Welcome to toha_nearest_neighbor’s documentation!

Installation

Install with pip:

$ pip3 install toha_nearest_neighbor

Api Reference

toha_nearest_neighbor.brute_force_location(line_points: np.ndarray, point_cloud: np.ndarray, parallel=False) → Tuple[np.ndarray, np.ndarray]

For every point in point_cloud, find it’s nearest neighbor in line_points using a brute-force algorithm. Returns a tuple of the closest line_points locations and their associated distances.

Parameters:

line_points (np.ndarray) – a 2D numpy array with each point being described by a row, and columns of coordinates of that point. Array should be in the shape (NUM_LINE_POINTS, POINT DIMENSION)
point_cloud (np.ndarray) – a 2D numpy array with each point being described by a row, and columns of coordinates of that point. Array should be in the shape (NUM_POINT_CLOUD_POINTS, POINT DIMENSION)
parallel (bool) – enable parallel processing for the dataset. If you have more than 2,000 line points and 2,000 point cloud points this may be useful.

this function returns a (NUM_POINT_CLOUD_POINTS, POINT DIMENSION) shaped array of the points where each i th row of the returned array is a row of line_points that is closest to the i th row of point_cloud

Example:

import toha_nearest_neighbor
import numpy as np

line_points = np.array(
    [
        [0.0, 0.0],
        [1.0, 1.0],
        [2.0, 2.0],
    ]
)

point_cloud = np.array(
    [
        [0.1, -0.1], #closest to the 0-th index of line_points rows
        [2.2, 3.0], # closest to the 2-nd index of line_points rows
    ]
)

closest_line_points, distances = toha_nearest_neighbor.brute_force(line_points, point_cloud)

# [[0. 0.]
#  [2. 2.]]
print(closest_line_points)
# [0.02 1.04]
print(distances)

toha_nearest_neighbor.brute_force_index(line_points: np.ndarray, point_cloud: np.ndarray, parallel=False) → Tuple[np.ndarray, np.ndarray]

For every point in point_cloud, find it’s nearest neighbor in line_points using a brute-force algorithm. Returns a tuple of indicies of the closest line_points rows and their associated distances.

Parameters:

line_points (np.ndarray) – a 2D numpy array with each point being described by a row, and columns of coordinates of that point. Array should be in the shape (NUM_LINE_POINTS, POINT DIMENSION)
point_cloud (np.ndarray) – a 2D numpy array with each point being described by a row, and columns of coordinates of that point. Array should be in the shape (NUM_POINT_CLOUD_POINTS, POINT DIMENSION)
parallel (bool) – enable parallel processing for the dataset. If you have more than 2,000 line points and 2,000 point cloud points this may be useful.

Example:

import toha_nearest_neighbor
import numpy as np

line_points = np.array(
    [
        [0.0, 0.0],
        [1.0, 1.0],
        [2.0, 2.0],
    ]
)

point_cloud = np.array(
    [
        [0.1, -0.1], #closest to the 0-th index of line_points rows
        [2.2, 3.0], # closest to the 2-nd index of line_points rows
    ]
)

indicies, distances = toha_nearest_neighbor.brute_force_index(line_points, point_cloud)
# [0 2]
print(indicies)
# [0.02 1.04]
print(distances)

toha_nearest_neighbor.kd_tree_location(line_points: np.ndarray, point_cloud: np.ndarray, parallel=False) → Tuple[np.ndarray, np.ndarray]

For every point in point_cloud, find it’s nearest neighbor in line_points using a brute-force algorithm. Returns a tuple of the closest line_points locations and their associated distances.

Parameters:

line_points (np.ndarray) – a 2D numpy array with each point being described by a row, and columns of coordinates of that point. Array should be in the shape (NUM_LINE_POINTS, POINT DIMENSION)
point_cloud (np.ndarray) – a 2D numpy array with each point being described by a row, and columns of coordinates of that point. Array should be in the shape (NUM_POINT_CLOUD_POINTS, POINT DIMENSION)
parallel (bool) – enable parallel processing for the dataset. If you have more than 2,000 line points and 2,000 point cloud points this may be useful.

this function returns a (NUM_POINT_CLOUD_POINTS, POINT DIMENSION) shaped array of the points where each i th row of the returned array is a row of line_points that is closest to the i th row of point_cloud

Example:

import toha_nearest_neighbor
import numpy as np

line_points = np.array(
    [
        [0.0, 0.0],
        [1.0, 1.0],
        [2.0, 2.0],
    ]
)

point_cloud = np.array(
    [
        [0.1, -0.1], #closest to the 0-th index of line_points rows
        [2.2, 3.0], # closest to the 2-nd index of line_points rows
    ]
)

(closest_line_points, distances) = toha_nearest_neighbor.kd_tree_location(line_points, point_cloud, parallel = True)

# [[0. 0.]
#  [2. 2.]]
print(closest_line_points)
# [0.02 1.04]
print(distances)

toha_nearest_neighbor.kd_tree_index(line_points: np.ndarray, point_cloud: np.ndarray, parallel=False) → Tuple[np.ndarray, np.ndarray]

For every point in point_cloud, find it’s nearest neighbor in line_points using a kd-tree algorithm. Returns a tuple of indicies of the closest line_points rows and their associated distances.

Parameters:

line_points (np.ndarray) – a 2D numpy array with each point being described by a row, and columns of coordinates of that point. Array should be in the shape (NUM_LINE_POINTS, POINT DIMENSION)
point_cloud (np.ndarray) – a 2D numpy array with each point being described by a row, and columns of coordinates of that point. Array should be in the shape (NUM_POINT_CLOUD_POINTS, POINT DIMENSION)
parallel (bool) – enable parallel processing for the dataset. If you have more than 2,000 line points and 2,000 point cloud points this may be useful.

this function returns a (NUM_POINT_CLOUD_POINTS, POINT DIMENSION) shaped array of the points where each i th row of the returned array is a row of line_points that is closest to the i th row of point_cloud

Example:

import toha_nearest_neighbor
import numpy as np

line_points = np.array(
    [
        [0.0, 0.0],
        [1.0, 1.0],
        [2.0, 2.0],
    ]
)

point_cloud = np.array(
    [
        [0.1, -0.1], #closest to the 0-th index of line_points rows
        [2.2, 3.0], # closest to the 2-nd index of line_points rows
    ]
)

indicies, distances = toha_nearest_neighbor.kd_tree_index(line_points, point_cloud, parallel = True)
# [0 2]
print(indicies)
# [0.02 1.04]
print(distances)