Skip to content

Latest commit

 

History

History
738 lines (528 loc) · 48.5 KB

CHANGELOG.md

File metadata and controls

738 lines (528 loc) · 48.5 KB

Changelog

All notable changes to this project will be documented in this file.

The format is based on Keep a Changelog, and this project adheres to Semantic Versioning.

Unreleased

0.14.0

Note: this version includes several changes that mean runs performed using previous versions cannot be resumed using this version.

Added

  • Add experimental support for discrete parameters (#401)
  • Support remapping to the unit hypercube with the standard sampler (#398)
  • Support reparameterisations that are not one-to-one (#418)
  • Support user-defined for flow proposal classes via the nessai.proposals entry point (#411)
  • Add an inverse method to FlowModel (#419)
  • Add support for using numpy random generators as inputs to samplers and other classes (#435)
  • Add include_logger_name to configure_logger (#445)
  • Add delta logL to the state plot produced by the standard sampler (#442)
  • Support resuming the sampler from a different directory to which it was run (#429)

Changed

  • Rework progress bar to no longer use tqdm (#422)
  • nessai.proposal.flowproposal.FlowProposal submodule has been refactored into two classes (#419)
  • N has been renamed to n_samples in FlowProposal.populate (#419)
  • Use new numpy random number generation for all RNG (#435, #437, #441, )
  • Change logging output to only include the name of the base logger (nessai) (#445)
  • Renamed setup_logger to configure_logger (#445)
  • Tweaks to experimental MCMCFlowproposal (#432, #438)

Fixed

  • Fix outputs of FlowProposal.backwards_pass when discard_nans and return_z are both True (#419)
  • Fix AugmentedFlowProposal (#421)

Deprecated

  • The flow_class argument for NestedSampler is deprecated in favour of flow_proposal_class (#411)
  • setup_logger is now deprecated in favour of configure_logger (#445)

Removed

  • The deprecated FlowProposal.rejection_sampling method has been removed (#419)

0.13.2

Fixed

  • Handle error when linear_transform='None' which occurs when using bilby_pipe after the flow_config changes (#414)

0.13.1

Changed

  • Make tests that require faiss are optional in the test suite (#408)

0.13.0

Added

  • Add p-value and additional panel to indices plot (#391)
  • Add support for numpy 2.0 (#387)
  • Add support for arbitrary priors in unit hypercube with the importance sampler (#377)

Changed

  • Simplify rescaling/reparameterisation configuration #395)
  • The default reparameterisation has been changed from rescale-to-bounds to z-score standardisation (#395)
  • Change default seaborn style to avoid plotting issues on some systems (#397)
  • Rework flow configuration to use flow_config and training_config keyword arguments (#394)
  • Skip nested sampling loop and populating live points if run is already finalised (#393, #400)

Deprecated

  • Specifying model_config in the flow_config dictionary is now deprecated (#394)
  • FlowProposal.names, FlowProposal.rescaled_names and FlowProposal.rescaled_dims are now deprecated (#395)

Removed

  • rescale_parameters, boundary_inversion, inversion_type, rescale_bounds update_bounds, detect_edges, detect_edges_kwargs, have all been removed in favour of using the reparameterisations directly (#395)
  • Drop support for Python 3.8 (#396)

Experimental

  • Add experimental ClusteringFlowProposal (#380)
  • Add experimental support for using flows directly from glasflow (#386)

0.12.0

This release reworks large parts of the importance nested sampler to enable drawing i.i.d samples during sampling.

The high-level API remains unchanged but the APIs for the ImportanceNestedSampler and ImportanceFlowProposal classes have changed. Existing runs of the importance nested sampler cannot be resumed with this version.

