siddharth2016 / chatbot-structure

A Simple Chatbot Structure using RASA Framework. Just add your intents, entity and actions, train it and see it working !

Geek Repo:Geek Repo

Github PK Tool:Github PK Tool

NOT UPDATED - DEPRECATED - KEEPING IT JUST AS A LOG

chatbot-structure-deprecated

Deprecated.

A Simple Chatbot Structure using RASA Framework. Just add your intents, entity and actions, train it and see it working !

This project features a simple easy to use chatbot that I did for some weekend activity. The task was to have chatbot create a Jira ticket and do some minor tasks like update logs and update its status.

It was just for fun activity kind of thing.

Please use this chatbot foundation to build further bots using Rasa framework.

Below are the steps for installation of rasa framework and getting the bot working !

Steps

  1. Create virtual env - python3 -m venv venv
  2. Install dependencies

Make sure to install these dependencies in your virtual env

To use virtual env

source venv/bin/activate

Then

pip3 install -r requirements.txt

If this does not seem to work, then

pip3 install rasa==2.1.3

pip3 install spacy==2.3.4

  1. Download en_core in venv

python -m spacy download en-core-web-md

  1. Train Model

rasa train

Run this command in terminal, where the venv is activated and all dependencies are installed.

  1. Run Actions server

rasa run actions

Open a terminal, say, T1 and type this command in same repo and environment where rasa chatbot code is present. This will run the actions server, which will communicate with the client server to predict the actions results and to display them in terminal.

  1. Run Client Server

rasa shell

Open a new terminal, say, T2 and type this command in same repo and environment where rasa chatbot code is present. This will run the Client server and communicate with the actions server to predict and show the results.

Thanks for using this repository !

Namaste 🙏

About

A Simple Chatbot Structure using RASA Framework. Just add your intents, entity and actions, train it and see it working !

License:MIT License


Languages

Language:Python 59.6%Language:HTML 40.4%