bilby issueshttps://git.ligo.org/lscsoft/bilby/-/issues2019-08-21T11:06:12Zhttps://git.ligo.org/lscsoft/bilby/-/issues/403Improve validate filename2019-08-21T11:06:12ZGregory Ashtongregory.ashton@ligo.orgImprove validate filenameWe have used a rather inelegant way to check if something is a file [when reading in a PSD](https://git.ligo.org/lscsoft/bilby/blob/master/bilby/gw/detector/psd.py#L291):
```python
if file is not None and '/' not in file:
...We have used a rather inelegant way to check if something is a file [when reading in a PSD](https://git.ligo.org/lscsoft/bilby/blob/master/bilby/gw/detector/psd.py#L291):
```python
if file is not None and '/' not in file:
file = os.path.join(os.path.dirname(__file__), 'noise_curves', file)
return file
```
This should be replaced by something like
```python
if file is not None and os.path.isfile(file):
file = os.path.join(os.path.dirname(__file__), 'noise_curves', file)
return file
```
This needs to be implemented and checked that it works. We should also look for any other instances which do something similar.0.5.5https://git.ligo.org/lscsoft/bilby/-/issues/398Emcee chains error2019-08-21T11:05:36ZGitLab Support BotEmcee chains errorDear developers:
I am a graduate student at the Institute of Theoretical Physics, Chinese Academy of Sciences. I use bilby for Bayesian analysis. The sampler emcee can't give the correct contour and I found a bug in bilby.
In the file ...Dear developers:
I am a graduate student at the Institute of Theoretical Physics, Chinese Academy of Sciences. I use bilby for Bayesian analysis. The sampler emcee can't give the correct contour and I found a bug in bilby.
In the file core/sampler/emcee.py, line 359 and line 371
```
chain = self.sampler.chain.reshape((-1, self.ndim))
self.result.samples = chain[n_samples:, :]
```
The chain is reshaped before dropping burn-in steps(slice after
n_samples). But this will throw away n_samples in the first chain
instead of the first nburns in each chains. So the samples stored in the
result is wrong. On the other hand, ptemcee.py do the slice before
reshape. I modified emcee.py according to ptemcee.py and the result is
correct.
I hope that you can correct this bug, and thank you for your contribution.
best wishes
Xia Chen
EDIT: I reformatted to use the code-formatting (Greg Ashton)https://git.ligo.org/lscsoft/bilby/-/issues/397Nightly scheduled tests failing2019-07-29T23:21:51ZGregory Ashtongregory.ashton@ligo.orgNightly scheduled tests failingThe nightly scheduled tests are failing due to the "sample from the prior" test. It still passes when I run it locally, but when run on the C.I. (https://git.ligo.org/lscsoft/bilby/-/jobs/405008) all the P-values come out to 1 and it fai...The nightly scheduled tests are failing due to the "sample from the prior" test. It still passes when I run it locally, but when run on the C.I. (https://git.ligo.org/lscsoft/bilby/-/jobs/405008) all the P-values come out to 1 and it fails. I can't figure out why.0.5.4https://git.ligo.org/lscsoft/bilby/-/issues/396Issue when sampling using BasicGravitationalWaveTransient2019-08-20T00:56:27ZJonathan DaviesIssue when sampling using BasicGravitationalWaveTransientI've been trying to modify the bilby code so that uncertainty in the form of the PSD can be folded into the analysis i.e. multiple PSD files are loaded in and the indices of these are implemented as parameters to be fitted. The error sho...I've been trying to modify the bilby code so that uncertainty in the form of the PSD can be folded into the analysis i.e. multiple PSD files are loaded in and the indices of these are implemented as parameters to be fitted. The error shown below got thrown up in run_sampler, as I am using BasicGravitationalWaveTransient rather than GravitationalWaveTransient in bilby/bilby/gw/likelihood.py. It gets called like this:
result = bilby.run_sampler(
likelihood, prior, outdir=outdir, label=label,
sampler=sampler, npoints=npoints, use_ratio=False,
injection_parameters=None,
conversion_function=bilby.gw.conversion.generate_all_bbh_parameters, resume=False, n_check_point=5000)
It seems that the issue is that the generate_all_bbh_parameters function is expecting the likelihood to have the various marginalization flags as attributes, and it fails for the Basic version which doesn't have those options.
