The @ZoekNieuws Twitterbot aims to give people easy access to recent news about a certain topic.
When you tag @ZoekNieuws on Twitter, it will check your tweet for search terms, and otherwise search the tweet you replied to for search terms. Terms between quote marks will get priority when searching for search terms.
It can look at your own tweet, regardless of whether it is a reply or not
@ZoekNieuws test-> The bot will search for Belgian news using "test" as search term@ZoekNieuws dit is een "tweede test"-> The bot will search the news using "tweede test" as search termDit is een "derde test" voor @ZoekNieuws-> The bot will search the news using "derde test" as search term
But also at the tweet of someone else by replying to it
Zijn er nu nergens artikels over deze "hittegolf"?- reply:@ZoekNieuws-> The bot will search for "hittegolf"Waarom staat er niks in de krant over het sportpaleis optreden van de kreuners deze avond?- reply:@ZoekNieuws-> The bot will search for "krant sportpaleis kreuners"Waarom staat er niks in de krant over het sportpaleis optreden van de kreuners deze avond?- reply:@ZoekNieuws "sportpaleis kreuners"-> The bot will search for " sportpaleis kreuners"
The bot will thus look for the "rarest" words to use in its search query. Note that this is often not optimal, and using quote brackets is very much encouraged.
If you're also a developer and feeling adventurous, feel free to optimize the algorithm for better search word extraction and send a pull request with your proposal!
This bot was created out of a need @ArbiterOfTweets had for easily showing relevant news articles.
- Make sure you have NodeJS installed
- Run
npm installin the main directory - Run
node src/bot.js 'any text here'to query the bot with any test query
New news sources can easily be added by adding a new Javascript file in the src/news_sources/ folder.
The module should export a function called search(topic) that receives a string topic as input, and returns a list of article objects as results.
These articles have five fields, namely title (string), date (Date class), summary (string), url (string) and source (string, likely fixed for your news source to be the name of the news outlet).
See other modules in the src/news_sources/ folder as an example!
The unigram data is calculated using the Dutch Twitter Ngram Trends dataset.
Contributions are very welcome.
Adding new news sources (/src/news_sources) is especially appreciated!
Fork the repository and update the code, and send a pull request to this repository!