lharikrishnan1993/kdtree-cpp

A C++ implementation of Kd Tree

To compile the code and generate executable, please run the run.sh using the command ./run.sh

If it requests execution rights, enter this in the commandline within the folder.

chmod +x run.sh ./run.sh

To run the code,

cd build

./kdtree This demonstrates few tasks from building, querying, serializing, deserializing, checking and so on. Details provided in the documentation. (or)

./kdtree -b /data/sample_data.csv This builds the tree from the specified .csv file. (or)

./kdtree -d /data/tree.kd This deserializes a tree which was already built

Find the generated output files at src/data/output.txt This file contains the data as instructed in the assignment question paper in the format of

index_of_nearest_neighbor_in_input_file, euclidean_distance_between_the_nn_and_query_point

For detailed documentation, please find the file doc/html/index.html

kdtree-cpp's Issues

split data using strategies rather than inserting by order.

1. Add values as a bunch rather than one after the other.
2. Add insertion strategy by changing split axis and split position
3. Provide facility to change the axis split and position split in run time.

Search error

First, implement at least the approximate search recursively and return the pointer to approximate neighbor. It can later be easily updated to accurate neighbor.