Traceback (most recent call last):
File "GW150914_advanced.py", line 130, in <module>
conversion_function=bilby.gw.conversion.generate_all_bbh_parameters, resume=True, n_check_point=5000)
File "/home/jonathan.davies/bilby/bilby/core/sampler/__init__.py", line 203, in run_sampler
conversion_function=conversion_function)
File "/home/jonathan.davies/bilby/bilby/core/result.py", line 1072, in samples_to_posterior
data_frame = conversion_function(data_frame, likelihood, priors)
File "/home/jonathan.davies/bilby/bilby/gw/conversion.py", line 703, in generate_all_bbh_parameters
likelihood=likelihood, priors=priors)
File "/home/jonathan.davies/bilby/bilby/gw/conversion.py", line 667, in _generate_all_cbc_parameters
samples=output_sample, likelihood=likelihood)
File "/home/jonathan.davies/bilby/bilby/gw/conversion.py", line 1015, in generate_posterior_samples_from_marginalized_likelihood
if not any([likelihood.phase_marginalization,
AttributeError: 'BasicGravitationalWaveTransient' object has no attribute 'phase_marginalization'FutureSylvia BiscoveanuSylvia Biscoveanuhttps://git.ligo.org/lscsoft/bilby/-/issues/382CDF plots with priors show the PDF2019-06-28T01:35:05ZGregory Ashtongregory.ashton@ligo.orgCDF plots with priors show the PDFThe 1D CDF plots, if generated with `priors=True` will overplot the pdf, e.g.
![image](/uploads/a921e8f0411a2a3e7cd1a7e3c6ba35e6/image.png)The 1D CDF plots, if generated with `priors=True` will overplot the pdf, e.g.
![image](/uploads/a921e8f0411a2a3e7cd1a7e3c6ba35e6/image.png)Gregory Ashtongregory.ashton@ligo.orgGregory Ashtongregory.ashton@ligo.orghttps://git.ligo.org/lscsoft/bilby/-/issues/372Bias in the ROQ geocent_time posterior2019-05-28T00:45:42ZGregory Ashtongregory.ashton@ligo.orgBias in the ROQ geocent_time posteriorRunning the fiducial BBH test for a 128s ROQ basis, there seems to be a bias in the geocent_time:
![image](/uploads/2bd91474291e4e436d997b8043218d02/image.png)
Full corner plot:
![image](/uploads/5b367b63e89276e2bb299851d8e7d86e/image...Running the fiducial BBH test for a 128s ROQ basis, there seems to be a bias in the geocent_time:
![image](/uploads/2bd91474291e4e436d997b8043218d02/image.png)
Full corner plot:
![image](/uploads/5b367b63e89276e2bb299851d8e7d86e/image.png)
Further more, a pp plot also suggests a bias
![image](/uploads/a5c0860e026508d956ab5f6e720efc68/image.png)
where the individual p-values are given by
```
09:37 bilby INFO : Key: KS-test p-value
09:37 bilby INFO : a_1: 0.004353009481190579
09:37 bilby INFO : a_2: 0.03985168897683587
09:37 bilby INFO : chirp_mass: 0.6807714838241979
09:37 bilby INFO : dec: 0.10544070562244892
09:37 bilby INFO : geocent_time: 0.0032471782742264045
09:37 bilby INFO : luminosity_distance: 0.8240397996477468
09:37 bilby INFO : mass_2: 0.048316085706487676
09:37 bilby INFO : mass_ratio: 0.05841381950543859
09:37 bilby INFO : phase: 0.8647187964800038
09:37 bilby INFO : phi_12: 0.13550324406264547
09:37 bilby INFO : phi_jl: 0.059429838161255764
09:37 bilby INFO : psi: 0.5781828834354825
09:37 bilby INFO : ra: 0.13952060698874352
09:37 bilby INFO : theta_jn: 0.002548388019280191
09:37 bilby INFO : tilt_1: 0.5544800028886142
09:37 bilby INFO : tilt_2: 0.16335729310487174
09:37 bilby INFO : Combined p-value: 1.0659765250158535e-05
```
Notably, the geocent time (along with spins) is small.0.5.1Gregory Ashtongregory.ashton@ligo.orgGregory Ashtongregory.ashton@ligo.orghttps://git.ligo.org/lscsoft/bilby/-/issues/370Distance marginalisation lookup doesn't store phase marginalisation2019-05-14T02:34:15ZColm Talbotcolm.talbot@ligo.orgDistance marginalisation lookup doesn't store phase marginalisationThe distance marginalisation lookup table depends on whether phase marginalisation was assumed when constructing it.
We should add phase marginalisation as an extra flag to the checks done when loading the lookup table.
