Skip to content

Latest commit

 

History

History
42 lines (31 loc) · 1.3 KB

README.md

File metadata and controls

42 lines (31 loc) · 1.3 KB
Raw

Locality Sensitive Hashing Library on Distributed Stores

This library implements the LSH algorithm in a way that the hash tables can be stored in distributed systems such as Redis or Memcache. For smaller use cases it can also store the vectors and hashes in memory.

To read more about the algorithm see http://en.wikipedia.org/wiki/Locality-sensitive_hashing

Building

There is a script (called sbt) in the root that loads the correct sbt version to build:

  1. ./sbt update
  2. ./sbt test
  3. ./sbt assembly

Quick Start

The easiest place to start would be a simple LSH Server that uses an in memory hash tables.

Below is an example that constructs the LSH table and queries it for one vector.

  val vectors = Array(
    IndexedVector(1L, LshVector(Array(1.0, 2.0, 3.0, 4.0))),
    IndexedVector(2L, LshVector(Array(3.0, 2.0, 1.0, 9.0))),
    IndexedVector(3L, LshVector(Array(19.0, 22.0, 13.0, 13.0)))
  )
  
  val lshParams = 
    LshParams(
      hashTables = 1,
      hashFunctions = 5,
      dimensions = 4
    )

  val lsh = CosineLsh.build(lshParams, vectors)

  val firstVector = vectors.head
  val scoredResults = Await.result(lsh.queryVector(firstVector.vector, maxResults = 10))

  val firstResult = scoreResults.head
  println("Key " + firstResult.key)
  println("Score " + firstResult.score)