I have created a pure Nim k-d tree implementation
([https://github.com/jblindsay/kdtree)](https://github.com/jblindsay/kdtree\)).
It's rather flexible compared with many existing k-d tree libraries, allowing
for single and multiple nearest neighbour searches, within-radius (circular
search areas), and within-range (rectangular search areas) spatial queries. I
plan to use this library for GIS analysis of vector (shapefiles) and LiDAR data
but thought that it may be of more general appeal. I'm still learning Nim and
so it's entirely possible that it could be improved for efficiency and code
style. All comments are welcome.
