib_phase_%s_%d_var = %.3g \n"%(ifo.lower(),ii,phase_sigma_nodes**2))
+            text_file.write(" \n")
+    text_file.write(" \n")
+    text_file.write(" \n")
+    text_file.close()
+else:
+    raise ValueError(
+        'Calibration file already present. '
+        'Either delete it or rename it. Overwriting '
+        'this file is not allowed!!!'
+    )
+
+# Sanity check plots
+plots_dir = '%s/plots'%output_dir
+os.system('mkdir -p %s'%plots_dir)
+if opts.plot_sanity_checks:
+    for ifo in opts.ifos:
+        calib_env_path = input_dict['calibration_envelop_file_%s'%ifo]
+        min_freq = input_dict['min_freq_%s'%ifo]
+        max_freq = input_dict['max_freq_%s'%ifo]
+        log_nodes = prior_dict['log_nodes_%s'%ifo]
+        amp_median_nodes = prior_dict['amplitude_median_nodes_%s'%ifo]
+        amp_sigma_nodes = prior_dict['amplitude_sigma_nodes_%s'%ifo]
+        d1 = numpy.loadtxt(calib_env_path)
+        plt.figure(figsize=(8,5))
+        plt.plot(d1[:,0],d1[:,1],label=r'$\mu$')
+        plt.plot(d1[:,0],d1[:,3],label=r'$\mu-\sigma$')
+        plt.plot(d1[:,0],d1[:,5],label=r'$\mu+\sigma$')
+        plt.plot(numpy.exp(log_nodes),amp_median_nodes+1,'.',label=r'$\mu$(Prior)')
+        plt.plot(numpy.exp(log_nodes),amp_median_nodes+1-amp_sigma_nodes,'.',label=r'$\mu-\sigma$ (Prior)')
+        plt.plot(numpy.exp(log_nodes),amp_median_nodes+1+amp_sigma_nodes,'.',label=r'$\mu+\sigma$ (Prior)')
+        plt.xscale('log')
+        plt.xlabel('Frequency (Hz)')
+        plt.ylabel('Amplitude')
+        plt.title('Calibration envelop (Amplitude) for %s'%ifo)
+        plt.legend()
+        plt.savefig('%s/calibration_envelop_amplitude_%s_%s.png'%(plots_dir,tag,ifo))
+        plt.clf()
+
+        phase_median_nodes = prior_dict['phase_median_nodes_%s'%ifo]
+        phase_sigma_nodes = prior_dict['phase_sigma_nodes_%s'%ifo]
+        plt.figure(figsize=(8,5))
+        plt.plot(d1[:,0],d1[:,2],label=r'$\mu$')
+        plt.plot(d1[:,0],d1[:,4],label=r'$\mu-\sigma$')
+        plt.plot(d1[:,0],d1[:,6],label=r'$\mu+\sigma$')
+        plt.plot(numpy.exp(log_nodes),phase_median_nodes,'.',label=r'$\mu$ (Prior)')
+        plt.plot(numpy.exp(log_nodes),phase_median_nodes-phase_sigma_nodes,'.',label=r'$\mu-\sigma$ (Prior)')
+        plt.plot(numpy.exp(log_nodes),phase_median_nodes+phase_sigma_nodes,'.',label=r'$\mu+\sigma$')
+        plt.xscale('log')
+        plt.xlabel('Frequency (Hz)')
+        plt.ylabel('Phase')
+        plt.title('Calibration envelop (Phase) for %s'%ifo)
+        plt.legend()
+        plt.savefig('%s/calibration_envelop_phase_%s_%s.png'%(plots_dir,tag,ifo))
+        plt.clf()
+

