DM_spatial_knn.py

## @package python.demo.DM_spatial_knn
# 
#   example script showing how to fill a kdtree and
#   perform a kNN query
# 
 
import sys, math
import random
from opals import pyDM
 
 
# create empty kdtree object
kdtree = pyDM.PointIndexLeaf(pyDM.IndexType.kdtree,2,True)  # index type, dimension, thread safety
# The dimension is relevant for the knn queries. If the dimension is 2d, than the nn-query is done
# based on 2d distance. For selection based on 3d distances, create kdtree with dimension 3
 
 
# create 100 points with random coordinate ranges between -50 and 50
coord_range = [-50, 50]
for i in range(100):
    x = random.uniform(coord_range[0], coord_range[1])
    y = random.uniform(coord_range[0], coord_range[1])
    z = random.uniform(coord_range[0], coord_range[1])
    kdtree.addPoint( pyDM.Point(x,y,z) )
 
limit = kdtree.getLimit()
 
# output some tree statistics
print("kdtree contains", kdtree.sizePoint(), "points")
print("2D-limit (%.3f," % limit.xmin, "%.3f) -" % limit.ymin, "(%.3f," % limit.xmax,
      "%.3f)" % limit.ymax)
 
# set parameters for nn selection
nnCount = 10    # find 10 neighbours
searchPt = pyDM.Point((limit.xmin+limit.xmax)/2.,(limit.ymin+limit.ymax)/2., 0.)    # search at the center of the limit window 
maxSearchDist = -1  # use -1 for no distance limit 
searchMode = pyDM.SelectionMode.nearest
 
# perform actual knn query
pts = kdtree.searchPoint(nnCount,searchPt,maxSearchDist,searchMode)
 
# output found neighours
print(len(pts), "neighbours found")
print("Output neighbours...")
 
for i, pt in enumerate(pts):
    dist = math.sqrt( (searchPt.x-pt.x)**2 + (searchPt.y-pt.y)**2 )
    print("%3d. Point   (%8.3f, %8.3f, %8.3f)  dist=%5.3f  " % ((i+1), pt.x, pt.y, pt.z, dist) )