Search searches the VP-tree for the k nearest neighbours of target. It returns the up to k narest neighbours and the corresponding distances in order

New3 returns a Store for searching hamming distance <= 3

actual type is *Config CurrentConfig atomically returns the current configuration

New creates a new VP-tree using the metric and items provided. The metric measures the distance between two items, so that the VP-tree can find the ne

distance returns the hamming distance between v1 and v2

Hash returns a simhash value for the document returned by the scanner

UpdateConfig atomically swaps the current configuration

This helper function compares two sets of coordinates/distances to make sure they are the same.

https://stackoverflow.com/questions/24562942/golang-how-do-i-determine-the-number-of-lines-in-a-file-efficiently

This helper function finds the k nearest neighbours of target in items. It's slower than the VPTree, but its correctness is easy to verify, so we can

Add inserts a signature and document id into the store

Add inserts a signature and document id into the store

Find searches the store for all hashes hamming distance 3 or less from the query signature. It returns the associated list of document ids.

Find searches the store for all hashes hamming distance 6 or less from the query signature. It returns the associated list of document ids.

Finish prepares the store for searching. This must be called once after all the signatures have been added via Add().

NewSliceScanner creates a scanner that returns the byte slices in tokens

This test makes sure vptree's behavior is sane with no input items

This test creates a small VPTree and makes sure its search function returns the right results