Feature/mesh refactor

[Jammy2211]

This pull request updates the codebase to replace the use of SettingsInversion with a more general Settings object throughout the autogalaxy package. This change standardizes configuration handling for inversions, simplifies parameter naming, and updates documentation and references accordingly. Additionally, there are updates to import paths for mesh-related classes to reflect new module structures.

Settings and Configuration Standardization:

• Replaced all references and parameters named settings_inversion (and the type SettingsInversion) with settings (type Settings) in constructors, function signatures, and docstrings across modules such as abstract_fit.py, fit_imaging.py, fit_interferometer.py, and dataset.py. This includes updating default values, property assignments, and documentation for clarity and consistency. [1] [2] [3] [4]
• Updated all usages of dataset settings files from settings_inversion.json to settings.json in documentation and code comments, ensuring consistency with the new settings structure. [1] [2] [3] [4]

Aggregator and Fit Classes:

• Updated aggregator classes (FitImagingAgg, FitInterferometerAgg) and their helper functions to use the new settings parameter and property instead of settings_inversion, including internal variable names and method calls. [1] [2] [3] [4] [5] [6]

Imports and Module Structure:

• Changed import paths for mesh classes to reflect new module locations, specifically importing Mesh2DRectangular and Mesh2DDelaunay from autoarray.inversion.pixelization.mesh_grid instead of autoarray.structures.mesh.
• Updated the import for the settings object in autogalaxy/__init__.py to use autoarray.settings.Settings instead of autoarray.inversion.inversion.settings.SettingsInversion.

Documentation and Comments:

• Revised docstrings and inline comments throughout the codebase to refer to the new settings parameter, improving clarity and reflecting the updated configuration approach. [1] [2] [3] [4] [5] [6] [7] [8]

Initialization Defaults:

• Updated default assignments to ensure that if settings is not provided, a default Settings() object is instantiated, replacing previous logic that defaulted to SettingsInversion(). [1] [2]

These changes collectively improve the maintainability and clarity of inversion settings handling throughout the codebase.

[Copilot]

Pull request overview

This PR refactors the autogalaxy package to replace SettingsInversion / settings_inversion with a more general Settings / settings throughout fits, analyses, aggregators, tests, and docs, alongside updates to mesh / mapper-related APIs and import paths.

Changes:

• Replaced SettingsInversion usage with Settings across core fit/analysis code and updated persisted JSON attribute naming to settings.
• Updated mesh/mapper usage to newer APIs (e.g. mesh.mapper_from, mapper attributes) and adjusted imports / public exports accordingly.
• Updated tests, docs, and config to reflect the renamed settings object and new file naming (settings.json).

Reviewed changes

Copilot reviewed 19 out of 19 changed files in this pull request and generated 7 comments.

Show a summary per file

File	Description
test_autogalaxy/test_regression.py	Updates regression test to pass settings=aa.Settings() instead of settings_inversion.
test_autogalaxy/interferometer/test_simulate_and_fit_interferometer.py	Updates fit construction to use settings=ag.Settings(...).
test_autogalaxy/interferometer/test_fit_interferometer.py	Updates tests for new settings param and new mesh→mapper construction path.
test_autogalaxy/config/general.yaml	Removes an inversion config entry consistent with settings refactor.
docs/api/pixelization.rst	Renames documented settings type from SettingsInversion to Settings.
autogalaxy/util/__init__.py	Removes mesh util import consistent with mesh module restructuring.
autogalaxy/profiles/light/linear/abstract.py	Switches default settings type used by linear light profile inversion helper to Settings.
autogalaxy/interferometer/model/analysis.py	Renames analysis parameter/property from settings_inversion to settings.
autogalaxy/interferometer/fit_interferometer.py	Renames fit parameter/property from settings_inversion to settings and updates downstream call sites.
autogalaxy/imaging/model/analysis.py	Renames analysis parameter/property from settings_inversion to settings.
autogalaxy/imaging/fit_imaging.py	Renames fit parameter/property from settings_inversion to settings and updates downstream call sites.
autogalaxy/galaxy/to_inversion.py	Refactors inversion setup to pass settings and use mesh.mapper_from(...) directly.
autogalaxy/analysis/plotter_interface.py	Updates plotting to use mapper’s new mesh-grid attribute access pattern.
autogalaxy/analysis/analysis/dataset.py	Renames persisted JSON from settings_inversion.json to settings.json.
autogalaxy/aggregator/interferometer/fit_interferometer.py	Updates aggregator loading/override argument to settings and reads settings from fit outputs.
autogalaxy/aggregator/imaging/fit_imaging.py	Updates aggregator loading/override argument to settings and reads settings from fit outputs.
autogalaxy/aggregator/ellipse/fit_ellipse.py	Updates docstring language to refer to settings.
autogalaxy/abstract_fit.py	Renames stored attribute to self.settings and updates doc references accordingly.
autogalaxy/__init__.py	Updates public exports to Settings and updates mesh class import paths.

