Redis OM Python Search Demo

This repository contains a basic demonstration of the object modeling and search capabilities of the Redis OM Python client for Redis.

The demo requires you to have RediSearch 2.2 or higher installed on your Redis server. The easiest way to get this in order to try out the code here is to use the supplied docker-compose.yml file, so you'll want to have Docker Desktop installed.

You'll also require a reasonably up to date version of Python 3 - I think you need 3.8 or better. I've tested this on the following version of Python on macOS Big Sur 11.6.1:

$ python3 --version
Python 3.9.5

Overview

We'll take a look at how to model domain objects in Python using Redis OM. You'll find the model in adoptable.py. These objects can then be persisted to Redis Hashes by calling save() on them. By defining which fields should be indexed and how to index them, we can tell Redis OM to build and maintain a RediSearch index for us. This allows us to leverage the fluent querying API to retrieve objects matching multiple search criteria.

In this example, we'll use a small data set of dogs and cats that are available for adoption at an animal shelter. This data is in the file animal_data.csv.

Setup

To try the code out, you'll want to clone the repo, create a Python virtual environment, install dependencies and start Redis using Docker:

$ git clone https://github.com/simonprickett/redis-om-python-search-demo.git
$ cd redis-om-python-search-demo
$ python3 -m venv venv
$ . ./venv/bin/activate
$ pip install -r requirements.txt

Then, in a second terminal:

$ cd redis-om-python-search-demo
$ docker-compose down

Important: At the time of writing, there was a bug in Redis OM Python that affects the full text search functionality. I have proposed a fix for this via a pull request, and to make the "find cats that are good with children" part of this demo work, you'll need to apply my fix to your local copy of the Redis OM Python code. When the fix has been released, I'll update this README to remove this note!

Loading the Data into Redis and Indexing it

Now it's time to load the data into Redis and setup the RediSearch index:

$ python load_adoptables.py

You should see output showing the name and Redis key of each animal loaded from animal_data.csv.

Running the Code

To query the data, run query_adoptables.py:

$ python query_adoptables.py

You can change which query runs by editing the final four lines of query_adoptables.py to add/remove comments as needed.

The example queries are as follows:

find_by_name: Find all animals whose name field is Luna
find_male_dogs: Find all animals where sex is m and species is dog
find_dogs_in_age_range: Find all animals where species is dog and age is 9 or 10
find_cats_good_with_children: Final all animals where species is cat and children (good with children) is y and where the free text description field does not contain anxious or nervous but does contain words similar to play e.g. play, playful

Shutting Down

When you're done with the demo, shut down the Redis server: