autolens.AnalysisInterferometer#

class AnalysisInterferometer[source]#

Bases: AnalysisDataset

Analysis classes are used by PyAutoFit to fit a model to a dataset via a non-linear search.

The Analysis class defines the log_likelihood_function which fits the model to the dataset and returns the log likelihood value defining how well the model fitted the data.

It handles many other tasks, such as visualization, outputting results to hard-disk and storing results in a format that can be loaded after the model-fit is complete.

This Analysis class is used for all model-fits which fit galaxies (or objects containing galaxies like a Tracer) to an interferometer dataset.

This class stores the settings used to perform the model-fit for certain components of the model (e.g. a pixelization or inversion), the Cosmology used for the analysis and adapt images used for certain model classes.

Parameters:

dataset – The interferometer dataset that the model is fitted too.
positions_likelihood_list (Optional[PositionsLH]) – Alters the likelihood function to include a term which accounts for whether image-pixel coordinates in arc-seconds corresponding to the multiple images of each lensed source galaxy trace close to one another in their source-plane. This is a list, as it may support multiple planes, where a positions likelihood object is input for each plane (e.g. double source plane lensing).
adapt_images (Optional[AdaptImages]) – Contains the adapt-images which are used to make a pixelization’s mesh and regularization adapt to the reconstructed galaxy’s morphology.
cosmology (LensingCosmology) – The Cosmology assumed for this analysis.
settings (Settings) – Settings controlling how an inversion is fitted, for example which linear algebra formalism is used.
raise_inversion_positions_likelihood_exception (bool) – If an inversion is used without the positions_likelihood_list it is likely a systematic solution will be inferred, in which case an Exception is raised before the model-fit begins to inform the user of this. This exception is not raised if this input is False, allowing the user to perform the model-fit anyway.
title_prefix (str) – A string that is added before the title of all figures output by visualization, for example to put the name of the dataset and galaxy in the title.
shared_preloads (bool) – Opts this analysis into the cross-factor shared-state mechanism of a FactorGraphModel (see shared_state_from). Set this to True only when this analysis is one of many datacube channels that share an identical lens model, so the channel-invariant inversion quantities (e.g. the curvature_matrix) can be computed once and reused by every channel. False by default, leaving the standard per-analysis behaviour unchanged.

Methods

`adapt_images_via_instance_from`
`compute_latent_samples`	Compute latent variables from a model instance.
`compute_latent_variables`	Override to compute latent variables from the instance.
`dataset_model_via_instance_from`	Create a dataset model from a model instance, which is used to fit the dataset.
`fit_for_visualization`	Build the fit used by the visualizer.
`fit_from`	Given a model instance create a FitInterferometer object.
`galaxies_via_instance_from`	Create a list of galaxies from a model instance, which is used to fit the dataset.
`log_likelihood_function`	Given an instance of the model, where the model parameters are set via a non-linear search, fit the model instance to the interferometer dataset.
`log_likelihood_penalty_from`	Call the positions overwrite log likelihood function, which add a penalty term to the likelihood if the positions of the multiple images of the lensed source do not trace close to one another in the source plane.
`make_result`	After the non-linear search is complete create its Result, which includes:
`modify_after_fit`	Overwrite this method to modify the attributes of the Analysis class before the non-linear search begins.
`modify_before_fit`	This function is called immediately before the non-linear search begins and performs final tasks and checks before it begins.
`modify_model`
`perform_quick_update`	Perform a quick visualization update during non-linear search fitting.
`print_vram_use`	Print JAX VRAM use for a given batch size.
`raise_exceptions`
`save_attributes`	Before the model-fit begins, this routine saves attributes of the Analysis object to the files folder
`save_results`	At the end of a model-fit, this routine saves attributes of the Analysis object to the files folder such that they can be loaded after the analysis using PyAutoFit's database and aggregator tools.
`save_results_combined`
`shared_state_from`	Compute the channel-invariant inversion quantities once so they can be shared across the factors of a datacube FactorGraphModel (see autofit.Analysis.shared_state_from).
`tracer_via_instance_from`	Create a Tracer from the galaxies contained in a model instance.
`with_model`	Associate an explicit model with this analysis.

Attributes

`LATENT_BATCH_MODE`
`LATENT_KEYS`
`interferometer`
`supports_background_update`	Whether this analysis supports background quick updates.
`supports_jax_visualization`	Whether the visualizer can work directly with JAX arrays.