Added

  • Add option to accumulate weights during rejection sampling (#358)
  • Add option to draw i.i.d samples during sampling when using the importance nested sampler (#362)
  • Add the OrderedSamples class for handling samples in the importance nested sampler (#362)
  • Add the in_unit_hypercube and sample_unit_hypercube methods to the model class Model (#362)
  • Add log-posterior-weights to nessai.samplers.importance.ImportanceNestedSampler (#382)
  • Add explicit support for Python 3.12 (#374)
  • Add fractional evidence stopping criterion to the importance nested sampler (#371)
  • Add option to recompute log_q when resuming the importance nested sampler instead of saving it (#368)

Changed

  • Standardize how sampling history (run statistics) are stored (#364)
  • The importance nested sampler no longer requires the to_unit_hypercube method to run (#362)
  • The ratio stopping criterion is now computed using the log-likelihood threshold instead of the live points (#362)
  • Change various defaults related to the importance nested sampler (#362)
  • Random seed is now randomly set if not specified and saved in the result file (#378)
  • Rework how weights are handled in the importance nested sampler (#376)

Fixed

  • Fix bug in with legend in nessai.plot.plot_1d_comparison (#360)
  • Fix bug with truths argument in nessai.plot.corner_plot (#375)

Removed

  • Remove the deprecated max_threads argument from nessai.flowsampler.FlowSampler and nessai.utils.threading.configure_threads (#363)

0.11.0

Added

  • Add log-posterior weights to the result dictionary and file (#341)
  • Add support for checkpoint callbacks (#355)

Changed

  • Explicitly support and test against Python 3.11 (#352)

0.10.1

Fixed

  • Relax tolerance used when checking if the log-prior is vectorised such that bilby priors are treated as vectorised (#343)

0.10.0

Added

  • birth_log_likelihoods to NestedSampler and logL_birth to the result dictionary (#318)
  • Support for non-vectorised log-prior functions (#330)
  • Add the live points to the trace plot for the standard nested sampler (#334)
  • Add an option to resume from a pickle object rather than a resume file (#337)

Changed

  • Nested samples are now stored as an array in the result object rather than a dictionary (#318)
  • Reduce the size of importance nested sampling checkpoints (#327)
  • Rename nessai.utils.bilbyutils to nessai.utils.settings (#332)
  • Changed name of dZ to dlogZ, this does not change how the stopping criterion is calculated (#333)

Fixed

  • Fix a bug with the prior bounds that occurred when bounds and names had different orders (#329)
  • Fix a bug with close_pool that lead to the pool being closed irrespective of the value (#331)

Deprecated

  • nessai.utils.bilbyutils is deprecated in favour on nessai.utils.settings and will be removed in a future release (#332)

0.9.1

Fixed

  • Fix duplicate parameters when adding reparameterisations (see #320 for details) (#321)

0.9.0

Added

  • Add importance nested sampler (#285)
  • Add support for using regex for specifying parameters in the reparametersations dictionary (#312)

Changed

  • Enable constant volume mode with uniform nball latent prior (#306)
  • Pass kwargs in RealNVP to the coupling class (#307)
  • Use log-scale on state plot (#308)
  • Support forkserver and spawn multiprocessing start methods (#313)

Fixed

  • Fix resume bug with fallback reparameterisation (#302)
  • Fix bugs caused by numpy 1.25 (#311)

0.8.1

Fixed

  • Fix incorrect sign in delta phase reparameterisation (#292)
  • Remove maximum scipy version (#295)
  • Specify three quantiles in default corner kwargs as required by corner 2.2.2 (#298)

0.8.0

Added

  • Add DeltaPhaseReparameterisation for GW analyses. (#244)
  • Add nessai.utils.sorting. (#244)
  • Add log_posterior_weights and effective_n_posterior_samples to the integral state object. (#248)
  • Add a check for the multiprocessing start method when using n_pool. (#250)
  • Add option to reverse reparameterisations in FlowProposal.
  • Add disable_vectorisation to FlowSampler. (#254)
  • Add likelihood_chunksize which allows the user to limit how many points are passed to a vectorised likelihood function at once. (#256)
  • Add allow_multi_valued_likelihood which allows for multi-valued likelihoods, e.g. that include numerical integration. (#257)
  • Add parameters keyword argument to nessai.plot.plot_trace and pass additional keyword arguments to the plotting function. (#259)
  • Add option to construct live points without non-sampling parameters. (#266)
  • Add option to use a different estimate of the shrinkage. Default remains unchanged. (#248, #269)
  • Add ScaleAndShift reparameterisation which includes Z-score normalisation. (#273)
  • Add option to specify default result file extension. (#274)

