Pizza order chatbot

A chat-bot to order pizza.

Requirements

• Python 3.6+
• Django
• DjangoRestFramework
• SQLite3
• NLTK

Easy Run

Windows

• Run setup.bat file to setup and run server

Linux

• Run setup.sh file to setup and run server

Run python manage.py createsuperuser to create Admin User

Manual Setup

• Create python virtual environment if needed.
• Run pip install -r requirement.txt to download dependencies.
• Run python manage.py migrate to setup database.
• Run python manage.py loaddata db.json to populate database with basic responses and other data.
• Run python manage.py createsuperuser to create Admin User
• Run python manage.py runserver to run server.

Usage

• Ask bot to take your order and start telling him the requirements

  • Eg: I would like to order pizza.

• Ask bot that you want to know your order status.

  • Eg: Can I my pizza status?

• Access admin page on <hostname>/admin

  • You can see orders and update order status from Admin Panel.
  • You can add more Chat-bot responses from Admin Panel.
  • You can add more pizza toppings and size from Admin Panel.

Sample Workflow

• Bot: Hello, Welcome to YoYo Pizza.
• User: Hi
• Bot: Hi there, How may I help you?
• User: I would like to order pizza.
• Bot: Here is our Menu. We have Farmhouse, Margherita, Indian spice, Desi chutney.
• User: I like Farmhouse.
• Bot: Sure, What pizza size do you want?
• User: I prefer large pizza.
• Bot: Nice, How many pizza do you want?
• User: I want 2 pizza. (Provide quantity in integer)
• Bot: Great, Can you tell me your name?
• User: Sam (Provide Name only)
• Bot: Where do you want it to be delivered?
• User: 401, Shayam Nagar (Provide Address only)
• Bot: That will be: 800Rs. Shall I confirm it.
• User: Sure.
• Bot: Thanks for ordering with us. Your Order No is: 1380165
• User: What is my order status?
• Bot: Can you provide me your order number.
• User: Here it is. 1380165
• Bot: Your Order is Placed.

Train Model

• You can add more user speeches to trainingSet.csv file and run create_classifier.py to create classifier.
• Run create_classifier.py file separately when server is not running.
• create_classifier.py will create classifier and save it in media folder so it can be used by Django server.

Application Architecture

Naive Bayes Classifier is used to predict user's text context. Server predicts the context and based on that chooses a random response from database. App uses REST API to communicate to server and uses session to remember context during multiple REST requests. Regular expression is used to take order details.