Fruit-Seller-Chatbot-with-Recommendation-System 🤖

Fruits/Vegetable Seller Bot With Recommendation System Integrated (Using Apriori Algorithm) And MySql Database to record all the transactions and products/stock details.

Install dependencies

pip install rasa

Installation docs: https://rasa.com/docs/rasa/installation/

Create a new python environment

conda create --name rasa python=3.6

Install all the requirements

pip install -r requirements.txt

Import SQL database "fruit_rasa" to your local server (Xampp)

Optional Installation

pip install rasa-x

Import and setup the Mysql database

Run the bot

Use rasa train to train a model.

Then, to run, first set up your action server in one terminal window:

rasa run actions

In another window, run the following command to talk to the bot:

rasa shell --debug

Note that --debug mode will produce a lot of output meant to help you understand how the bot is working under the hood. To simply talk to the bot, you can remove this flag.

You can also try out your bot locally using Rasa X by running

rasa x

Integrated Features

• Sign In and Sign Up functionality with DB support
• User can check if the company sells a particular fruit or not
• Check the price of fruits
• Fruit menu which display fruit and price per kg
• User can verify the product detail which he want to purchase before checkout
• User can place order, which will then be stored in DB
• User can get details of his last purchase as well as all purchases
• Product Recommendation using Apriori Algorithm integrated

Admin Features

• Questions
  • Who are my best customers
  • Area from, where i get max profit
  • Best selling product
  • What is my current stock of apples in mumbai
  • What is my total revenue
  • Montly revenue
  • Profit earn from selling apples in nagpur
  • Which products are not selling
  • Which fruits are giving me loss
  • What is the most asked fruit and i don't have it

Future Improvements

• Question => Price of 2 kg of apple
• Subsription feature for a particular or variety of fruits
• Give out to actual user, to make the model learn and handle extreme cases
• Check if the user already exists during sign up
• Use encryption to save password
• Check for user valid input (Sql injection)