If you load a p...The distance marginalisation lookup table depends on whether phase marginalisation was assumed when constructing it.
We should add phase marginalisation as an extra flag to the checks done when loading the lookup table.
If you load a phase marignalised table without using phase marginalisation (or vice versa) things go very wrong.https://git.ligo.org/lscsoft/bilby/-/issues/364Errors with sampling frequency for open data2019-05-09T00:32:15ZShanika GalaudageErrors with sampling frequency for open dataThere is a problem when setting values of sampling frequency that are not equal to a power of 2.
Fails with the following,
```
~/anaconda3/lib/python3.6/site-packages/bilby-0.4.4-py3.6.egg/bilby/gw/source.py in _base_lal_cbc_fd_wavefo...There is a problem when setting values of sampling frequency that are not equal to a power of 2.
Fails with the following,
```
~/anaconda3/lib/python3.6/site-packages/bilby-0.4.4-py3.6.egg/bilby/gw/source.py in _base_lal_cbc_fd_waveform(frequency_array, mass_1, mass_2, luminosity_distance, theta_jn, phase, a_1, a_2, tilt_1, tilt_2, phi_12, phi_jl, lambda_1, lambda_2, eccentricity, **waveform_kwargs)
263 h_cross = np.zeros_like(frequency_array, dtype=np.complex)
264
--> 265 h_plus[:len(hplus.data.data)] = hplus.data.data
266 h_cross[:len(hcross.data.data)] = hcross.data.data
267
ValueError: could not broadcast input array from shape (16385) into shape (8197)
```
Until a permanent fix, setting sampling frequency equal to a power of 2 should avoid this problem.
@colm.talbot @gregory.ashtonhttps://git.ligo.org/lscsoft/bilby/-/issues/362Fix ROQ time resolution for quiet signals.2019-04-29T07:13:38ZColm Talbotcolm.talbot@ligo.orgFix ROQ time resolution for quiet signals.It looks like there's a bias in the time posteriors for ROQ runs with quiet injections. This is coming from https://git.ligo.org/lscsoft/bilby/blob/master/bilby/gw/likelihood.py#L986.
We should set a minimum for the network SNR used to ...It looks like there's a bias in the time posteriors for ROQ runs with quiet injections. This is coming from https://git.ligo.org/lscsoft/bilby/blob/master/bilby/gw/likelihood.py#L986.
We should set a minimum for the network SNR used to calculate the time resolution.
Also, SNRs add in quadrature.https://git.ligo.org/lscsoft/bilby/-/issues/359Calibration names/prefix mismatch2019-05-27T04:02:37ZGregory Ashtongregory.ashton@ligo.orgCalibration names/prefix mismatchFor reading in calibration files, there is some confusion. The calibration model `CubicSpline` takes `prefix`, but the `bilby.gw.prior.CalibrationPriorDict` takes `label`. To get this to work, I think you need `"prefix=recalib_{}_".forma...For reading in calibration files, there is some confusion. The calibration model `CubicSpline` takes `prefix`, but the `bilby.gw.prior.CalibrationPriorDict` takes `label`. To get this to work, I think you need `"prefix=recalib_{}_".format(ifo.name)` and `label=ifo.name`. The `recalib` is hard-coded.
I think we should ideally just have them both take the same argument since they mean the same thing.Gregory Ashtongregory.ashton@ligo.orgGregory Ashtongregory.ashton@ligo.orghttps://git.ligo.org/lscsoft/bilby/-/issues/358Distance posterior reconstruction broken2019-04-16T03:35:54ZColm Talbotcolm.talbot@ligo.orgDistance posterior reconstruction brokenThe reconstruction of the distance posterior after using the distance marginalised likelihood appears to be broken.
The marginalised likelihood _is_ working. It's just the post-processing that is broken.The reconstruction of the distance posterior after using the distance marginalised likelihood appears to be broken.
The marginalised likelihood _is_ working. It's just the post-processing that is broken.0.4.5Colm Talbotcolm.talbot@ligo.orgColm Talbotcolm.talbot@ligo.orghttps://git.ligo.org/lscsoft/bilby/-/issues/357`gwpy` breaking legend plots2019-05-15T03:54:04ZGregory Ashtongregory.ashton@ligo.org`gwpy` breaking legend plotsThis script
```python
import matplotlib.pyplot as plt
import bilby
plt.plot([0, 0], [1, 1], label='test')
plt.legend()
plt.savefig('test.png')
```
Generates this image
![test](/uploads/f175def69c6ee30f627410b9d1c993ac/test.png)
*if*...This script
```python
import matplotlib.pyplot as plt
import bilby
plt.plot([0, 0], [1, 1], label='test')
plt.legend()
plt.savefig('test.png')
```
Generates this image
![test](/uploads/f175def69c6ee30f627410b9d1c993ac/test.png)
*if* `gwpy` is installed, whereas if `gwpy` is not installed, it will generate this.