Changed

  • Refactor nessai.reparameterisations into a submodule. (#241)
  • Use torch.inference_mode instead of torch.no_grad. (#245)
  • Changed CombinedReparameterisations to sort and add reparameterisations based on their requirements. (#244, #253)
  • Refactor nessai.evidence._NSIntegralState to inherit from a base class. (#248)
  • Revert default logging level to INFO. (#249)
  • Rework logging statements to reduce the amount of information printed by default. (#249)
  • Refactor nessai.proposal.FlowProposal.verify_rescaling to be stricter. (#253)
  • Truth input in nessai.plot.corner_plot can now be an iterable or a dictionary. (#255)
  • Tweak how the prior volume is computed for the final nested sample. This will also change the evidence and posterior weights. (#248, #269)
  • Stricter handling of keyword arguments passed to NestedSampler. Unknown keyword arguments will now raise an error. (#270)
  • Rework nessai.config to have config.livepoints and config.plot which contain global settings. Some of the setting names have also changed. (#272)
  • Rescale reparameterisation is now an alias for ScaleAndShift. (#273)
  • Change the default result file extension to hdf5, old result file format can be recovered by setting it to json. (#274)
  • Optimisations to FlowProposal.populate, including changes to Model.in_bounds and how sampling from the latent prior is handled. (#277)
  • Add a maximum figure size (nessai.config.plotting.max_figsize) to prevent very large trace plots when the number of dimensions is very high. (#282)

Fixed

  • Fix a bug where setting the livepoint precision (e.g. f16) did not work. (#272)
  • Fix plotting failing when sampling large number of parameters. (#281, #282)

Removed

  • Removed nessai._NSIntegralState.reset. (#248)
  • Removed nessai.gw.legacy. (#267)
  • Removed support for changing the variance of the latent distribution via draw_latent_kwargs from FlowProposal. (#277)

0.7.1

Fixed

  • Fix bug that led to the multiprocessing pool not being used when resuming. (#261)

0.7.0

Important: in this release the flow backend changed from nflows to glasflow which increased the minimum version of PyTorch to 1.11.0.

Added

  • Add explicit support for Python 3.10. (#224)
  • Add more structure utils (get_subset_arrays, isfinite_struct). (#209)
  • Add nessai.sampler.base.BaseNestedSampler class. (#210)
  • Add option to use multinomial resampling to nessai.posterior.draw_posterior_samples. (#213, #214)
  • Add features (log_prob, sample, end_iteration, finalise, training with weights) to FlowModel. (#216)
  • Add option to checkpoint based on elapsed time. (#225)
  • Add stream option to setup_logger for setting the stream for logging.StreamHandler. (#229)
  • Add configurable periodic logging based on either the iteration or elapsed time. (#229)
  • Add glasflow dependency. (#228)
  • Add posterior_sampling_method to FlowSampler.run. (#233)
  • Add options plot_{indices, posterior, logXlogL} for disabling plots in FlowSampler.run. (#233)
  • Add FlowSampler.terminate_run. (#233)
  • Add FlowSampler.log_evidence and FlowSampler.log_evidence_error. (#233)
  • Add nessai.utils.bilbyutils. (#236)
  • Add a warning for when the final p-value for the insertion indices is less than 0.05. (#235)
  • Add reset_flow to NestedSampler for resetting the entire flow. (#238)

Changed

  • Change how threading is handled to no longer use max_threads. (#208)
  • Refactor nessai.nestedsampler into the nessai.samplers submodule. (#210)
  • Refactor nessai.flowmodel into a submodule with nessai.flowmodel.{base, utils, config}. (#216)
  • Change how noise_scale is configured FlowModel. User can now specify noise_type and noise_scale. (#216)
  • Change nessai.utils.rescaling.{logit, sigmoid} to match torch.{logit, sigmoid}. (#218)
  • Change default checkpoint interval to 10 minutes rather than after training. (#225)
  • Change flows to use glasflow.nflows instead of nflows. (#228)
  • Change close_pool to be called at the end of FlowSampler.run rather than at the end of NestedSampler.nested_sampling_loop. (#233)
  • Bump minimum PyTorch version to 1.11.0. (#230)

