https://github.com/jqhoogland/open-dictionary

Wiktionary for machines (and exacting people).
https://github.com/jqhoogland/open-dictionary

Last synced: 10 days ago
JSON representation

Wiktionary for machines (and exacting people).

• Host: GitHub
• URL: https://github.com/jqhoogland/open-dictionary
• Owner: jqhoogland
• License: gpl-3.0
• Created: 2022-07-06T05:42:44.000Z (over 2 years ago)
• Default Branch: main
• Last Pushed: 2022-07-25T06:58:52.000Z (over 2 years ago)
• Last Synced: 2024-11-19T18:55:13.920Z (2 months ago)
• Language: Python
• Homepage: open-dictionary.vercel.app
• Size: 608 KB
• Stars: 3
• Watchers: 2
• Forks: 0
• Open Issues: 0
• Metadata Files:
  • Readme: README.md

Awesome Lists containing this project

README

        
# Open Dictionary

📖 [Open Dictionary](https://open-dictionary.vercel.app/) is structured, machine-readable [Wiktionary](https://www.wiktionary.org/). 

## 🚀 Start here

- Check out the [Python Client](/wiktionary).

Come back soon to check out:

- A REST API

- Or GraphQL API.

- And Typescript Client.

With these, you'll find:

- Better tools for NLP — built-in semantic web visualization, lemmatizers & tokenizers, etc. 

- Better tools for language-learning — finding translations, pronunciations, and word frequency lists.

## ☠️ The Problem

Wiktionary has no API.[^1] 

[^1]: It [actually has an api](https://en.wiktionary.org/w/api.php), but the api returns an unworkable sludge of html or [wikitext](https://en.wikipedia.org/wiki/Help:Wikitext).

As a result, querying Wiktionary programmatically has required a parser like [this one](https://github.com/Suyash458/WiktionaryParser). 

This is not ideal since these parsers work with outputted html rather than raw [wikitext](https://en.wikipedia.org/wiki/Help:Wikitext).

There's also the problem that Wiktionaries vary in standards across languages, so existing parsers are Wiktionary-specific. They throw out a lot of valuable data. 

## 🏗 The Solution

We provide tools to convert [wikitext](https://en.wikipedia.org/wiki/Help:Wikitext) into queryable ["semantic triples"](https://en.wikipedia.org/wiki/Semantic_triple)

and then expose these via easy-to-use APIs.

On top of that, we (will eventually) offer a few other niceties like tools to visualize the graph of words, lemmatize inflected forms, etc. It's going to be the best dictionary on the planet.

These triples serve as a kind of language-agnostic compile target for different wiktionaries, which means we can eventually consolidate all wiktionaries into one consistent format. 

For more, check out the [wiktionary parser](./wiktionary).

(The tools translate readily to Wikipedia, so feel free to intiative with that follow-up.)