Neighborhoods

MaxPerCategoryNeighborhood

Bases: MinMaxPointsNeighborhood

Min Max Per Category Neighbors model.

Neighborhood based on minimum and maximum number of closest points to be selected and maximum per category.

Parameters:

Name	Type	Description	Default
dimension	int	Neighborhood dimensions.	3
metric	Optional[Metric]	Metric.	None
min_n	int	Minimum number of closest points in ball.	1
max_n	int	Maximum number of closest points in ball.	100
max_per_category	Optional[int]	Maximum number of neighbors for each category.	None
category	Optional[ndarray]	Attribute to determine uniqueness of each category.	None

Warning

Categories must be defined in consecutive groups, like drillhole.

Attributes:

Name	Type	Description
dimension		Neighborhood dimensions.
metric		Metric.
min_n		Minimum number of closest points in ball.
max_n		Maximum number of closest points in ball.
max_per_category		Maximum number of neighbors for each category.
category		Attribute to determine uniqueness of each category.

select(points, reference=np.zeros(3))

Neighborhood selection function.

Parameters:

Name	Type	Description	Default
points	ndarray	A numpy matrix where each row is a vector in the Euclidean Space.	required
reference	ndarray	A numpy vector indicating the position (center of the ball).	zeros(3)

Returns:

Type	Description
ndarray	This function returns the indices of the N rows of x which are the closest to the vector p according to the distance

MaxPerOctantNeighborhood

Bases: MinMaxPointsNeighborhood

Max Per Octant Neighbors model.

Neighborhood based on minimum and maximum number of closest points to be selected in a specified number of octants.

Octant are equal parts of an ellipsoid.

Parameters:

Name	Type	Description	Default
dimension	int	Neighborhood dimensions.	3
metric	Optional[Metric]	Metric.	None
min_n	int	Minimum number of closest points in ball.	1
max_per_octant	int	Maximum number of neighbors for each octant.	15
n_octant	int	Number of octants dividing the ellipsoid.	8

Attributes:

Name	Type	Description
dimension		Neighborhood dimensions.
metric		Metric.
min_n		Minimum number of closest points in ball.
max_per_octant		Maximum number of neighbors for each octant.
n_octant		Number of octants dividing the ellipsoid.

Warning

Octant must number divisible by 2 in order to have the same amount of octant in the upper and lower part of the ellipsoid.

MinMaxPointsNeighborhood

Bases: Neighborhood

MimMax Neighbors model.

Parameters:

Name	Type	Description	Default
dimension	int	Neighborhood dimension.	3
metric	Optional[Metric]	Metric.	None
min_n	int	Minimum number of closest points in ball.	1
max_n	int	Maximum number of closest points in ball.	100

Attributes:

Name	Type	Description
dimension		Neighborhood dimension.
metric		Metric.
min_n		Minimum number of closest points in ball.
max_n		Maximum number of closest points in ball.

select(points, reference=np.zeros(3))

Neighborhood selection function.

Parameters:

Name	Type	Description	Default
points	ndarray	A numpy matrix where each row is a vector in the Euclidean Space.	required
reference	ndarray	A numpy vector indicating the position (center of the ball).	zeros(3)

Returns:

Type	Description
ndarray	This function returns the indices of the N rows of x which are the closest to the vector p according to the distance.

Neighborhood

Bases: Entity

Base class for unique neighborhood. Sub-class it to have a custom selection mechanism.

select(points, reference=np.zeros(3))

Neighborhood selection function.

Parameters:

Name	Type	Description	Default
points	ndarray	A numpy matrix where each row is a vector in the Euclidean Space.	required
reference	ndarray	A numpy vector indicating the position (center of the ball).	zeros(3)

Returns:

Type	Description
ndarray	This function returns the indices of the N rows of x which are the closest to the vector p according to the distance.

which_in_ball(points, reference=np.zeros(3), radius=1.0, metric=Metric())

Function to return the indexes of the points which are in a (closed) ball in the Euclidean space.

Parameters:

Name	Type	Description	Default
points	ndarray	A numpy matrix whose rows represent vectors in the Euclidean space.	required
reference	ndarray	Position vector indicating the center of the ball.	zeros(3)
radius	float	Radius of the ball.	1.0
metric	Metric	A metric object.	Metric()

Returns:

Type	Description
ndarray	A numpy arrray containing indices of point x to keep.

which_nearest_in_ball(points, reference=np.zeros(3), radius=1.0, metric=Metric(), min_n=1, max_n=100)

Function to obtain the N closest points of a set of points to the center of a ball which are on that ball.

Parameters:

Name	Type	Description	Default
points	ndarray	A numpy matrix where each row is a vector in the Euclidean Space.	required
reference	ndarray	A numpy vector indicating the position (center of the ball).	zeros(3)
radius	float	Radius of the ball.	1.0
metric	Metric	Metric object defining the ball.	Metric()
min_n	int	Minimum number of points in the ball to be considered otherwise returns no points.	1
max_n	int	Maximum number of points in the ball to be considered (the N closest).	100

Returns:

Type	Description
ndarray	A numpy array. This function returns the indices of the N rows of x which are the closest to the vector p according to the distance Metric and which are in the ball of center p and radius r (according to that distance). If there is less points in the ball than N, the indexes of all those points are returned. Points returned are ordered by ascending distance.

which_nearest_in_ball_with_max_per_category(points, reference=np.zeros(3), radius=1.0, metric=Metric(), min_n=1, max_n=100, max_per_category=None, category=None)

Function to obtain the N closest points of a set of points to the center of a ball which are on that ball with a maximum per category.

Parameters:

Name	Type	Description	Default
points	ndarray	A numpy matrix where each row is a vector in the Euclidean Space.	required
reference	ndarray	A numpy vector indicating the position (center of the ball).	zeros(3)
radius	float	Radius of the ball.	1.0
metric	Metric	Metric object defining the ball.	Metric()
min_n	int	Minimum number of points in the ball to be considered otherwise returns no points.	1
max_n	int	Maximum number of points in the ball to be considered (the N closest).	100
max_per_category	Optional[int]	Maximum number of neighbors for each category.	None
category	Optional[ndarray]	Attribute to determine uniqueness of each category.	None

Returns:

Type	Description
ndarray	A numpy array. This function returns the indices of the N rows of x which are the closest to the vector p according to the distance Metric and which are in the ball of center p and radius r (according to that distance). If there is less points in the ball than N, the indexes of all those points are returned. Points returned are ordered by ascending distance.

---

Last update: 2022-01-06