Result#: alias of ResultInterferometer

Visualizer#: alias of VisualizerInterferometer

property interferometer#

log_likelihood_function(instance, shared=None)[source]#

Given an instance of the model, where the model parameters are set via a non-linear search, fit the model instance to the interferometer dataset.

This function returns a log likelihood which is used by the non-linear search to guide the model-fit.

For this analysis class, this function performs the following steps:

If the analysis has a adapt image, associated the model galaxy images of this dataset to the galaxies in the model instance.
Extract attributes which model aspects of the data reductions, like the scaling the background sky and background noise.
Extracts all galaxies from the model instance and set up a Tracer, which includes ordering the galaxies by redshift to set up each Plane.
Use the Tracer and other attributes to create a FitInterferometer object, which performs steps such as creating model images of every galaxy in the plane, transforming them to the uv-plane via a Fourier transform and computing residuals, a chi-squared statistic and the log likelihood.

Certain models will fail to fit the dataset and raise an exception. For example if an Inversion is used, the linear algebra calculation may be invalid and raise an Exception. In such circumstances the model is discarded and its likelihood value is passed to the non-linear search in a way that it ignores it (for example, using a value of -1.0e99).

Parameters:

instance – An instance of the model that is being fitted to the data by this analysis (whose parameters have been set via a non-linear search).
shared – The cross-factor shared state of a FactorGraphModel, computed once per evaluation by the lead factor’s shared_state_from (see that method). For this analysis it is a PreloadsInterferometer carrying the channel-invariant inversion quantities; when provided it is reused by the fit instead of being recomputed. None (the default, e.g. a standalone fit) leaves behaviour unchanged.

Returns:

The log likelihood indicating how well this model instance fitted the interferometer data.

Return type:

float

shared_state_from(instance)[source]#

Compute the channel-invariant inversion quantities once so they can be shared across the factors of a datacube FactorGraphModel (see autofit.Analysis.shared_state_from).

When shared_preloads is set, every factor of the graph is an interferometer channel sharing the same lens model, so the inversion’s curvature_matrix (F = LᵀW̃L) — the dominant inversion-setup cost — is identical for every channel. This builds it once on the lead factor and returns it inside a PreloadsInterferometer, which FactorGraphModel forwards as the shared argument to every factor’s log_likelihood_function, so each channel reuses it instead of rebuilding it.

Returns None when the analysis has not opted in (shared_preloads=False) or when the model performs no inversion, in which case no state is shared and every factor fits as normal.

The caller is responsible for the invariance contract: only enable shared_preloads when the inversion quantities really are channel-invariant (e.g. the narrow-emission-line regime where uv_wavelengths and noise_map are ~channel-invariant). Outside it, leave shared_preloads=False so each channel computes its own inversion.

fit_from(instance, preloads=None)[source]#

Given a model instance create a FitInterferometer object.

This function is used in the log_likelihood_function to fit the model to the interferometer data and compute the log likelihood.

Parameters:

instance (ModelInstance) – An instance of the model that is being fitted to the data by this analysis (whose parameters have been set via a non-linear search).
preloads – An optional PreloadsInterferometer carrying channel-invariant inversion quantities (e.g. the curvature_matrix) computed once and reused by the fit instead of being rebuilt. Supplied by the datacube shared-state path (see shared_state_from); None (the default) fits as normal.

Returns:

The fit of the plane to the interferometer dataset, which includes the log likelihood.

Return type:

FitInterferometer

save_attributes(paths)[source]#

Before the model-fit begins, this routine saves attributes of the Analysis object to the files folder such that they can be loaded after the analysis using PyAutoFit’s database and aggregator tools.

For this analysis, it uses the AnalysisDataset object’s method to output the following:

The settings associated with the inversion.

The settings associated with the pixelization.

The Cosmology.

The adapt image’s model image and galaxy images, as adapt_images.fits, if used.

This function also outputs attributes specific to lens modeling:

The positions of the brightest pixels in the lensed source which are used to discard mass models.

The following .fits files are also output via the plotter interface:

The real space mask applied to the dataset, in the PrimaryHDU of dataset.fits.
The interferometer dataset as dataset.fits (data / noise-map / uv_wavelengths).

It is common for these attributes to be loaded by many of the template aggregator functions given in the aggregator modules. For example, when using the database tools to perform a fit, the default behaviour is for the dataset, settings and other attributes necessary to perform the fit to be loaded via the pickle files output by this function.

Parameters:: paths (DirectoryPaths) – The paths object which manages all paths, e.g. where the non-linear search outputs are stored, visualization, and the pickled objects used by the aggregator output by this function.