This is an example bilingual rasa chatbot implementation that can be connected to your Facebook page using chatfuel.
Chatfuel comes with a bunch of features one of which is providing a messenger channel to receive messages from facebook along with other attributes including what we need, user’s profile language. It also gives you a nice get started button so you can provide a nice welcome message for the user. Chatfuel also provides a backend integration webhook namely JSON API to pass the user message to the Rasa servers that i am running in my backend. Goal is to interpret the message from Chatfuel and provide a response back
The project is a quick implementation to demonstrate a desired architecture of Rasa components using the latest tensforflow embeddings in French and English
Requirements
- Docker
docker-compose up --build
You will have two servers up at localhost:5005 - serving the english model at localhost:5006 - serviing the french model
Redis tracker is running at port 6379