Games Tag Search - powered by Typesense 🎮🔍

This demonstration showcases a tag-based search implementation powered by Typesense, using a publicly available dataset of video games from the GOG store on Kaggle here. The search interface enables querying across multiple dimensions including game titles, genres, publishers, developers, and supported operating systems.

In a search context, 'Tags' function as searchable keywords that enable precise document filtering and retrieval. In this implementation, any indexed field within a document can serve as a tag to refine search results. The interface provides an autocompleting search input that suggests tag completions across multiple fields in real-time. Users may select suggested completions from the dropdown menu or manually enter tags as "All Fields" queries, which are processed as conventional text search terms across all indexed fields.

The live demo can be found here.

Setup

1. Install Node.js dependencies (for build scripts):

npm install

2. Set up Python virtual environment and install dependencies:

npm run setup

This may take a while to finish as it sets up a new python venv

Get started

To run this project locally:

1. Start the typesense server

docker compose up

2. Download the dataset and index it into Typesense

# For local development (uses default 'xyz' key):
npm run index

# For production Typesense cluster, set the admin API key and connection details before running:
export TYPESENSE_ADMIN_API_KEY=your_admin_key_here
export TYPESENSE_PORT=443
export TYPESENSE_PROTOCOL=https

# For single-node cluster
export TYPESENSE_HOST=your-cluster.a1.typesense.net  
# For HA cluster
export TYPESENSE_HOST_NEAREST=your-cluster.a1.typesense.net  
export TYPESENSE_HOST=your-cluster-1.a1.typesense.net
export TYPESENSE_HOST_2=your-cluster-2.a1.typesense.net
export TYPESENSE_HOST_3=your-cluster-3.a1.typesense.net

3. Open the index.html file in the root of this repo in a web browser.