LytVector

LytVector is a simple vector database for storing and searching vectors. It provides a REST API for adding vectors, retrieving vectors by ID, and performing similarity searches.

Features

• Add vectors with metadata
• Retrieve vectors by ID
• Perform similarity searches with configurable thresholds and top-K results
• Caching for efficient repeated queries
• Dockerized for easy deployment

Clone the repository:

git clone https://github.com/IndGeek/lyt-vector.git
cd lyt-vector

Copy the enviroment variables

cp .env.example .env

Running the Project

1. Running node locally

npm install

npm run build

node dist/index.js

2. Using Docker

docker-compose up --build

API Endpoints

• Add Vector: POST /add

  • Request Body:

    { 
      "content": "your content", 
      "metadata": { "key": "value" } 
    }

  • Response:

    { "id": "vector-id" }

• Get Vector by ID: GET /vector/:id

  • Response:

    { 
      "id": "vector-id", 
      "vector": [0.234, 0.678, ] 
    }

• Search Vectors: POST /search

  • Request Body:

    { 
      "query": "your query", 
      "threshold": 0.5, 
      "topK": 5 
    }

  • Response:

    [
      { 
        "id": "vector-id", 
        "score": 0.9, 
        "metadata": { "key": "value" } 
      },