TextRazor now supports 17 languages, with new Entity Linking, Topic Tagging and Classification in Korean, Arabic, Greek and Ukrainian - https://t.co/qdhqJMLfx6
“How we used @TextRazor’s NLP AI to save our client money and time” - https://t.co/eOU1mARJ7b - Great read from our friends @browserlondon on their TextRazor classifier integration.
@pigsonthewing Good idea, though to keep the request size overhead down we prefer to return links and allow users to join that with other data. Multilingual labels are actually our most requested field from Wikidata. I do pass through Birmingham from time to time, would be good to chat!
@jsbullockdotnet It's best to make sure your data is consistently UTF8 at the source, but something like https://t.co/dA9skDXnv6 might help if that’s out of your control.
@jsbullockdotnet Thx James, good to hear! The SDK will throw a PHP Exception with HTTP status 400 for any bad encoding in the "analyze" call, you can catch this and log the error in your code.