High-level driver for the fast alignment of Gaussian shapes against a set of reference shapes. More...

Inheritance diagram for CDPL.Shape.FastGaussianShapeAlignment:

Public Member Functions
None	__init__ ()
	Constructs the `FastGaussianShapeAlignment` instance.

None	__init__ (GaussianShape ref_shape)
	Constructs the `FastGaussianShapeAlignment` instance with ref_shape as the single reference shape. More...

None	__init__ (GaussianShapeSet ref_shapes)
	Constructs the `FastGaussianShapeAlignment` instance with the shapes in ref_shapes as the reference set. More...

int	getObjectID ()
	Returns the numeric identifier (ID) of the wrapped C++ class instance. More...

None	setResultCompareFunction (BoolAlignmentResult2Functor func)
	Specifies the function used to compare two alignment results for sorting and filtering. More...

BoolAlignmentResult2Functor	getResultCompareFunction ()
	Returns the currently configured result-compare function. More...

None	setScoringFunction (DoubleAlignmentResultFunctor func)
	Specifies the function used to score an alignment result. More...

DoubleAlignmentResultFunctor	getScoringFunction ()
	Returns the currently configured scoring function. More...

None	setResultSelectionMode (int mode)
	Sets the alignment-result selection mode (see namespace Shape.AlignmentResultSelectionMode). More...

int	getResultSelectionMode ()
	Returns the currently configured alignment-result selection mode. More...

None	setMaxNumOptimizationIterations (int max_iter)
	Sets the maximum number of overlap-optimization iterations. More...

int	getMaxNumOptimizationIterations ()
	Returns the currently configured maximum number of overlap-optimization iterations. More...

None	setOptimizationStopGradient (float grad_norm)
	Sets the gradient norm at which the overlap optimization is stopped. More...

float	getOptimizationStopGradient ()
	Returns the currently configured overlap-optimization stop gradient. More...

None	performAlignment (bool perf_align)
	Specifies whether the actual alignment shall be performed (vs. More...

bool	performAlignment ()
	Tells whether the actual alignment is performed. More...

None	optimizeOverlap (bool optimize)
	Specifies whether the overlap shall be optimized iteratively after the initial alignment. More...

bool	optimizeOverlap ()
	Tells whether the overlap is optimized iteratively. More...

None	greedyOptimization (bool greedy)
	Specifies whether the overlap optimization shall use a greedy strategy that stops at the first local maximum. More...

bool	greedyOptimization ()
	Tells whether the overlap optimization uses a greedy strategy. More...

None	setSymmetryThreshold (float thresh)
	Sets the relative threshold for treating two principal moments as equal. More...

float	getSymmetryThreshold ()
	Returns the currently configured symmetry threshold. More...

None	genShapeCenterStarts (bool generate)
	Enables or disables the generation of a starting transformation at the shape centroid. More...

bool	genShapeCenterStarts ()
	Tells whether a starting transformation at the shape centroid is generated. More...

None	genColorCenterStarts (bool generate)
	Enables or disables the generation of starting transformations at color (pharmacophore) feature centers. More...

bool	genColorCenterStarts ()
	Tells whether starting transformations at color (pharmacophore) feature centers are generated. More...

None	genNonColorCenterStarts (bool generate)
	Enables or disables the generation of starting transformations at non-color (shape) element centers. More...

bool	genNonColorCenterStarts ()
	Tells whether starting transformations at non-color (shape) element centers are generated. More...

None	genRandomStarts (bool generate)
	Enables or disables the generation of random starting transformations. More...

bool	genRandomStarts ()
	Tells whether random starting transformations are generated. More...

None	genForAlignedShapeCenters (bool generate)
	Specifies whether element-/color-center starts shall be generated for centers of the aligned shape. More...

bool	genForAlignedShapeCenters ()
	Tells whether element-/color-center starts are generated for centers of the aligned shape. More...

None	genForReferenceShapeCenters (bool generate)
	Specifies whether element-/color-center starts shall be generated for centers of the reference shape. More...

bool	genForReferenceShapeCenters ()
	Tells whether element-/color-center starts are generated for centers of the reference shape. More...

