Possible bug in delta_lambda prior key in binary_neutron_star_example.py
@philippe.landry and I have been trying to run the binary neutron star example and suspect there is a typo with the "delta_lambda" prior key.
When running the example as is, we are finding recovered lambda_1 and lambda_2 values which are not consistent with the injected values. This inconsistency vanishes if the two instances of "delta_lambda" in the prior keys are changed to "delta_lambda_tilde" in line 105.
In the plots below, the first two are the example run without changes, and the second two are the same with the aforementioned change in the prior key.
In both cases, the "lambda_tilde" and "delta_lambda_tilde" recovery is as expected, but when using "delta_lambda" as the prior key, the mapping from the sampled parameters to "lambda_1" and "lambda_2" seems to be incorrect. We notice that "delta_lambda" does not appear in the prior dictionaries in priors.py or conversion.py, whereas "delta_lambda_tilde" does. We are unsure whether this is by design, but we are proposing to edit the example to use the "delta_lambda_tilde" prior key instead of "delta_lambda", and wanted to bring the issue of conversion.py compatibility with the "delta_lambda" key to your attention.
If desired, we can prepare a merge request on our fork of bilby to fix binary_neutron_star_example.py.