s1lv3rj1nx/rasa-nlg

A single rasa nlg server for all rasa chatbot

rasa-nlg

• Make sure you have set up git and github repo for the project and have heroku cli installed for your specific os
• Create google sheet api. Ref Analytics Vidya
• Save the API json in your project folder and make sure to add it to .gitignore

Steps:

• heroku login
• heroku create rasa-nlg
• set env variable GSHEET_KEY in your heroku app and the value will be the entire dictionary presnt in your json file
• git push heroku main
• Check logs for build

Automate deployment:

• You can go to deploy tab in your heroku app on broweser
• Connect to github (Deployment Method)
• Serach for repo, corresponding branch and enable automatic deployment.
• So next time you push to repo app will be automatically deployed.

Automatic Ping:

• We all know heroku apps sleep after 30 min of inactivity hence we can ping it to prevent from sleeping
• Go to https://kaffeine.herokuapp.com/
• Put name of your your app, and you are done!!
• e.g http://rasa-nlg.herokuapp.com

Finally you can put this as your endpoint in your rasa bot. e.g:

nlg:

url: https://rasa-nlg.herokuapp.com/nlg

How to remove your app from Kaffeine

• http://kaffeine.herokuapp.com/#decaf