Stardust87 / wordsea

Open source codebase powering the WordSea dictionary app.
https://wordsea.xyz
54 stars 2 forks source link
huggingface llm python pytorch svelte svelte-kit sveltekit tailwindcss typescript typesense

WordSea logo

WordSea

Let written be seen

WordSea is a SvelteKit web application designed to enhance English vocabulary. You can check it out and dive into WordSea on wordsea.xyz.

The core idea is to improve the process of learning new words by employing mnemonic techniques that associate words with their definition-based visualizations.

While some words naturally lend themselves to visual representations (think "apple" or "parachute"), this project's approach addresses the challenge of memorizing abstract concepts (such as "imagine" or "eon"). To accomplish this, we've utilized state-of-the-art LLMs and Text-to-Image models to generate definition-based visual representations of a subset of the English vocabulary. These visualizations are later combined with word definitions, IPA pronunciation, audio recording, and derivative information to create compact word cards, forming comprehensive dictionary entries.

Embark on your journey with WordSea and let the written be seen.

Setup

If you want to run the application locally, follow the instructions below.

Prerequisites

Start by cloning the repository

git clone https://github.com/Stardust87/wordsea
cd wordsea

Application

cd app
npm install
cd ..
docker compose up

Database

Data is stored in MongoDB and its dump is available for download here.

Check the container id and restore the dump

docker exec -i [container_id] mongorestore --gzip --nsInclude="wordsea.*" --archive < [dump_path]

Word cards

cache meaning

slapdash meaning

amenity meaning