Building a chatbot from scratch thanks to RASA NLU. Most consequent project at the Institute of Technology of Dijon
The purpose was to build a smart bot that would be able to answer to specific questions about giving word's :
-
definitions
-
translations in many languages
-
related words / synonyms.
-
This bot first queries an NLU service (RASA NLU in this repo, but the online version http://chatbot.dorian-naaji.fr/ queries LUIS.AI)
-
According to the NLU service response, the bot will query DBPEDIA with the SPARQL syntax
-
Finally, the answer will be built and sent to the user.
With not much experience in Software Development / Engineering, my team and I still did a great job for beginners. It was a nice experience to build a chatbot from scratch. Nevertheless, the actual code is not really usable at it is, and is a bit spaghetti. If you need any help at setting up a chatbot, I'm sure there's nice websites and tutorials over the web.
Definition of 'apple'
Words related to 'boat'
Translation of 'apple'. Some UI components are quite buggy ! 😅