gwcelery.tasks.em_bright module¶
This module computes the probabilities that there is a neutron star in the binary, and that the coalescence event resulted in creation of tidally disrupted matter.
The result is returned in the form of a JSON file:
‘{“HasNS”: 1.0, “HasRemnant”: 1.0}’
HasNS
: The probability that at least one of the component masses- in the binary is a neutron star. The definition of a neutron star in this context simply means an object with mass less than 3.0 solar mass.
HasRemnant
: The probability that the binary system can produce- tidally disrupted matter during coalescence. This is
computed using the fitting formula in arXiv:1807.00011
We are currently using an extremely stiff equation of
state (2H) to compute the compactness of the neutron
star. This results in a higher chance of labelling
a systems with non-zero
HasRemnant
value.
Qualitative source classification for CBC events.
-
(task)
gwcelery.tasks.em_bright.
em_bright_posterior_samples
(posterior_file_content)[source]¶ Returns the probability of having a NS component and remnant using LALInference posterior samples.
Parameters: posterior_file_content (hdf5 posterior file content) – Returns: JSON formatted string storing HasNS
andHasRemnant
probabilitiesReturn type: str Example
>>> em_bright_posterior_samples(GraceDb().files('S190930s', ... 'LALInference.posterior_samples.hdf5').read()) {"HasNS": 0.014904901243599122, "HasRemnant": 0.0}
-
(task)
gwcelery.tasks.em_bright.
classifier_other
(args, graceid)[source]¶ Returns the boolean probability of having a NS component and the probability of having non-zero disk mass. This method is used for pipelines that do not provide the data products necessary for computation of the source properties probabilities.
Parameters: Returns: JSON formatted string storing
HasNS
andHasRemnant
probabilitiesReturn type: Example
>>> em_bright.classifier_other((2.0, 1.0, 0.0, 0.0, 10.), 'S123456') '{"HasNS": 1.0, "HasRemnant": 1.0}'
-
(task)
gwcelery.tasks.em_bright.
classifier_gstlal
(args, graceid)[source]¶ Returns the probability of having a NS component and the probability of having non-zero disk mass in the detected event. This method will be using the data products obtained from the weekly supervised learning runs for injections campaigns. The data products are in pickle formatted RandomForestClassifier objects. The method predict_proba of these objects provides us the probabilities of the coalesence being EM-Bright and existence of neutron star in the binary.
Parameters: Returns: JSON formatted string storing
HasNS
andHasRemnant
probabilitiesReturn type: Notes
This task would only work from within the CIT cluster.