None	genForLargerShapeCenters (bool generate)
	Specifies whether element-/color-center starts shall be generated for centers of the shape with more elements (instead of both shapes). More...

bool	genForLargerShapeCenters ()
	Tells whether element-/color-center starts are generated for centers of the shape with more elements (instead of both shapes). More...

None	setMaxRandomTranslation (float max_trans)
	Sets the maximum random translation magnitude applied to random starts. More...

float	getMaxRandomTranslation ()
	Returns the currently configured maximum random translation. More...

None	setNumRandomStarts (int num_starts)
	Sets the number of random starting transformations. More...

int	getNumRandomStarts ()
	Returns the currently configured number of random starts. More...

None	setRandomSeed (int seed)
	Sets the seed used by the random number generator that produces the random starts. More...

None	clearReferenceShapes ()
	Removes all reference shapes and reference shape sets.

None	addReferenceShape (GaussianShape shape, bool new_set=True)
	Adds shape to the reference shapes. More...

None	addReferenceShapes (GaussianShapeSet shapes, bool new_set=True)
	Adds the shapes in shapes to the reference shapes. More...

int	getNumReferenceShapes ()
	Returns the total number of reference shapes. More...

bool	align (GaussianShape shape)
	Aligns shape against all reference shapes. More...

bool	align (GaussianShapeSet shapes)
	Aligns each shape in shapes against all reference shapes. More...

int	getNumResults ()
	Returns the number of stored alignment results. More...

AlignmentResult	getResult (int idx)
	Returns the alignment result at index idx. More...

int	__len__ ()

AlignmentResult	__getitem__ (int idx)

Static Public Attributes
float	DEF_OPTIMIZATION_STOP_GRADIENT = 1.0
	Default gradient norm at which the overlap optimization is stopped.

int	DEF_MAX_OPTIMIZATION_ITERATIONS = 20
	Default maximum number of overlap-optimization iterations.

int	DEF_RESULT_SELECTION_MODE = 3
	Default alignment-result selection mode.

float	DEF_SYMMETRY_THRESHOLD = 0.15
	Default relative threshold for treating two principal moments as equal.

int	DEF_NUM_RANDOM_STARTS = 4
	Default number of random starting transformations.

float	DEF_MAX_RANDOM_TRANSLATION = 2.0
	Default maximum random translation magnitude applied to random starts.

Properties
	objectID = property(getObjectID)

	numResults = property(getNumResults)

	resultCompareFunction = property(getResultCompareFunction, setResultCompareFunction)

	scoringFunction = property(getScoringFunction, setScoringFunction)

	resultSelectionMode = property(getResultSelectionMode, setResultSelectionMode)

	maxNumOptIterations = property(getMaxNumOptimizationIterations, setMaxNumOptimizationIterations)

	optStopGradient = property(getOptimizationStopGradient, setOptimizationStopGradient)

	perfAlignment = property(performAlignment, performAlignment)

	optOverlap = property(optimizeOverlap, optimizeOverlap)

	greedyOpt = property(greedyOptimization, greedyOptimization)

	numReferenceShapes = property(getNumReferenceShapes)

	symmetryThreshold = property(getSymmetryThreshold, setSymmetryThreshold)

	maxRandomTranslation = property(getMaxRandomTranslation, setMaxRandomTranslation)

	numRandomStarts = property(getNumRandomStarts, setNumRandomStarts)

	shapeCenterStarts = property(genShapeCenterStarts, genShapeCenterStarts)

	colorCenterStarts = property(genColorCenterStarts, genColorCenterStarts)

	nonColorCenterStarts = property(genNonColorCenterStarts, genNonColorCenterStarts)

	randomStarts = property(getNumRandomStarts, setNumRandomStarts)

	genForAlignedShapeCtrs = property(genForAlignedShapeCenters, genForAlignedShapeCenters)

	genForReferenceShapeCrs = property(genForReferenceShapeCenters, genForReferenceShapeCenters)

	genForLargerShapeCtrs = property(genForLargerShapeCenters, genForLargerShapeCenters)

Detailed Description

High-level driver for the fast alignment of Gaussian shapes against a set of reference shapes.