Fixed

  • Fixed a bug in nessai.flows.utils.configure_model that only occurred when the specified device_tag is invalid. (#216)
  • Fixed a bug in nessai.utils.sampling.draw_truncated_gaussian where the input was being changed by an in-place operation. (#217)
  • Fixed an infinite loop when resuming a run that was interrupted when switching proposal. (#237)

Deprecated

  • Setting max_threads is deprecated and will be removed in a future release. (#208)
  • nessai.nestedsampler is deprecated and will be removed in a future release. Use nessai.samplers.nestedsampler instead. (#226)
  • nessai.flows.transforms.LULinear is deprecated in favour of glasflow.nflows.transforms.LULinear and will be removed in a future release. (#228)

Removed

  • Removed unused code for saving live points in NestedSampler. (#210)
  • Removed nflows dependency. (#228)

0.6.0 - 2022-08-24

Added

  • Add a warning in Model.verify_model when Model.log_prior returns an array that has float16 precision. (#175)
  • Add more functionality for configuring live point fields and defaults. (#170)
  • Record iteration at which live points are drawn in it field of live points. (#170)
  • Add nessai.config for storing package wide defaults. (#170)
  • Add nessai.utils.testing submodule which contains functions to use during testing. (#170)
  • Add nessai.livepoint.unstructured_view and nessai.model.Model.unstructured_view for constructing unstructured views of live points. (#178)
  • Add nessai.plot.corner_plot as an alternative to plot_live_points that uses corner instead of seaborn. (#189)
  • Add new examples. (#195, #198)
  • Add filehandler_kwargs to nessai.utils.logging.setup_logger which allows the user to configure the FileHandler in the logger. (#204)
  • Add final_p_value and final_ks_statistic to NestedSampler and the result file.

Changed

  • Change default values for log-likelihood and log-prior in empty live points to be np.nan instead of zero. (#170)
  • nessai.livepoint.get_dtype now returns an instance of numpy.dtype. (#170)
  • Style for plots is no longer set globally and can be disabled completely. (#194)
  • Update examples. (#190)
  • Changed behaviour of from nessai import * to no longer imports any modules. (#201)

Fixed

  • Fixed a bug in FlowProposal.populate which occurred when the pool of samples was not empty (closes #176) (#177)
  • Fixed a bug in nessai.model.Model.new_point where the incorrect number of points were returned. (#200)

Removed

  • Drop support for Python 3.6. (#188)
  • Remove a temporary fix for #46 that was introduced in #47. (#202)

0.5.1 - 2022-06-20

Fixed

  • Fixed a bug where live points where added to the initial points with incorrect log-likelihood and log-prior. (#171)

0.5.0 - 2022-06-14

Added

  • Add dataframe_to_live_points function to nessai.livepoint for converting from a pandas.DataFrame to live points. (#133)
  • Add fallback_reparameterisation to FlowProposal. This allows the user to specify which reparameterisation to use for parameters that are not included in the reparameterisations dictionary. Default behaviour remains unchanged (defaults to no reparameterisation). (#134)
  • Add rolling_mean to nessai.utils.stats. (#136)
  • Add nessai.flows.utils.create_linear_transform as a common function for creating linear transforms in the flows. (#137)
  • Add nessai.flows.transforms.LULinear to address a bug in nflows that has not been patched and prevents the use of CUDA with LULinear. (#138)
  • Add calibration_example.py to the gravitational wave examples. (#139)
  • Add defaults keyword argument to nessai.reparameterisations.get_reparameterisation for overriding the dictionary of default reparameterisations. (#142)
  • Add explicit tests for nessai.flowsampler (#143)
  • Add more tests for nessai.reparameterisations (#145)
  • Add more tests for nessai.gw (#144)
  • Add support for vectorised likelihoods and automatically detect if the likelihood is vectorised. (#148, #166)
  • Add support for passing a user-defined pool instead of using n_pool. (#148)
  • Add an option to disable signal handling in FlowSampler. (#159)
  • Add support for ray.util.multiprocessing.Pool (fixes #162). (#163)

