Enable using multiple and multi-banded ROQ bases

This patch enables using multiple ROQ bases in a single PE run and computing ROQ weights from multi-banded bases. The new file format of ROQ basis and weights for this code change is described here. We discussed this new file format in a PE call and decided to continue discussions in this MR. This code change does not break the workflow of using conventional .npy bases, .npz weights or .json weights.

The main code changes are summarized below:

• self.weights["{ifo}_linear/quadratic"] is now not a single ndarray of ROQ weights but a list of ndarrays for the use of multiple ROQ bases. The property basis_number_linear/quadratic specifies which basis should be used, and is automatically updated based on self.parameters.
• self.weights contains frequency_nodes_linear/quadratic, which is a list of ndarrays of frequency nodes. When self.waveform_generator is called, its frequency nodes is automatically updated to self.weights["frequency_nodes_linear/quadratic"][self.basis_number_linear/quadratic].
• self.weights contains prior_range_linear/quadratic as it is necessary information to determine which basis should be used for a likelihood evaluation.
• Only weights with the .hdf5 format can contain multiple weights. If another format is used for multiple weights, ValueError is raised.
• Basis numbers and frequency nodes are not updated when self.parameters is the same as cached parameters or only a single basis is used.
• If conventional .npy basis, .npz weight or .json weight is passed, the likelihood automatically assumes it uses only a single basis.
• _set_weights_linear/quadratic_multiband are implemented to compute ROQ weights from multibanded bases using the algorithm introduced in arXiv:2104.07813.
• roq_params are automatically retrieved from .hdf5 basis if not specified.

cc: @colm.talbot @rory-smith @hong-qi

added Likelihood label

I just took an initial look. I think I'll have more specific requests later, but in order to merge this, it needs unit testing for all of the new features.

I think we needs tests that:

• the new file format can be read.
• the correct number of bases are loaded based on the provided priors
• the likelihood agrees with the non-ROQ case

Can you put together a small multi-basis file for single and multi-banded?

added 1 commit

• db6db380 - add unit test for multiple-multibanded-ROQ

Compare with previous version

added 1 commit

• bd7376da - add unit test for multiple-multibanded-ROQ

Compare with previous version

added 1 commit

• cd491913 - add unit test for multiple-multibanded-ROQ

Compare with previous version

added 1 commit

• 64adc8dd - add unit test for multiple-multibanded-ROQ

Compare with previous version

added 1 commit

• 782ab4ef - add unit test for multiple-multibanded-ROQ

Compare with previous version

added 13 commits

• 782ab4ef...84675b8e - 10 commits from branch lscsoft:master
• eb97ab94 - bilby/gw/likelihood.py: make ROQGravitationalWaveTransient compatible with the new .hdf5 ROQ format
• f474e369 - follow naming convention in spec doc
• cb5ea513 - add unit test for multiple-multibanded-ROQ

Compare with previous version

added 1 commit

• ec203b6d - bilby/gw/likelihood/roq.py: add doc string for parameter_conversion

Compare with previous version

added 6 commits

• 42ee3a03 - bilby/gw/likelihood.py: make ROQGravitationalWaveTransient compatible with the new .hdf5 ROQ format
• 819dac78 - follow naming convention in spec doc
• cc134f40 - add unit test for multiple-multibanded-ROQ
• 074a9da7 - bilby/gw/likelihood/roq.py: add doc string for parameter_conversion
• 44550d62 - bilby/gw/likelihood/roq.py: retrieve roq_params from .hdf5 basis file
• a6e8a295 - test/gw/likelihood_test.py: add more cases for TestROQLikelihoodHDF5

Compare with previous version

added 2 commits

• 6210e515 - bilby/gw/likelihood/roq.py: retrieve roq_params from .hdf5 basis file
• dee6a1e2 - test/gw/likelihood_test.py: add more cases for TestROQLikelihoodHDF5

Compare with previous version

changed the description

Hi @colm.talbot , I am just checking in with the status of review of this MR. If you could give us any suggestions/comments/caveats etc., it would be very helpful for us, and I am happy to address them. Thank you!

Is there any way the hdf5 files can be stored on the test images rather than in the repository? This is currently done in https://git.ligo.org/lscsoft/bilby/-/blob/master/containers/dockerfile-template#L61-L67.

added 6 commits

• 0b795a92 - bilby/gw/likelihood.py: make ROQGravitationalWaveTransient compatible with the new .hdf5 ROQ format
• 16e599a3 - follow naming convention in spec doc
• 2bed8de1 - add unit test for multiple-multibanded-ROQ
• 71d56b77 - bilby/gw/likelihood/roq.py: add doc string for parameter_conversion
• a66cad6a - bilby/gw/likelihood/roq.py: retrieve roq_params from .hdf5 basis file
• 7888afa4 - test/gw/likelihood_test.py: add more cases for TestROQLikelihoodHDF5

Compare with previous version

added 6 commits

• 12fbf829 - bilby/gw/likelihood.py: make ROQGravitationalWaveTransient compatible with the new .hdf5 ROQ format
• e0aca1b7 - follow naming convention in spec doc
• 8358f3d3 - add unit test for multiple-multibanded-ROQ
• 5b706502 - bilby/gw/likelihood/roq.py: add doc string for parameter_conversion
• ad6716d2 - bilby/gw/likelihood/roq.py: retrieve roq_params from .hdf5 basis file
• 2893f328 - test/gw/likelihood_test.py: add more cases for TestROQLikelihoodHDF5

Compare with previous version