Customising data analysis

It is possible to specify alternative likelihoods, default priors, source models, and suchlike through the ini file. Each of these are done by passing a full python path, for example, to use a WaveformGenerator class and source model from some external package the following should be included in the ini file.

waveform-generator-class = my_package.submodule.CustomWaveformGenerator
frequency-domain-source-model = my_package.submodule.custom_source_model

In order to be compatible with the bilby_pipe analysis scripts, custom classes should take the same arguments as their parent classes. If you do not want to hard code all such arguments, you can use **kwargs to capture (and ignore) additional arguments. An exception to this is when passing custom likelihood classes. In this case additional keyword arguments can be passed through the ini file, as below

likelihood-type = my_package.submodule.CustomLikelihood
extra-likelihood-kwargs = {new_argument: value, other_argument: other_value}

Specifically, the following options allow one to pass a python path:

analysis-executable-parser
conversion-function
default-prior
frequency-domain-source-model
generation-function
likelihood-type
waveform-generator-class