Changed

  • NestedSampler.plot_state now includes the log-prior volume in one of the subplots and the rolling mean of the gradient (|dlogL/dLogX|) is plotted instead of the gradient directly. (#136)
  • The figure produced by NestedSampler.plot_state now includes a legend for the different vertical lines that can appear in the subplots. (#136)
  • RealNVP and NeuralSplineFlow now use nessai.flows.utils.create_linear_transform. (#137)
  • Updated all of the examples to reflect the new defaults. (#139)
  • Rework nessai.gw.reparameterisations.get_gw_reparameterisation to use get_reparameterisation with the defaults keyword argument. (#142)
  • Switch to os.path.join for joining paths. (#143, #161)
  • Context is now passed to the transform in nessai.flows.base.NFlow enabling the use of flows with conditional transforms. (#146)
  • Add context_features to RealNVP and NeuralSplineFlows (#146)
  • Rework MaskedAutoregressiveFlow to add context_features (#146)
  • Rework how likelihood parallelisation is handled. The model now contains the pool instead of the sampler and proposals. (#148)
  • Update parallelisation_example.py to show use of n_pool and pool for parallelisation. (#148)
  • Simplify how the normalising flow is reset in FlowModel and NestedSampler. (#150)
  • Reduce logging level a some statements in FlowProposal. (#160)

Fixed

  • Fixed a bug in RescaleToBounds when using pre_rescaling without boundary inversion. (#145)
  • Fixed slow integration tests not running if a quick integration test is reran after failing. (#153)
  • Fixed a bug that prevented the use of prior_sampling=True with FlowSampler. (#156)
  • Fix issue when creating multiple instances of FlowSampler with the same output directory when resuming is enabled as reported in #155. (#157)
  • Fixed missing square-root in nessai.flows.distributions.MultivariateGaussian._sample and fix the corresponding unit test. (#158)
  • Fix issue with cosmology in ComovingDistanceConverter caused by changes to astropy.cosmology in version 5.1. (#168)
  • Fixed bug with caching in LULinear transform when reloading a checkpointed flow. (#167)

Removed

  • Removed legacy_gw_example.py (#139)
  • Removed keep_samples from FlowProposal. (#140)

0.4.0 - 2021-11-23

Added

  • Add a constant volume mode to FlowProposal. In this mode the radius of the latent contour is fixed to the q'th quantile, which by default is 0.95. (#125)
  • Add a check for resume_file when resume=True. (#126)
  • Change default logging level to WARNING. (#126)
  • Add angle-cosine reparameterisation. (#126)
  • Added an explicit check for one-dimensional models that raises a custom exception OneDimensionalModelError. (#121)
  • RealNVP and NeuralSplineFlow now raise an error if features<=1. (#121)
  • Add option in nessai.reparameterisations.Angle to set scale=None, the scale is then set as 2 * pi / angle_prior_range. (#127)
  • Add 'periodic' reparameterisation that uses scale=None in nessai.reparameterisations.Angle. (#127)
  • Add the use_default_reparameterisations option to FlowProposal to allow the use of the default reparameterisations in GWFlowProposal without specifying any reparameterisations. (#129)
  • Add chi_1, chi_2 and time_jitter to known parameters in GWFlowProposal with corresponding defaults. (#130)

Changed

  • Reparameterisation angle-sine is now an alias for RescaledToBounds instead of Angle with specific keyword arguments. (#126)
  • maximum_uninformed=None now defaults to 2 times nlive instead of np.inf. (#126)
  • nlive=2000 by default. (#126)
  • Default batch_size is now 1000. (#126)
  • Default n_neurons is now 2 times the dimensions of the normalising flow. (#126)
  • Default mode for FlowProposal is constant_volume_mode=True. (#126)
  • Proposal plots are now disabled by default. (#126)
  • cooldown now defaults to 200 to reflect the change in nlive. (#126)
  • Default optimiser is now adamw. (#126)
  • Rework AugmentedFlowProposal to work with the new defaults. (#126)
  • Model.names and Model.bounds are now properties by default and their setters include checks to verify the values provided are valid and raise errors if not. (#121)
  • Logger now has propagation enabled by default. (#128)
  • FlowProposal.configure_reparameterisations can now handle an input of None. In this case only the default reparameterisations will be added. (#129)
  • Changed default reparameterisation for gravitational-wave parameters a_1 and a_2 to 'default'. (#130)

Fixed

  • Fixed a bug where the parameters list passed to Reparameterisation (or its child classes) wasn't being copied and changes made within the reparameterisation would change the original list. (#127)

Deprecated

  • keep_samples in FlowProposal will be removed in the next minor release.

