@hitmiss/pinecone v1.0.5

@hitmiss/pinecone - Semantic Caching with Pinecone

Overview

@hitmiss/pinecone is a TypeScript-based caching system that leverages Pinecone and vector embeddings to store and retrieve data based on semantic similarity. Instead of traditional key-value storage, it allows querying data using meaning-based searches.

Features

✅ Stores values as vector embeddings for efficient similarity-based retrieval.
✅ Uses Pinecone for scalable, high-performance vector storage.
✅ Supports semantic queries, retrieving the most relevant stored data.
✅ Adjustable minimum similarity threshold for refined search results.

Installation

Ensure you have Node.js installed, then install the required dependencies:

pnpm add @pinecone-database/pinecone dotenv

Setup

Create a .env file and add your Pinecone API key:

PINECONE_API_KEY=your_pinecone_api_key_here

Usage

1. Import and Initialize

import SemanticCache from '@hitmiss/pinecone';

const cache = new SemanticCache();

2. Store Data

await cache.set("greeting", "Hello, how are you?");

3. Retrieve Data

const result = await cache.get("hello", 0.75);
console.log(result); // Returns the most relevant stored value

Error Handling

• Throws an error if Pinecone API key is missing.
• Returns null if no relevant data is found.

License

This project is licensed under the MIT License.