Open thaingu opened 5 years ago
I wouldn't convert the entire decision tree in one go I would do a small part then test it out with the chatbot first
Given the decision tree is quite small, why not just keep it as json and embed in the widget or even serve from a json file? It should perform better then. I definitely agree with @noobling regarding just make sample data for now
As you may have heard the client needs to be altering the decision tree over time so it needs to be persisted in a database so it can be altered.
I have assigned this to @pradjones as I have assigned him to creating the data structure. He will be responsible for storing the data on our end and working with @Joash95 to set up an API to send the JSON to the bot.
We need to convert the decision tree provided into something readable by machine, putting it into a JSON object is probably the best way to do this for early stages of the project. Once we implement a backend we can read the decision tree off the database instead. Example provided here: https://stackoverflow.com/questions/20052660/how-can-i-create-a-json-object-that-can-handle-conditionals