Unlike Shape.GaussianShapeAlignment, which delegates the alignment to a separate Shape.GaussianShapeFunctionAlignment instance, FastGaussianShapeAlignment integrates the entire pipeline (start-transform generation, BFGS overlap optimization, result selection) into one class and operates on a self-contained Gaussian-product representation. The configuration combines settings exposed by Shape.PrincipalAxesAlignmentStartGenerator (start-transform options) and the iterative overlap optimization.

Constructor & Destructor Documentation

◆ init() [1/2]

None CDPL.Shape.FastGaussianShapeAlignment.__init__ ( GaussianShape ref_shape )

Constructs the FastGaussianShapeAlignment instance with ref_shape as the single reference shape.

Parameters

ref_shape The reference shape.

◆ init() [2/2]

None CDPL.Shape.FastGaussianShapeAlignment.__init__ ( GaussianShapeSet ref_shapes )

Constructs the FastGaussianShapeAlignment instance with the shapes in ref_shapes as the reference set.

Parameters

ref_shapes The reference shapes.

Member Function Documentation

◆ getObjectID()

int CDPL.Shape.FastGaussianShapeAlignment.getObjectID ( )

Returns the numeric identifier (ID) of the wrapped C++ class instance.

Different Python FastGaussianShapeAlignment instances may reference the same underlying C++ class instance. The commonly used Python expression a is not b thus cannot tell reliably whether the two FastGaussianShapeAlignment instances a and b reference different C++ objects. The numeric identifier returned by this method allows to correctly implement such an identity test via the simple expression a.getObjectID() != b.getObjectID().

Returns: The numeric ID of the internally referenced C++ class instance.

◆ setResultCompareFunction()

None CDPL.Shape.FastGaussianShapeAlignment.setResultCompareFunction ( BoolAlignmentResult2Functor func )

Specifies the function used to compare two alignment results for sorting and filtering.

Parameters

func	The result-compare function.

◆ getResultCompareFunction()

BoolAlignmentResult2Functor CDPL.Shape.FastGaussianShapeAlignment.getResultCompareFunction ( )

Returns the currently configured result-compare function.

Returns: A reference to the result-compare function.

◆ setScoringFunction()

None CDPL.Shape.FastGaussianShapeAlignment.setScoringFunction ( DoubleAlignmentResultFunctor func )

Specifies the function used to score an alignment result.

Parameters

func	The scoring function.

◆ getScoringFunction()

DoubleAlignmentResultFunctor CDPL.Shape.FastGaussianShapeAlignment.getScoringFunction ( )

Returns the currently configured scoring function.

Returns: A reference to the scoring function.

◆ setResultSelectionMode()

None CDPL.Shape.FastGaussianShapeAlignment.setResultSelectionMode ( int mode )

Sets the alignment-result selection mode (see namespace Shape.AlignmentResultSelectionMode).

Parameters

mode	The new result-selection mode.

◆ getResultSelectionMode()

int CDPL.Shape.FastGaussianShapeAlignment.getResultSelectionMode ( )

Returns the currently configured alignment-result selection mode.

Returns: The result-selection mode (see namespace Shape.AlignmentResultSelectionMode).

◆ setMaxNumOptimizationIterations()

None CDPL.Shape.FastGaussianShapeAlignment.setMaxNumOptimizationIterations ( int max_iter )

Sets the maximum number of overlap-optimization iterations.

Parameters

max_iter The new maximum number of iterations.

◆ getMaxNumOptimizationIterations()

int CDPL.Shape.FastGaussianShapeAlignment.getMaxNumOptimizationIterations ( )

Returns the currently configured maximum number of overlap-optimization iterations.

Returns: The maximum number of iterations.

◆ setOptimizationStopGradient()

None CDPL.Shape.FastGaussianShapeAlignment.setOptimizationStopGradient ( float grad_norm )

Sets the gradient norm at which the overlap optimization is stopped.

