Getting started

floky exposes Python bindings to the LSH implementation of lsh-rs. An LSH implementation in Rust. floky is therefore blazingly fast.

Below shows how you can get started with floky.

from floky import SRP
import numpy as np

N = 10000
n = 100
dim = 10

# Generate some random data points
data_points = np.random.randn(N, dim)

# Do a one time (expensive) fit.
lsh = SRP(n_projections=19, n_hash_tables=10)
lsh.fit(data_points)

# Query approximated nearest neigbors in sub-linear time
query = np.random.randn(n, dim)
results = lsh.predict(query)