WikiJs is a node.js library which serves as an interface to Wikipedia (or any MediaWiki).
- Search wiki articles
- Fetch article content
- Find all links/images/categories in a article page
- Get parsed information about articles
- Find articles by geographical location
- and much more!
npm install wikijs
import wiki from 'wikijs';
// const wiki = require('wikijs').default;
wiki()
.page('Batman')
.then(page => page.info('alterEgo'))
.then(console.log); // Bruce Wayne
You can run these commands in order to build and test WikiJs:
git clone [email protected]:dijs/wiki.git
cd wiki
npm install
npm run build
npm test
In order for webpack to build wikijs properly, you must add an option to your webpack configuration file. Documentation
externals: {
"isomorphic-fetch": "fetch"
}
You can use the API options configuration:
wiki({
apiUrl: 'https://awoiaf.westeros.org/api.php',
origin: null
}).search('Winterfell');
You just need to change the API to the proper URL. This is normally just changing the subdomain of wikipedia.
wiki({ apiUrl: 'https://es.wikipedia.org/w/api.php' })
.page('Cristiano Ronaldo')
.then(page => page.info())
.then(console.log);
Read more about Cross Domain Requests here
If you need to pass authentication headers or anything else.
wiki({
headers: {
Cookie: 'name=value; name2=value2; name3=value3'
}
}).search('Winterfell');
You can use proxy with https-proxy-agent
const HttpsProxyAgent = require('https-proxy-agent')
wiki({agent: new HttpsProxyAgent('http://address:port')})
.page('Batman')
.then(page => page.info('alterEgo'))
.then(console.log) // Bruce Wayne
Query a specific page:
wiki()
.page('albert einstein')
.then(page =>
page
.chain()
.summary()
.image()
.links()
.request()
);
Or query many pages at once:
wiki()
.chain()
.geosearch(52.52437, 13.41053)
.summary()
.image()
.coordinates()
.request();
The code Wikipedia uses for infobox data is strange and complex. So I have split the parsing code into another library. You can find it here.
We not only parse out the information, but also try to transform the data into a convenient structure for data processing.
I always welcome help. Please just stick to the lint rules and write tests with each feature/fix.
Thanks to Heather van der Dys for the awesome logo!