![test](/uploads/3205905ea04e140b248b772f60ee9412/test.png)
Note that the legend label is a box when `gwpy` is installed. For dashed lines etc this looks very confused. I can't figure out exactly, but I think `gwpy` is setting some rcparam. @duncanmmacleod do you understand this at all?https://git.ligo.org/lscsoft/bilby/-/issues/353CoupledTimeFrequencySeries bug?2019-04-05T22:52:11ZColm Talbotcolm.talbot@ligo.orgCoupledTimeFrequencySeries bug?I think there's a bug in the `CoupledTimeAndFrequencySeries` that are hurting performance.
- `_frequency_array_updated` is never set to `True` after an update, I think this should be done around L47
The result of these is that we spend...I think there's a bug in the `CoupledTimeAndFrequencySeries` that are hurting performance.
- `_frequency_array_updated` is never set to `True` after an update, I think this should be done around L47
The result of these is that we spend a _lot_ of time creating the frequency array.
I may be misunderstanding the code, @moritz.huebner?https://git.ligo.org/lscsoft/bilby/-/issues/351`samples_to_posterior` fails if the prior is not given2019-04-10T03:12:40ZGregory Ashtongregory.ashton@ligo.org`samples_to_posterior` fails if the prior is not givenCurrently, the `prior` default to None. Then in the code it does `for _ in prior` which causes an exception of `NoneType object is not iterable. This just needs a quick bug fix.Currently, the `prior` default to None. Then in the code it does `for _ in prior` which causes an exception of `NoneType object is not iterable. This just needs a quick bug fix.https://git.ligo.org/lscsoft/bilby/-/issues/348Time samples should be saved with ROQ weights2019-04-17T09:57:02ZColm Talbotcolm.talbot@ligo.orgTime samples should be saved with ROQ weightsCurrently, there's an option to save the ROQ weights to disk so they can be loaded later. This currently doesn't store the time samples.
This means that the time samples are recomputed when the weights are loaded, but if the IFO network...Currently, there's an option to save the ROQ weights to disk so they can be loaded later. This currently doesn't store the time samples.
This means that the time samples are recomputed when the weights are loaded, but if the IFO network has changed (e.g. for coherence tests) these will not match leading to bias in the time posterior and other parameters as a consequence.
We should save the `time_samples` in the same dictionary as the weights.https://git.ligo.org/lscsoft/bilby/-/issues/345Resume file removed before result file output2019-03-28T01:16:56ZTsun-Ho PangResume file removed before result file outputI was having a parameter estimation run on the cluster (CIT) as a condor job. Because the generation of posterior samples with the conversion function takes too long, my job got idle. As the resume file got removed, the parameters estima...I was having a parameter estimation run on the cluster (CIT) as a condor job. Because the generation of posterior samples with the conversion function takes too long, my job got idle. As the resume file got removed, the parameters estimation restarted from the beginning with all the progress lost. May I ask can we have the resume removed only after the posterior samples saved in result.h5?https://git.ligo.org/lscsoft/bilby/-/issues/344Reconstructed time looks odd2019-03-26T22:14:41ZGregory Ashtongregory.ashton@ligo.orgReconstructed time looks oddI did a run on GW150914 and recovered the following posterior:
![image](/uploads/88b8e3cf58c03aeaf344100369f12a37/image.png)
I'm guessing this is just due to the choice of binning or so, any thoughts @colm.talbot @eric.thrane ?I did a run on GW150914 and recovered the following posterior:
![image](/uploads/88b8e3cf58c03aeaf344100369f12a37/image.png)
I'm guessing this is just due to the choice of binning or so, any thoughts @colm.talbot @eric.thrane ?https://git.ligo.org/lscsoft/bilby/-/issues/340Issue with open data for GW1512262019-08-28T22:44:23ZVivien RaymondIssue with open data for GW151226There seem to be an issue with the PSD for H1 when running the following:
```python
from __future__ import division, print_function
import bilby
outdir = 'outdir'
label = 'GW151226'
time_of_event = bilby.gw.utils.get_event_time(label)
i...There seem to be an issue with the PSD for H1 when running the following:
```python
from __future__ import division, print_function
import bilby
outdir = 'outdir'
label = 'GW151226'
time_of_event = bilby.gw.utils.get_event_time(label)