Parameters

grad_norm The new stop gradient norm.

◆ getOptimizationStopGradient()

float CDPL.Shape.FastGaussianShapeAlignment.getOptimizationStopGradient ( )

Returns the currently configured overlap-optimization stop gradient.

Returns: The stop gradient norm.

◆ performAlignment() [1/2]

None CDPL.Shape.FastGaussianShapeAlignment.performAlignment ( bool perf_align )

Specifies whether the actual alignment shall be performed (vs.

only evaluating overlaps in the initial pose).

Parameters

perf_align True to perform the alignment, and False to only evaluate the initial pose.

◆ performAlignment() [2/2]

bool CDPL.Shape.FastGaussianShapeAlignment.performAlignment ( )

Tells whether the actual alignment is performed.

Returns: True if the alignment is performed, and False otherwise.

◆ optimizeOverlap() [1/2]

None CDPL.Shape.FastGaussianShapeAlignment.optimizeOverlap ( bool optimize )

Specifies whether the overlap shall be optimized iteratively after the initial alignment.

Parameters

optimize True to optimize the overlap, and False to skip optimization.

◆ optimizeOverlap() [2/2]

bool CDPL.Shape.FastGaussianShapeAlignment.optimizeOverlap ( )

Tells whether the overlap is optimized iteratively.

Returns: True if the overlap is optimized, and False otherwise.

◆ greedyOptimization() [1/2]

None CDPL.Shape.FastGaussianShapeAlignment.greedyOptimization ( bool greedy )

Specifies whether the overlap optimization shall use a greedy strategy that stops at the first local maximum.

Parameters

greedy True to use the greedy strategy, and False to use the full optimization.

◆ greedyOptimization() [2/2]

bool CDPL.Shape.FastGaussianShapeAlignment.greedyOptimization ( )

Tells whether the overlap optimization uses a greedy strategy.

Returns: True if the greedy strategy is used, and False otherwise.

◆ setSymmetryThreshold()

None CDPL.Shape.FastGaussianShapeAlignment.setSymmetryThreshold ( float thresh )

Sets the relative threshold for treating two principal moments as equal.

Parameters

thresh The new symmetry threshold.

◆ getSymmetryThreshold()

float CDPL.Shape.FastGaussianShapeAlignment.getSymmetryThreshold ( )

Returns the currently configured symmetry threshold.

Returns: The symmetry threshold.

◆ genShapeCenterStarts() [1/2]

None CDPL.Shape.FastGaussianShapeAlignment.genShapeCenterStarts ( bool generate )

Enables or disables the generation of a starting transformation at the shape centroid.

Parameters

generate True to enable the shape-centroid start, and False to disable it.

◆ genShapeCenterStarts() [2/2]

bool CDPL.Shape.FastGaussianShapeAlignment.genShapeCenterStarts ( )

Tells whether a starting transformation at the shape centroid is generated.

Returns: True if a shape-centroid start is generated, and False otherwise.

◆ genColorCenterStarts() [1/2]

None CDPL.Shape.FastGaussianShapeAlignment.genColorCenterStarts ( bool generate )

Enables or disables the generation of starting transformations at color (pharmacophore) feature centers.

Parameters

generate True to enable the color-center starts, and False to disable them.

◆ genColorCenterStarts() [2/2]

bool CDPL.Shape.FastGaussianShapeAlignment.genColorCenterStarts ( )

Tells whether starting transformations at color (pharmacophore) feature centers are generated.

Returns: True if color-center starts are generated, and False otherwise.

◆ genNonColorCenterStarts() [1/2]

None CDPL.Shape.FastGaussianShapeAlignment.genNonColorCenterStarts ( bool generate )

Enables or disables the generation of starting transformations at non-color (shape) element centers.

Parameters

generate True to enable the non-color-center starts, and False to disable them.

◆ genNonColorCenterStarts() [2/2]

bool CDPL.Shape.FastGaussianShapeAlignment.genNonColorCenterStarts ( )

Tells whether starting transformations at non-color (shape) element centers are generated.

