Skip to content

Visual and semantic vector similarity with Redis Stack, FastAPI, PyTorch and Huggingface.

License

Notifications You must be signed in to change notification settings

redis-developer/redis-product-search

Repository files navigation

Redis Vector Search Demo Application

This demo showcases the vector search similarity (VSS) capability within Redis Stack and Redis Enterprise. Through the RediSearch module, vector types and indexes can be added to Redis. This turns Redis into a highly performant vector database which can be used for all types of applications.

The following Redis Stack capabilities are available in this demo:

  • Vector Similarity Search
    • by image
    • by text
  • Multiple vector indexing types
    • HNSW
    • Flat (brute-force)
  • Hybrid Queries
    • Apply tags as pre-filter for vector search

Application

This app was built as a Single Page Application (SPA) with the following components:

Some inspiration was taken from this Cookiecutter project and turned into a SPA application instead of a separate front-end server approach.

General Project Structure

Much inspiration taken from tiangelo/full-stack-fastapi-template

/backend
    /productsearch
        /api
            /routes
                product.py # primary API logic lives here
        /db
            load.py # seeds Redis DB
            redis_helpers.py # redis util
        /schema
            # pydantic models for serialization/validation from API
        /tests
        /utils
        config.py
        spa.py # logic for serving compiled react project
        main.py # entrypoint
/frontend
    /public
        # index, manifest, logos, etc.
    /src
        /config
        /styles
        /views
            # primary components live here

        api.ts # logic for connecting with BE
        App.tsx # project entry
        Routes.tsk # route definitions
        ...
/data
    # folder mounted as volume in Docker
    # load script auto populates initial data from S3

Datasets

The dataset was taken from the the following Kaggle links.

Running the App with docker-compose

Before running the app, install Docker Desktop.

Using Redis Cloud (recommended)

  1. Get your Redis Cloud Database (if needed).

  2. Export Redis Endpoint Environment Variables:

    $ export REDIS_HOST=your-redis-host
    $ export REDIS_PORT=your-redis-port
    $ export REDIS_PASSOWRD=your-redis-password
  3. Run the App:

    $ docker compose -f docker-cloud-redis.yml up

The benefit of this approach is that the db will persist beyond application runs. So you can make updates and re run the app without having to provision the dataset or create another search index.

Using Redis Docker

$ docker compose -f docker-local-redis.yml up

Running without docker-compose

Run frontend

  1. Install NPM packages
    $ cd frontend/
    $ npm install
  2. Use npm to serve the application from your machine
    $ npm run start
  3. Navigate to http://localhost:3000 in a browser.

All changes to your local code will be reflected in your display in semi realtime.

Run backend

Pre-step: install poetry.

  1. cd backend
  2. poetry install to get necessary python deps
  3. poetry run start to launch uvicorn server with FastAPI app

vscode debugger

Included in the project is a ./vscode/launch.json for local debugging purposes.

Troubleshooting

Sometimes you need to clear out some Docker cached artifacts. Run docker system prune, restart Docker Desktop, and try again.

Open an issue here on GitHub and we will try to be responsive to these. Additionally, please consider contributing.