{PyCBC-2.9.0 → pycbc-2.10.0}/bin/inference/pycbc_inference_extract_samples

@@ -43,6 +43,8 @@ import pycbc
 from pycbc.inference.io import (ResultsArgumentParser, results_from_cli,
                                 PosteriorFile, loadfile)
 from pycbc.inference.io.base_hdf import format_attr
+from scipy.special import logsumexp
+import warnings
 
 
 def isthesame(current_val, val):
@@ -83,6 +85,35 @@ parser.add_argument("--skip-groups", default=None, nargs="+",
                          "to write all groups if only one file is provided, "
                          "and all groups from the first file except "
                          "sampler_info if multiple files are provided.")
+parser.add_argument("--combine-via-sampling", action='store_true', 
+                    default=False,
+                    help="Specify whether to combine the posteriors by " 
+                         "sampling. By default, extract_samples will dump all "
+                         "samples from multiple inputs into one output file. "
+                         "If this option is specified, the samples will "
+                         "be randomly sampled with weighting based on the "
+                         "evidence in each input. For example, if the "
+                         "evidence of input 1 is twice that of input 2, "
+                         "the resulting posterior file with this option "
+                         "specified will have twice as many points from input "
+                         "1 than input 2. The output will have the same "
+                         "number of samples as the smallest input file. "
+                         "An error is thrown if this option is specified and "
+                         "any of the input files does not have a log_evidence "
+                         "attribute (e.g., the file used a sampler like emcee "
+                         "that does not report evidence). This option "
+                         "assumes that the priors of all files do not "
+                         "overlap; we cannot properly combine posteriors if "
+                         "the priors do overlap.")
+parser.add_argument("--mutually-exclusive-priors", action='store_true',
+                    default=False,
+                    help="If specifying --combine-via-sampling, specify "
+                         "whether to treat priors as mutually exclusive. By "
+                         "default, the provided input files are assumed to "
+                         "have identical priors, and their evidences will be "
+                         "averaged. If this option is specified, the priors "
+                         "are assumed to be non-overlapping across files, and "
+                         "their evidences will be summed together.")
 
 opts = parser.parse_args()
 
@@ -101,10 +132,52 @@ if len(opts.input_file) == 1:
 # convert samples to a dict in which the keys are the labels
 # also stack results if multiple files were provided
 if len(opts.input_file) > 1:
-    samples = {labels[p]: numpy.concatenate([s[p] for s in samples])
-               for p in params}
+    if opts.combine_via_sampling:
+        raw_samples = {labels[p]: numpy.concatenate([s[p] for s in samples])
+                   for p in params}
+        logz_list = []
+        dlogz_list = []
+        len_list = []
+        raw_samps_list = []
+        weights_list = []
+        for file in opts.input_file:
+            fp = loadfile(file, 'r')
+            # get evidence from each file if possible
+            try:
+                logz, dlogz = fp.log_evidence
+            except KeyError:
+                raise ValueError(f"Cannot combine evidences; file {file} "
+                                 "does not have a log_evidence attr")
+            logz_list.append(logz)
+            dlogz_list.append(dlogz)
+            # get samples from each file
+            file_samps = fp.read_samples(list(fp['samples'].keys()))
+            raw_samps_list.append(file_samps)
+            # get the number of samples from each file
+            len_list.append(len(file_samps))
+        # compute sampling weights from evidences
+        logz_net = logsumexp(logz_list)
+        len_net = sum(len_list)
+        out_size = min(len_list)
+        for i in range(len(opts.input_file)):
+            # weight each file's samples according to logz
+            logwt = logz_list[i] - logz_net
+            weights_list.append([numpy.exp(logwt)/len_list[i] for j in 
+                                 range(len_list[i])])
+        # randomly sample indices from all samples
+        weights = numpy.concatenate(weights_list)
+        idx = numpy.random.choice(int(len_net), size=out_size, replace=True, 
+                                  p=weights)
+        samples = {param: raw_samples[param][idx] for param in 
+                   raw_samples.keys()}
+    else:
+        samples = {labels[p]: numpy.concatenate([s[p] for s in samples])
+                   for p in params}
 else:
     samples = {labels[p]: samples[p] for p in params}
+    if opts.combine_via_sampling:
+        warnings.warn("Specified combine_via_sampling with only one input "
+                      "file. This option will have no effect.")
 
