Page MenuHomePhabricator

Evaluate the French-Corsican translation API 'TraduttoriCorsu' for integration with Content Translation
Open, Needs TriagePublic

Description

Paul Franceschi, the creator of a French-Corsican translation API 'TraduttoriCorsu' reached out to offer the API for use on Content Translation. This will need evaluation of technical and legal considerations. This API is currently accessible from an Android application, a Windows application and the website www.okchakko.com.

Event Timeline

Please find below the Traduttori_corsu API specifications:

  • the API is currently hosted on www.okchakko.com site: it would be available for wikimedia.org via an authentification key for wikimedia.org (it wouldn't be a problem if several keys were needed)
  • The API is accessible via a POST: no identification other than the wikimedia.org authentication key would be required
  • for wikimedia.org, the size of the text transmitted by the POST will not be limited (so it will accept above 5000 characters per request)
  • the API is hosted by a VPS, so that a large number of calls per second would be possible. As far as I can see, it could accept 'at east 10,000 requests per day, at least 10 million characters per day'
  • the API translates plain text in French into plain text in Corsican language (any of the three main variants — cismuntincu, sartinesu or taravesu— of Corsican language can be chosen)
  • the api would be available for free to wikimedia.org
  • if necessary and in order to preserve privacy, no information other than the length of the text and the date and time of the POST will be saved on the server log (neither the IP address nor the content of the text would be saved).
  • the API is regularly evaluated using an 'open test' that bears on the first 100 words of the article of the day from the French wikipedia. Its current score is 95% (5% errors).

I recently realized, as a test, three articles for wikipedia in Corsican language (co.wikipedia.org) from the 'TraduttoriCorsu' API: Romulu è Remu (Romulus and Remus), Plutarcu (Plutarch), Disertu libicu (Libyan Desert)

Traduttori_corsu is a translator dedicated to French-Corsican translation. It is the first of a more general project which aims to provide translation pairs for endangered languages.

I remain at your disposal.

At best,
Paul

Paul Franceschi