spheni 0.1.1

A tiny CPU-first, in-memory vector search library in C++ with Python bindings.

spheni

A tiny CPU-first, in-memory vector search library in C++ with Python bindings.

• Source: https://github.com/datavorous/spheni
• License: Apache-2.0

Install

pip install spheni

Quick Example

import numpy as np
import spheni

spec = spheni.IndexSpec(4, spheni.Metric.L2, spheni.IndexKind.Flat)
engine = spheni.Engine(spec)

base = np.random.rand(10, 4).astype(np.float32)
engine.add(base)

query = np.random.rand(4).astype(np.float32)
hits = engine.search(query, 3)

for h in hits:
    print(h.id, h.score)

Features

• Indexes: Flat, IVF
• Metrics: Cosine, L2
• Storage: F32, INT8
• Operations: add, search, search_batch, train, save, load

Build From Source

python -m pip wheel . --no-deps -w dist

Then install the wheel from dist/.

Links

• Full README: https://github.com/datavorous/spheni/blob/master/README.md
• Python API: https://github.com/datavorous/spheni/blob/master/docs/wiki/python_api.md
• Build Guide: https://github.com/datavorous/spheni/blob/master/docs/wiki/building.md