Returns: True if non-color-center starts are generated, and False otherwise.

◆ genRandomStarts() [1/2]

None CDPL.Shape.FastGaussianShapeAlignment.genRandomStarts ( bool generate )

Enables or disables the generation of random starting transformations.

Parameters

generate True to enable the random starts, and False to disable them.

◆ genRandomStarts() [2/2]

bool CDPL.Shape.FastGaussianShapeAlignment.genRandomStarts ( )

Tells whether random starting transformations are generated.

Returns: True if random starts are generated, and False otherwise.

◆ genForAlignedShapeCenters() [1/2]

None CDPL.Shape.FastGaussianShapeAlignment.genForAlignedShapeCenters ( bool generate )

Specifies whether element-/color-center starts shall be generated for centers of the aligned shape.

Parameters

generate True to consider aligned-shape centers, and False to ignore them.

◆ genForAlignedShapeCenters() [2/2]

bool CDPL.Shape.FastGaussianShapeAlignment.genForAlignedShapeCenters ( )

Tells whether element-/color-center starts are generated for centers of the aligned shape.

Returns: True if aligned-shape centers are considered, and False otherwise.

◆ genForReferenceShapeCenters() [1/2]

None CDPL.Shape.FastGaussianShapeAlignment.genForReferenceShapeCenters ( bool generate )

Specifies whether element-/color-center starts shall be generated for centers of the reference shape.

Parameters

generate True to consider reference-shape centers, and False to ignore them.

◆ genForReferenceShapeCenters() [2/2]

bool CDPL.Shape.FastGaussianShapeAlignment.genForReferenceShapeCenters ( )

Tells whether element-/color-center starts are generated for centers of the reference shape.

Returns: True if reference-shape centers are considered, and False otherwise.

◆ genForLargerShapeCenters() [1/2]

None CDPL.Shape.FastGaussianShapeAlignment.genForLargerShapeCenters ( bool generate )

Specifies whether element-/color-center starts shall be generated for centers of the shape with more elements (instead of both shapes).

Parameters

generate True to use only the centers of the larger shape, and False to use the centers of all configured shapes.

◆ genForLargerShapeCenters() [2/2]

bool CDPL.Shape.FastGaussianShapeAlignment.genForLargerShapeCenters ( )

Tells whether element-/color-center starts are generated for centers of the shape with more elements (instead of both shapes).

Returns: True if only the centers of the larger shape are used, and False otherwise.

◆ setMaxRandomTranslation()

None CDPL.Shape.FastGaussianShapeAlignment.setMaxRandomTranslation ( float max_trans )

Sets the maximum random translation magnitude applied to random starts.

Parameters

max_trans The new maximum random translation magnitude.

◆ getMaxRandomTranslation()

float CDPL.Shape.FastGaussianShapeAlignment.getMaxRandomTranslation ( )

Returns the currently configured maximum random translation.

Returns: The maximum random translation magnitude.

◆ setNumRandomStarts()

None CDPL.Shape.FastGaussianShapeAlignment.setNumRandomStarts ( int num_starts )

Sets the number of random starting transformations.

Parameters

num_starts The new number of random starts.

◆ getNumRandomStarts()

int CDPL.Shape.FastGaussianShapeAlignment.getNumRandomStarts ( )

Returns the currently configured number of random starts.

Returns: The number of random starts.

◆ setRandomSeed()

None CDPL.Shape.FastGaussianShapeAlignment.setRandomSeed ( int seed )

Sets the seed used by the random number generator that produces the random starts.

Parameters

seed	The new random seed.

◆ addReferenceShape()

None CDPL.Shape.FastGaussianShapeAlignment.addReferenceShape	(	GaussianShape	shape,
		bool	new_set = `True`
	)

Adds shape to the reference shapes.

Parameters

shape	The reference shape.
new_set	If `True`, shape is added to a new reference set. Otherwise, it is appended to the most recent reference set.