Comments suppressed due to low confidence (1)

autogalaxy/abstract_fit.py:86

• sparse_operator is accessed unconditionally via self.dataset.sparse_operator, even though the docstring states it should only be called when self.settings.use_sparse_operator is enabled. If dataset.sparse_operator is expensive to build, this defeats the intended optimization. Consider checking the setting before touching self.dataset.sparse_operator, or update the docstring to reflect the actual behavior.

    def sparse_operator(self) -> Optional[aa.ImagingSparseOperator]:
        """
        Only call the `sparse_operator` property of a dataset used to perform efficient linear algebra calculations if
        the Settings()` object has `use_sparse_operator=True`, to avoid unnecessary computation.

        Returns
        -------
        The sparse operator used for efficient linear algebra calculations for this inversion, if enabled.
        """
        if self.dataset.sparse_operator is not None:
            if self.total_mappers > 0:
                return self.dataset.sparse_operator

---

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.

[Copilot]

LightProfileLinearObjFuncList.__init__ uses settings=aa.Settings() as a default argument, which will create a shared Settings instance across calls (and can lead to cross-test / cross-instance state leakage if Settings is mutable). Use settings: Optional[aa.Settings] = None and set settings = settings or aa.Settings() inside the method instead. Also consider fixing the regularization parameter signature to a proper typed optional with a default (current regularization=Optional[...] is not a valid type annotation).

[Copilot]

This aggregator now only loads settings via fit.value(name="settings"). This will likely break loading older fit outputs that saved settings_inversion (pre-refactor). Consider falling back to fit.value(name="settings_inversion") when settings is missing, to preserve backwards compatibility when inspecting existing databases / output directories.

Suggested change

	settings = settings or fit.value(name="settings")
	if settings is None:
	try:
	settings = fit.value(name="settings")
	except (KeyError, AttributeError):
	settings = None
	if settings is None:
	try:
	settings = fit.value(name="settings_inversion")
	except (KeyError, AttributeError):
	settings = None

[Copilot]

This aggregator now only loads settings via fit.value(name="settings"). This will likely break loading older fit outputs that saved settings_inversion (pre-refactor). Consider falling back to fit.value(name="settings_inversion") when settings is missing, to preserve backwards compatibility when inspecting existing databases / output directories.

Suggested change

	settings = settings or fit.value(name="settings")
	if settings is None:
	try:
	settings = fit.value(name="settings")
	except (KeyError, AttributeError):
	settings = None
	if settings is None:
	try:
	settings = fit.value(name="settings_inversion")
	except (KeyError, AttributeError):
	settings = None

[Copilot]

The docstring says inversion settings can be overwritten by passing a settings object, but _fit_ellipse_from does not accept a settings parameter. Either remove this sentence or add a settings parameter (if ellipse fitting actually supports it).

Suggested change

	The settings of an inversion can be overwritten by inputting a `settings` object, for example
	if you want to use a grid with a different inversion solver.

[Copilot]

The settings parameter defaults to None but is annotated as aa.Settings. For accurate typing (and to match how it’s used in the body), this should be Optional[aa.Settings] = None.

Suggested change

	settings: aa.Settings = None,
	settings: Optional[aa.Settings] = None,

[Copilot]

The settings parameter defaults to None but is annotated as aa.Settings. For accurate typing (and to match how it’s used in the body), this should be Optional[aa.Settings] = None.

Suggested change

	settings: aa.Settings = None,
	settings: Optional[aa.Settings] = None,

[Copilot]

AbstractFitInversion.__init__ accepts settings but the implementation allows None (settings or aa.Settings()). The type annotation should be Optional[aa.Settings] (and ideally the signature should default settings to None) to match actual usage from callers like FitImaging / FitInterferometer.