A demo retail website, powered by Aerospike, showcasing Key-Value, Graph, and Vector Search.

To run locally:

1. Download the kaggle fashion dataset
   1. Place the contents of the /images/ directory in the data/images/ directory.
   2. Place the contents of the /styles/ directory in the data/styles/ directory.
2. Replace the config/aerospike/features.replace.conf and config/vector/features.replace.conf with a valid Aerospike feature key file.

   Note

   The feature key file must have a line item for vector-service

3. Create the containers:

   DOCKER_BUILDKIT=0 docker-compose up -d # using docker-compose standalone

   or

   DOCKER_BUILDKIT=0 docker compose up -d # using docker 

4. Load data into the database:

   docker exec -it -w /server aerospike-client python3 load_data.py

   Note

   This will take some time. It's loading ~44,000 fashion items, creating an embedding for each image, and 100,000 customer profiles

5. Access the site at http://localhost:4173