◆ addReferenceShapes()

None CDPL.Shape.FastGaussianShapeAlignment.addReferenceShapes	(	GaussianShapeSet	shapes,
		bool	new_set = `True`
	)

Adds the shapes in shapes to the reference shapes.

Parameters

shapes	The reference shapes.
new_set	If `True`, the shapes are added to a new reference set. Otherwise, they are appended to the most recent reference set.

◆ getNumReferenceShapes()

int CDPL.Shape.FastGaussianShapeAlignment.getNumReferenceShapes ( )

Returns the total number of reference shapes.

Returns: The number of reference shapes.

◆ align() [1/2]

bool CDPL.Shape.FastGaussianShapeAlignment.align ( GaussianShape shape )

Aligns shape against all reference shapes.

Parameters

shape The shape to align.

Returns: True if at least one alignment result was produced, and False otherwise.

◆ align() [2/2]

bool CDPL.Shape.FastGaussianShapeAlignment.align ( GaussianShapeSet shapes )

Aligns each shape in shapes against all reference shapes.

Parameters

shapes The shapes to align.

Returns: True if at least one alignment result was produced, and False otherwise.

◆ getNumResults()

int CDPL.Shape.FastGaussianShapeAlignment.getNumResults ( )

Returns the number of stored alignment results.

Returns: The number of alignment results.

◆ getResult()

AlignmentResult CDPL.Shape.FastGaussianShapeAlignment.getResult ( int idx )

Returns the alignment result at index idx.

Parameters

idx	The zero-based result index.

Returns: A reference to the result.

Exceptions

Base.IndexError if the number of results is zero or idx is not in the range [0, getNumResults() - 1].

◆ len()

int CDPL.Shape.FastGaussianShapeAlignment.__len__ ( )

Returns

◆ getitem()

AlignmentResult CDPL.Shape.FastGaussianShapeAlignment.__getitem__ ( int idx )

Parameters

idx

Returns

Public Member Functions

Static Public Attributes

Properties

Detailed Description

Constructor & Destructor Documentation

◆ __init__() [1/2]

◆ __init__() [2/2]

Member Function Documentation

◆ getObjectID()

◆ setResultCompareFunction()

◆ getResultCompareFunction()

◆ setScoringFunction()

◆ getScoringFunction()

◆ setResultSelectionMode()

◆ getResultSelectionMode()

◆ setMaxNumOptimizationIterations()

◆ getMaxNumOptimizationIterations()

◆ setOptimizationStopGradient()

◆ getOptimizationStopGradient()

◆ performAlignment() [1/2]

◆ performAlignment() [2/2]

◆ optimizeOverlap() [1/2]

◆ optimizeOverlap() [2/2]

◆ greedyOptimization() [1/2]

◆ greedyOptimization() [2/2]

◆ setSymmetryThreshold()

◆ getSymmetryThreshold()

◆ genShapeCenterStarts() [1/2]

◆ genShapeCenterStarts() [2/2]

◆ genColorCenterStarts() [1/2]

◆ genColorCenterStarts() [2/2]

◆ genNonColorCenterStarts() [1/2]

◆ genNonColorCenterStarts() [2/2]

◆ genRandomStarts() [1/2]

◆ genRandomStarts() [2/2]

◆ genForAlignedShapeCenters() [1/2]

◆ genForAlignedShapeCenters() [2/2]

◆ genForReferenceShapeCenters() [1/2]

◆ genForReferenceShapeCenters() [2/2]

◆ genForLargerShapeCenters() [1/2]

◆ genForLargerShapeCenters() [2/2]

◆ setMaxRandomTranslation()

◆ getMaxRandomTranslation()

◆ setNumRandomStarts()

◆ getNumRandomStarts()

◆ setRandomSeed()

◆ addReferenceShape()

◆ addReferenceShapes()

◆ getNumReferenceShapes()

◆ align() [1/2]

◆ align() [2/2]

◆ getNumResults()

◆ getResult()

◆ __len__()

◆ __getitem__()

◆ init() [1/2]

◆ init() [2/2]

◆ len()

◆ getitem()