interferometers = bilby.gw.detector.get_event_data(label)
```
![download](/uploads/128331e72bd6513e43091dbb120c404e/download.png)
The H1 ASD is missing, and is `NaN`:
```python
interferometers[0].power_spectral_density
```
```
PowerSpectralDensity(frequency_array=[0.00000e+00 2.50000e-01 5.00000e-01 ... 2.04750e+03 2.04775e+03
2.04800e+03], psd_array=[nan nan nan ... nan nan nan], asd_array=[nan nan nan ... nan nan nan])
```https://git.ligo.org/lscsoft/bilby/-/issues/339Issue in parsing component spins for aligned spin BNS run2019-03-11T19:55:23ZArunava MukherjeeIssue in parsing component spins for aligned spin BNS run# Issue in parsing component spins for BNS run to the WF call in LalSimulation
I tried to run aligned-spin BNS system with the [default example](https://git.ligo.org/lscsoft/bilby/blob/master/examples/injection_examples/binary_neutron_...# Issue in parsing component spins for BNS run to the WF call in LalSimulation
I tried to run aligned-spin BNS system with the [default example](https://git.ligo.org/lscsoft/bilby/blob/master/examples/injection_examples/binary_neutron_star_example.py), but it seemed to fail with the following error:
```
In [1]: %run binary_neutron_star_example.py
14:35 bilby INFO : Running bilby version: 0.4.1: (CLEAN) 8af0cce 2019-03-04 23:22:41 -0600
/Users/arunava/anaconda3/lib/python3.7/site-packages/bilby-0.4.1-py3.7.egg/bilby/gw/detector.py:1997: RuntimeWarning: invalid value encountered in multiply
frequency_domain_strain = self.__power_spectral_density_interpolated(frequencies) ** 0.5 * white_noise
XLAL Error - XLALSimInspiralChooseFDWaveform: Non-zero transverse spins were given, but this is a non-precessing approximant.
XLAL Error - XLALSimInspiralChooseFDWaveform (LALSimInspiral.c:1057): Invalid argument
---------------------------------------------------------------------------
RuntimeError Traceback (most recent call last)
```
```
/anaconda3/lib/python3.7/site-packages/bilby-0.4.1-py3.7.egg/bilby/gw/utils.py in lalsim_SimInspiralChooseFDWaveform(mass_1, mass_2, spin_1x, spin_1y, spin_1z, spin_2x, spin_2y, spin_2z, luminosity_distance, iota, phase, longitude_ascending_nodes, eccentricity, mean_per_ano, delta_frequency, minimum_frequency, maximum_frequency, reference_frequency, waveform_dictionary, approximant)
797 longitude_ascending_nodes, eccentricity, mean_per_ano, delta_frequency,
798 minimum_frequency, maximum_frequency, reference_frequency,
--> 799 waveform_dictionary, approximant)
800
801
RuntimeError: Invalid argument
In [2]:
```
* It seemed to be an error in parsing the spin components return values to the `lalsim.SimInspiralChooseFDWaveform` call in line no 794 of `bilby/gw/utils.py` for this version (bilby version: 0.4.1: (CLEAN) 8af0cce 2019-03-04 23:22:41 -0600): [here](https://git.ligo.org/lscsoft/bilby/blob/master/bilby/gw/utils.py#L794)
## Issue in computing the numbers in spin components
* While I am running it for the values fixed prior in `chi_1` = `chi_2` = 0.02
```
In [2]: list(priors.items())
Out[2]:
[('chi_1', DeltaFunction(peak=0.02, name=None, latex_label=None, unit=None)),
('chi_2', DeltaFunction(peak=0.02, name=None, latex_label=None, unit=None)),
('luminosity_distance',
DeltaFunction(peak=50.0, name=None, latex_label=None, unit=None)),
('dec', DeltaFunction(peak=-1.2108, name=None, latex_label=None, unit=None)),
('ra', DeltaFunction(peak=1.375, name=None, latex_label=None, unit=None)),
('theta_jn', DeltaFunction(peak=0.4, name=None, latex_label=None, unit=None)),
('psi', DeltaFunction(peak=2.659, name=None, latex_label=None, unit=None)),
('phase', DeltaFunction(peak=1.3, name=None, latex_label=None, unit=None)),
('geocent_time',
DeltaFunction(peak=1126259642.413, name=None, latex_label=None, unit=None)),
('chirp_mass',
Gaussian(mu=1.215, sigma=0.1, name='chirp_mass', latex_label='$\\mathcal{M}$', unit='$M_{\\odot}$')),
('symmetric_mass_ratio',
Uniform(minimum=0.1, maximum=0.25, name='symmetric_mass_ratio', latex_label='$\\eta$', unit=None)),
('lambda_tilde',
Uniform(minimum=0, maximum=5000, name='lambda_tilde', latex_label='$\\tilde{\\Lambda}$', unit=None)),
('delta_lambda',
Uniform(minimum=-5000, maximum=5000, name='delta_lambda', latex_label='$\\delta\\Lambda$', unit=None))]
```
I still see that spin components are passed to the `lalsim.SimInspiralChooseFDWaveform` call as:
```
spin_1x=-0.00493474, spin_1y=0.0193775, spin_1z=0.0004, spin_2x=-0.00493474, spin_2y=0.0193775, spin_2z=0.0004
```
* It seems to me to be incorrect.