 # create the file
 outtype = PosteriorFile.name
@@ -137,10 +210,13 @@ for fp in fps:
 skip_attrs = ['filetype', 'thin_start', 'thin_interval', 'thin_end',
               'thinned_by', 'cmd', 'resume_points', 'effective_nsamples',
               'run_start_time', 'run_end_time']
-# also skip evidence if multiple files are being combined, since that will
-# not be the same
+# also skip evidence if multiple files are being combined; this will be handled
+# via sampling if specified
 if len(opts.input_file) > 1:
     skip_attrs += ['log_evidence', 'dlog_evidence']
+    
+# make sure attrs are the same between files...
+cat_params = False
 for fp in fps:
     for key in map(format_attr, fp.attrs):
         if key not in skip_attrs:
@@ -151,12 +227,36 @@ for fp in fps:
                 out.attrs[key] = val
                 current_val = format_attr(out.attrs[key])
             if not isthesame(current_val, val):
-                raise ValueError("cannot combine all files; file attr {} is "
-                                 "not the same across all files ({} vs {})"
-                                 .format(key, current_val, val))
+                if key == 'remapped_params':
+                    # ...unless it's remapped_params; just save first file's
+                    # entries if they don't match between files
+                    warnings.warn("WARNING: remapped_params metadata does not "
+                                 "match between files; saving metadata from "
+                                 "first file")
+                else:
+                    raise ValueError("cannot combine all files; file attr {} is "
+                                     "not the same across all files ({} vs {})"
+                                     .format(key, current_val, val))
 
-# store what parameters were renamed
-out.attrs['remapped_params'] = list(labels.items())
+# store what parameters were renamed (if not already saved above)
+if not cat_params:
+    out.attrs['remapped_params'] = list(labels.items())
+
+# write combined evidence and dlog evidence if combining via sampling
+if opts.combine_via_sampling:
+    if len(opts.input_file) == 1:
+        # there's only one file; just return what's in the input
+        out.attrs['log_evidence'] = fps[0].log_evidence[0]
+        out.attrs['dlog_evidence'] = fps[0].log_evidence[1]
+    elif opts.mutually_exclusive_priors:
+        # add together the evidences; quadrature sum the residuals
+        out.attrs['log_evidence'] = logsumexp(logz_list)
+        out.attrs['dlog_evidence'] = numpy.sqrt(sum([i**2 for i in dlogz_list]))
+    else:
+        # average the evidences; quadrature sum and scale the residuals
+        n = len(opts.input_file)
+        out.attrs['log_evidence'] = logsumexp(logz_list) - numpy.log(n)
+        out.attrs['dlog_evidence'] = numpy.sqrt(sum([i**2 for i in dlogz_list])) / n
 
 # write the other groups using the first file
 fp = fps[0]

{PyCBC-2.9.0 → pycbc-2.10.0}/bin/inference/pycbc_inference_plot_posterior

@@ -79,6 +79,11 @@ parser.add_argument("--plot-maxl", action="store_true", default=False,
 parser.add_argument('--legend-location', default='upper right',
                     help='Where to put the legend (if multiple files are '
                          'provided). Default is "upper right".')
+parser.add_argument('--legend-bbox-to-anchor', default='(1.0, 1.0)',
+                    help='Box that is used to position the legend in conjunction with loc.')
+parser.add_argument('--legend-fontsize', default='medium',
+                    help="The font size of legend. Default is 'medium'."
+                         "Choose from int or {'xx-small', 'x-small', 'small', 'medium', 'large', 'x-large', 'xx-large'}.")
 parser.add_argument('--dpi', type=int, default=200,
                     help="Set the DPI of the plot. Default is 200.")
 # style option
@@ -113,6 +118,15 @@ if opts.plot_maxl:
 else:
     add_logl = False
 
+# parse legend bbox
+def parse_tuple(value):
+    try:
+        return tuple(map(float, value.strip("()").split(",")))
+    except Exception as e:
+        raise argparse.ArgumentTypeError(f"Invalid tuple format: {value}") from e
+
+opts.legend_bbox_to_anchor = parse_tuple(opts.legend_bbox_to_anchor)
+
 # set logging
 pycbc.init_logging(opts.verbose)
 
@@ -358,7 +372,8 @@ if len(opts.input_file) > 1:
     else:
         addto = fig
     addto.legend(loc=opts.legend_location, handles=handles,
-                 labels=labels)
+                 labels=labels, fontsize=opts.legend_fontsize,
+                 bbox_to_anchor=opts.legend_bbox_to_anchor)
 
 # set DPI
 fig.set_dpi(opts.dpi)