Skip to content

Prerequisites

For the complete documentation index see: llms.txt

All documentation pages available in markdown.

This tutorial requires Docker, Python, Git, and a Bash-compatible shell. It uses small local Parquet files as stand-ins for batch feature sources that would come from a data pipeline in production. You do not need a data warehouse or streaming pipeline.

Required

Docker

Docker runs Aerospike Database locally.

  1. Download, install, and launch Docker Desktop from the Docker website.

  2. Verify Docker is running:

    Terminal window
    docker --version
Python

The pinned Feast commit requires Python 3.10 or later. Use Python 3.10 for this tutorial. Newer interpreters may work, but this tutorial was verified with Python 3.10.

  1. Install Python 3.10.

  2. Verify Python is available:

    Terminal window
    python3 --version

    The output must show Python 3.10 or newer.

Git and Bash-compatible shell

The installation command downloads Feast from a pinned Git commit. The tutorial commands use Bash syntax.

  1. Install Git.

  2. Verify Git and Bash are available:

    Terminal window
    git --version
    bash --version

    On Windows, run the documented shell commands in WSL 2.