0.3.3 - 2021-11-04

Fixed

  • Fixed a bug in nessai.livepoint.dict_to_live_points when passing a dictionary where the entries contained floats instead of objects with a length raised an error. (#119)

0.3.2 - 2021-10-12

Added

  • Added more checks to the init method for nessai.reparameterisations.AnglePair to catch invalid combinations of priors and/or angle conventions. Now supports RA or azimuth defined on [-pi, pi] in addition to [0, 2pi]. (#114)
  • Add a check in nessai.flowmodel.update_config for 'noise_scale', a ValueError is now raised if noise_scale is not a float or 'adaptive'. (#115)
  • Add codespell to the pre-commit checks. (#116)

Changed

  • The dtype for tensors passed to the flow is now set using torch.get_default_dtype() rather than always using float32. (#108)
  • Incorrect values for mask in nessai.flows.realnvp.RealNVP now raise ValueError and improved the error messages returned by all the exceptions in the class. (#109)
  • Change scale of y-axis of the log-prior volume vs. log-likelihood plot from symlog to the default linear axis. (#110)
  • nessai.plot.plot_trace now includes additional parameters such as logL and logP default, previously the last two parameters (assumed to be logL and logP were always excluded). (#111)

Fixed

  • Fixed an issue where nessai.reparameterisations.AnglePair would silently break when the prior range for RA or azimuth was set to a range that wasn't [0, 2pi]. It now correctly handles both [0, 2pi] and [-pi, pi] and raises an error for any other ranges. (#114)
  • Fixed various spelling mistakes throughout the source code and documentation. (#116)

0.3.1 Minor improvements and bug fixes - 2021-08-23

This release has a few minor improvements and bug fixes. It also explicitly adds support for python 3.9, which worked previously but was not tested.

Added

  • Add in_bounds, parameter_in_bounds and sample_parameter methods to nessai.model.Model. (#90)
  • Implemented the option to specify the cosmology in nessai.gw.utils.ComovingDistanceConverter using astropy. Previously changing the value had no effect of the transformation. (#91)
  • Improve test coverage for nessai.proposal.base.Proposal (#92)
  • Add 'logit' to the default reparameterisations (#98)
  • Add example using the Rosenbrock likelihood in two dimensions (#99)
  • Add a colours argument to nessai.plot.plot_1d_comparison (#102)
  • Explicitly support Python 3.9 (Added Python 3.9 to unit tests) (#103)

Changed

  • nessai.gw.utils.DistanceConverter now inherits from abc.ABC and to_uniform_parameter and from_uniform_parameter are both abstract methods. (#91)
  • nessai.proposal.base.Proposal now inherits from abc.ABC and draw is an abstract method. (#92)
  • nessai.proposal.rejection.RejectionProposal now inherits from nessai.proposal.analytic.AnalyticProposal. Functionality is the same but the code will be easier to maintain since this removes several methods that were identical. (#93)
  • noise_scale='adaptive' option in FlowModel now correctly uses a standard deviation of 0.2 times the mean nearest neighbour separation as described in Moss 2019. Note that this feature is disabled by default, so this does not change the default behaviour. (#95)
  • Refactor nessai.utils into a submodule. (#96)
  • Change behaviour of determine_rescaled_bounds so that rescale_bounds is ignored when inversion=True. This matches the behaviour in RescaledToBounds where when boundary inversion is enabled, values are rescaled to [0, 1] and then if no inversion if applied, changed to [-1, 1]. (#96)
  • Tweaked detect_edges so that both is returned in cases where the lower and upper regions contain zero probability. (#96)
  • NestedSampler no longer checks capitalisation of flow_class when determining which proposal class to use. E.g. 'FlowProposal' and 'flowproposal' are now both valid values. (#100)
  • NestedSampler.configure_flow_proposal now raises ValueError instead of RuntimeError if flow_class is an invalid string. (#100)
  • Raise a ValueError if nessai.plot.plot_1d_comparison is called with a labels list and the length does not match the number of sets of live points being compared. (#102)
  • nessai.flow.base.BaseFlow now also inherits from abc.ABC and methods that should be defined by the user are abstract methods. (#104)
  • Changed default to fuzz=1e-12 in nessai.utils.rescaling.logit and nessai.utils.rescaling.sigmoid and improved stability. (#105)