* It also explains why the aligned spin waveform like `TaylorF2` is failing. (see #334)https://git.ligo.org/lscsoft/bilby/-/issues/337Nightly tests are failing2019-03-26T21:42:55ZColm Talbotcolm.talbot@ligo.orgNightly tests are failingThe nightly tests are failing because three examples are run in "test mode", I'm not sure I understand what this does exactly, but it seems to lead to the posterior containing samples with `log_likelihood=-inf`.
Specifically, we get sam...The nightly tests are failing because three examples are run in "test mode", I'm not sure I understand what this does exactly, but it seems to lead to the posterior containing samples with `log_likelihood=-inf`.
Specifically, we get samples with m1 < m2 which means that when we try to reconstruct the marginalised parameters we get waveforms which are `None` and it fails with the following trace.
My question is, how useful is this testing mode? Can't we just construct a test which runs p.e. with very few live points which will still finish in a few minutes and give meaningful output.
@gregory.ashton @moritz.huebner
```
self = <test.gw_example_test.Test testMethod=test_examples>
def test_examples(self):
""" Loop over examples to check they run """
examples = ['examples/injection_examples/fast_tutorial.py',
'examples/injection_examples/marginalized_likelihood.py',
'examples/open_data_examples/GW150914.py',
]
for filename in examples:
print("Testing {}".format(filename))
> execfile(filename)
test/gw_example_test.py:52:
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
/opt/conda/lib/python3.7/site-packages/past/builtins/misc.py:82: in execfile
exec_(code, myglobals, mylocals)
examples/injection_examples/marginalized_likelihood.py:72: in <module>
conversion_function=bilby.gw.conversion.generate_all_bbh_parameters)
bilby/core/sampler/__init__.py:186: in run_sampler
conversion_function=conversion_function)
bilby/core/result.py:981: in samples_to_posterior
data_frame = conversion_function(data_frame, likelihood, priors)
bilby/gw/conversion.py:697: in generate_all_bbh_parameters
likelihood=likelihood, priors=priors)
bilby/gw/conversion.py:661: in _generate_all_cbc_parameters
samples=output_sample, likelihood=likelihood)
bilby/gw/conversion.py:1027: in generate_posterior_samples_from_marginalized_likelihood
signal_polarizations=signal_polarizations)
bilby/gw/conversion.py:1080: in generate_time_sample_from_marginalized_likelihood
signal = ifo.get_detector_response(signal_polarizations, sample)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
self = Interferometer(name='H1', power_spectral_density=PowerSpectralDensity(psd_file='/builds/lscsoft/bilby/bilby/gw/noise_c...76571388889, elevation=142.554, xarm_azimuth=125.9994, yarm_azimuth=215.9994, xarm_tilt=-0.0006195, yarm_tilt=1.25e-05)
waveform_polarizations = None
parameters = {'a_1': 0.4, 'a_2': 0.3, 'dec': -1.2108, 'geocent_time': 1126259639.413, ...}
def get_detector_response(self, waveform_polarizations, parameters):
""" Get the detector response for a particular waveform
Parameters
-------
waveform_polarizations: dict
polarizations of the waveform
parameters: dict
parameters describing position and time of arrival of the signal
Returns
-------
array_like: A 3x3 array representation of the detector response (signal observed in the interferometer)
"""
signal = {}
> for mode in waveform_polarizations.keys():
E AttributeError: 'NoneType' object has no attribute 'keys'
bilby/gw/detector.py:1271: AttributeError
```