Skip to content

Latest commit

 

History

History
executable file
·
64 lines (46 loc) · 1.53 KB

README.md

File metadata and controls

executable file
·
64 lines (46 loc) · 1.53 KB
Raw

RASA

The project was forked from rasa-bot.

This is a mentor bot that can help to introduce the discussion event and teach the user how to discussion, built with the rasa framework and LLM. We design ideas from ChatDev and try to build a step-by-step framework.

Design

actions:
- action_ActionAskGpt
- action_ActionGoNext

intents:
- confirm
- faq

UI Demo

img

Service

Start Service

  1. Update OpenAI API Key in docker-compose.yml
  2. Update Domain name in botroom/App.js:23 -> (frontend)
  3. up service with docker-compose
docker-compose up
  1. go to weeb
http://127.0.0.1

Ask bot by cmd

curl -o output.txt -X POST -H "Content-Type: application/json; charset=UTF-8" -d '{"sender": "user-001","message": "早安"}'  http://127.0.0.1:5005/webhooks/rest/webhook && echo -e "$(<output.txt)"
curl -o output.txt -X POST -H "Content-Type: application/json; charset=UTF-8" -d '{"sender": "user-001","message": "昆蟲是甚麼?"}'  http://127.0.0.1:5005/webhooks/rest/webhook && echo -e "$(<output.txt)"
curl -o output.txt -X POST -H "Content-Type: application/json; charset=UTF-8" -d '{"sender": "user-001","message": "是的"}'  http://127.0.0.1:5005/webhooks/rest/webhook && echo -e "$(<output.txt)"

Testing

  • with rasa tests
docker-compose up test-model
  • interactive

docker-compose up rasa-actionServer
  • Testing In docker Container
docker run -it -v ./actionsServer:/app --entrypoint=/bin/bash rasa/rasa:3.6.6-full