Fixed

  • Fixed a typo in nessai.gw.utils.NullDistanceConverter.from_uniform_parameter that broke the method. (#91)
  • Fixed a bug in nessai.reparameterisations.RescaleToBounds when using offset=True and pre_rescaling where the prime prior bounds were incorrectly set. (#97)
  • Fixed a bug that prevented disabling periodic checkpointing. (#101)
  • Fixed a bug when calling nessai.plot.plot_1d_comparison with live points that contain a field with only infinite values. (#102)
  • Fixed the log Jacobian determinant for nessai.utils.rescaling.logit and nessai.utils.rescaling.sigmoid which previously did not include the Jacobian for the fuzz when it was used. (#105)

0.3.0 Testing, testing and more testing - 2021-07-05

This release contains a large number of changes related to bugs and issues that were discovered when writing more tests for nessai.

It also adds a number of feature and examples.

Note: Runs produced with previous releases are incompatible with this release and cannot be resumed with out manual intervention.

Added

  • Added code to catch errors when calling plot_live_points when gwpy is installed.
  • Added tests for _NSIntegralState.
  • Add code coverage upload
  • Added an example of using unbounded priors, bilby_unbounded_priors.py
  • Added Rescale reparameterisation that just rescales by a constant and does not require prior bounds. Also add tests for this reparameterisation.
  • Added more GW examples.
  • Added tests for AugmentedFlowProposal.
  • Added an example using AugmentedFlowProposal.
  • Added eggbox example.
  • Added an error if calling FlowProposal.rejection_sampling with FlowProposal.truncate=True but worst_q=None.
  • Add option to train using dataloaders or directly with tensors. This is faster when using CUDA.
  • Add options to train with different optimisers: Adam, AdamW, SGD
  • Add tests for NestedSampler
  • Explicitly check prior bounds when using reparameterisations. This catches cases where infinite bounds are used and break some reparameterisations. (#82)
  • Add error when calling FlowProposal.populate without initialising the proposal.
  • Add NestedSampler.plot_insertion_indices to allow for easier plotting of insertion indices.
  • Add filename keyword argument to NestedSampler.plot_trace.
  • Added batch_norm_within_layers to NeuralSplineFlow

Changed

  • Plotting logX vs logL now returns the figure is filename=None
  • NestedSampler.plot_state now has the keyword argument filename and the figure is only saved if it is specified.
  • Changed name from _NSintegralState to _NSIntegralState.
  • nessai.model.Model now inherits from abc.ABC and log_prior and log_likelihood are now abstractmethods. This prevents the class from being used without redefining those methods.
  • Updated AumgentedFlowProposal to work with current version of FlowProposal
  • Fix random seed unit tests.
  • Improved FlowProposal.reset so that all attributes that are changed by calling draw are reset.
  • Move _NSIntegralState and some functions from posterior.py to evidence.py
  • NestedSampler.check_flow_model_reset will now NOT reset the flow it has never been trained (i.e proposal.training_count==0)
  • Moved all legacy gw functions to nessai/gw/legacy.py and removed them from the coverage report.
  • Minor improvements to NestedSampler
  • Better handling on NaNs in NestedSampler.populate_live_points
  • Minor improvements to plotting in FlowProposal and moved plotting to separate methods in FlowProposal.
  • Switch to using os.path.join when joins paths.
  • Improved FlowProposal.reset
  • Renamed FlexibleRealNVP to RealNVP, shouldn't affect most uses since the default way to specify a flow is via strings in configure_model.
  • Renamed nessai.flows.utils.setup_model to configure_model.
  • Renamed nessai.flows.utils.CustomMLP to MLP
  • Changed default value for tail_bound in NeuralSplineFlow to 5.

