"Iced": The Elm architecture, but in Rust

https://gitlab.com/NielsRenard/dict

Introduction

Last month I wanted to try out iced: a Rust framework for building cross platform desktop (and web) applications. I also needed a quick way to create japanese flash cards for use in anki (a free and open-source flashcard program using spaced repetition). This all seemed to fit so I put together a small dictionary application using the jisho search api (found here) and a huge set of japanese example sentences (found on tatoeba.org).

Iced

Iced uses The Elm Architecture (TEA) to structure your application. I have a sizable sideproject (BeerBinder.com) running in Elm so I’m familiar with TEA, which made picking up iced a breeze. Most of the work went into writing the parser for the example sentence format.

Parsing with nom

I used the very mature parser combinators library nom, and am quite happy with the way it turned out. I always think using parser combinators is a lot of fun. At startup it parses ~150K Japanese sentence strings into my ExampleSentence datastructure, which takes around 100ms on my 4 year old thinkpad. It then spends another 2 seconds indexing them using the default Rust HashMap.

What’s next

I’m using the application daily, because it’s a lot faster than navigating multiple web pages, and can create flashcards with 1 click. Nonetheless there’s a lot I want to improve still (words with multiple meanings, sorting by presence of example sentences, cross-platform installation, web version) and features I want to add (update example sentences, generate furigana, toggle translation visiblity, expanded flashcard generation, cloze sentences). It would also be nice to turn the parser into a crate, so other people can use it, and expand it with more sources.