Integrate Redis as a Vector Database with OpenAI

Using Redis as a Vector Database with OpenAI

This notebook provides an introduction to using Redis as a vector database with OpenAI embeddings. Redis is a scalable, real-time database that can be used as a vector database when using the RediSearch Module. The RediSearch module allows you to index and search for vectors in Redis. This notebook will show you how to use the RediSearch module to index and search for vectors created by using the OpenAI API and stored in Redis.

What is Redis?

Most developers from a web services background are probably familiar with Redis. At it's core, Redis is an open-source key-value store that can be used as a cache, message broker, and database. Developers choice Redis because it is fast, has a large ecosystem of client libraries, and has been deployed by major enterprises for years.

In addition to the traditional uses of Redis. Redis also provides Redis Modules which are a way to extend Redis with new data types and commands. Example modules include RedisJSON, RedisTimeSeries, RedisBloom and RediSearch.

What is RediSearch?

RediSearch is a Redis module that provides querying, secondary indexing, full-text search and vector search for Redis. To use RediSearch, you first declare indexes on your Redis data. You can then use the RediSearch clients to query that data. For more information on the feature set of RediSearch, see the README or the RediSearch documentation.

Deployment options

There are a number of ways to deploy Redis. For local development, the quickest method is to use the Redis Stack docker container which we will use here. Redis Stack contains a number of Redis modules that can be used together to create a fast, multi-model data store and query engine.

For production use cases, The easiest way to get started is to use the Redis Cloud service. Redis Cloud is a fully managed Redis service. You can also deploy Redis on your own infrastructure using Redis Enterprise. Redis Enterprise is a fully managed Redis service that can be deployed in kubernetes, on-premises or in the cloud.

Additionally, every major cloud provider (AWS Marketplace, Google Marketplace, or Azure Marketplace) offers Redis Enterprise in a marketplace offering.

Prerequisites

Before we start this project, we need to set up the following:

• start a Redis database with RediSearch (redis-stack)
• install libraries
  • Redis-py
• get your OpenAI API key

===========================================================

Start Redis

To keep this example simple, we will use the Redis Stack docker container which we can start as follows

$ docker-compose up -d

This also includes the RedisInsight GUI for managing your Redis database which you can view at http://localhost:8001 once you start the docker container.

You're all set up and ready to go! Next, we import and create our client for communicating with the Redis database we just created.

Install Requirements

Redis-Py is the python client for communicating with Redis. We will use this to communicate with our Redis-stack database.

===========================================================

Prepare your OpenAI API key

The OpenAI API key is used for vectorization of query data.

If you don't have an OpenAI API key, you can get one from https://beta.openai.com/account/api-keys.

Once you get your key, please add it to your environment variables as OPENAI_API_KEY by using following command:

Load data

In this section we'll load embedded data that has already been converted into vectors. We'll use this data to create an index in Redis and then search for similar vectors.

Connect to Redis

Now that we have our Redis database running, we can connect to it using the Redis-py client. We will use the default host and port for the Redis database which is localhost:6379.

Creating a Search Index in Redis

The below cells will show how to specify and create a search index in Redis. We will:

1. Set some constants for defining our index like the distance metric and the index name
2. Define the index schema with RediSearch fields
3. Create the index

Load Documents into the Index

Now that we have a search index, we can load documents into it. We will use the same documents we used in the previous examples. In Redis, either the HASH or JSON (if using RedisJSON in addition to RediSearch) data types can be used to store documents. We will use the HASH data type in this example. The below cells will show how to load documents into the index.

Simple Vector Search Queries with OpenAI Query Embeddings

Now that we have a search index and documents loaded into it, we can run search queries. Below we will provide a function that will run a search query and return the results. Using this function we run a few queries that will show how you can utilize Redis as a vector database.

Hybrid Queries with Redis

The previous examples showed how run vector search queries with RediSearch. In this section, we will show how to combine vector search with other RediSearch fields for hybrid search. In the below example, we will combine vector search with full text search.

HNSW Index

Up until now, we've been using the FLAT or "brute-force" index to run our queries. Redis also supports the HNSW index which is a fast, approximate index. The HNSW index is a graph-based index that uses a hierarchical navigable small world graph to store vectors. The HNSW index is a good choice for large datasets where you want to run approximate queries.

HNSW will take longer to build and consume more memory for most cases than FLAT but will be faster to run queries on, especially for large datasets.

The following cells will show how to create an HNSW index and run queries with it using the same data as before.