This repository provides a REST API for retrieving random words with optional filtering by language, length, count, and difficulty (based on Wikipedia word frequency). The application is built with Scala using the Cask web framework, loads word lists from JSON files at startup, and uses Redis for IP-based rate limiting. Key features include difficulty filtering via parallel Wikipedia API calls with caching and fallback mechanisms, efficient random word selection algorithms, and multi-language support through community-contributed JSON files. The architecture is monolithic with clear separation of concerns across service objects for language handling, word selection, rate limiting, and external API integration.
Ibis is a federated wiki platform built entirely in Rust, enabling users to create, edit, and discuss articles across ActivityPub-enabled instances. The application follows a full-stack architecture with a Leptos-based frontend (compiled to WASM for client-side hydration) and an Axum backend serving both the API and server-side rendering. PostgreSQL is used as the database via Diesel ORM, and federation is implemented using the `activitypub-federation` crate, allowing articles, edits, comments, and user profiles to be synchronized between independent instances. The codebase is organized as a Cargo workspace with separate crates for backend, frontend, federation logic, database access, API client, markdown rendering, and reusable UI components. Key features include markdown-based article editing with custom extensions (LaTeX, table of contents, article links), edit conflict resolution via three-way merge, user authentication (JWT, OAuth, email verification), in-app and email notifications, and Wikipedia article import. The frontend uses Tailwind CSS with DaisyUI components and supports internationalization via `leptos-fluent`. Ibis demonstrates strong software engineering practices: modular crate structure, comprehensive integration tests with multi-instance federation scenarios, CI pipeline with formatting and linting, and Docker-based deployment. While the codebase is actively maintained and well-architected, it contains some production `expect` calls and relies on beta versions of key dependencies, which introduces minor risk. Overall, it is a polished, production-ready federated application with reusable patterns in frontend components, database abstractions, and ActivityPub object handling.