Fixed

  • Fixed a bug when plotting the state plot from a saved instance of the sampler where the sampling time was changed based on the current time.
  • Fixed a bug when using plot_trace, plot_1d_comparison or plot_live_points with a single parameter
  • Total sampling time is now correctly displayed when producing a state plot from a saved sampler.
  • Fixed a bug when using unbounded priors related to Model.verify_model
  • Fix inversion-split with RescaleToBounds
  • Fixed AugmentedGWFlowProposal.
  • Fixed a bug with plot_live_points when the hue parameter (c) was constant.
  • Fixed a bug with the reparameterisation Rescale when scale was set to a negative number.
  • Fixed a bug where scale could not be changed in ToCartesian.
  • Fixed a error when specifying NullReparameterisation (!82)
  • Fix typo in FlowProposal.set_poolsize_scale when acceptance=0
  • Fixed unintended behaviour when rescale_parameters is a list and boundary_inversion=True, where the code would try apply inversion to all parameters in Model.names.
  • Fixed bug where z returned by FlowProposal.rejection_sampling was incorrect when using truncation (which is not recommended).
  • Fix prior_sampling
  • Fixed minor typos in nessai.proposal.flowproposal.py

Removed

  • Remove "clip" option in FlowProposal, this was unused and untested.

0.2.4 - 2021-03-08

This release includes a number of bug fixes, changes to make the GWFlowProposal consistent with LegacyGWFlowProposal and a number of new unit tests to improve test coverage.

Added

  • Add poolsize to AnalyticProposal
  • Add a test for sampling with multiprocessing.
  • Add a test for sampling with AnalyticProposal and RejectionProposal.
  • Add a test for using the proposal methods with n_pool
  • Add tests for reparameterisations.
  • Add a test for comparing GWFlowProposal and LegacyGWFlowProposal.

Changed

  • Changed prime priors in LegacyGWFlowProposal to not update. This improves efficiency.
  • Changes to the reparameterisations to the the proposal consistent with LegacayGWFlowProposal:
    • Use [-1, 1] when inversion is enabled but not applied
  • Improved errors when reparameterisations are configured incorrectly.

Fixed

  • Fixed a bug with saving results when multiprocessing is enabled.
  • Fixed a bug with AnalyticProposal introduced in the last release.
  • Fixed a bug with resuming when using certain reparameterisations.

0.2.3 - 2021-02-24

Add support for Python >= 3.6 and other minor changes and bug fixes

Added

  • Badges for DOI and PyPI versions.
  • Add support for Python >= 3.6.
  • Improve doc-strings and tweak settings for doc-strings in the documentation.
  • Add tests for plotting functions.
  • Added sections to README and docs on citing nessai.

Changed

  • Remove := operator to enable support for Python >= 3.6.
  • Plotting functions are now more consistent and all return the figure if filename=None.

Fixed

  • Fixed bug when plotting non-structured arrays with plot_1d_comparison and specifying parameters.
  • Fixed bug where plot_indices failed if using an empty array but worked with an empty list.

Removed

  • Remove plot_posterior because functionality is include in plot_live_points.
  • Remove plot_likelihood_evaluations because information is already contained in the state plot.
  • Remove plot_acceptance as it is only by augmented proposal which is subject to change.
  • Remove plot_flow.

0.2.2 - 2021-02-19

This release was added to trigger Zenodo for producing a DOI.

Added

  • Docs badge

0.2.1 - 2021-02-18

Minor repository related fixes. Core code remains unchanged.

Added

  • PyPI workflow to automatically release package to PyPI

Fixed

  • Fixed issue with README not rendering of PyPi

0.2.0 - 2021-02-18

First public release.

Added

  • Complete documentation
  • Use setup.cfg and pyproject.toml for installing package
  • reparemeterisations submodule for more specific reparameterisations
  • half_gaussian.py example

Changed

  • Change to use main instead of master
  • Default GWFlowProposal changed to used reparameterisations
  • Split proposal.py into various submodules
  • Minor updates to examples
  • max_threads default changed to 1.

Fixed

  • Fix a bug where maximum_uninformed did not have the expected behaviour.

Deprecated

  • Original GWFlowProposal method renamed to LegacyGWFlowProposal. Will be removed in the next release.