ChatBot

A Machine Learning ChatBot using Flask

Getting Started

• How it Works
• Running Locally
• How to use this chatbot as a FAQ/Support/Helpdesk chatbot for your website

How it Works

This bot uses a cosine similarity to find the most similar prompt to the user's prompt and then returns the response to that prompt. The bot uses a pre-defined list of prompts and responses to generate the responses.

Running Locally

• Run Backend

    # Clone the repository
    $ git clone <repo-url>
    # Navigate to the repository
    $ cd <repo-name>
    # Navigate to backend directory
    $ cd backend
    # Install the required packages
    $ pip install -r requirements.txt
    # Run flask api
    $ flask -app api/index.py run

• Then you'll be able to access the API at http://localhost:3000/predict

• Make sure to replace the http://localhost:3000 with the correct URL in the frontend code

• Run Frontend

    # Navigate to the repository
    $ cd <repo-name>
    # Navigate to frontend directory
    $ cd frontend
    # Install the required packages
    $ npm install
    # Run the frontend
    $ npm run dev

How to use this chatbot as a FAQ/Support/Helpdesk chatbot for your website

• Create a new json file in the backend/ directory with the name intents.json
• Add data in your json with following structure:

    {
        "prompt1": ["response1", "response2", "response3"],
        "prompt2": ["response1", "response2", "response3"],
        "prompt3": ["response1", "response2", "response3"],
    }

• Go to backend/api/index.py and change the path of the intents.json file to the new file you created

   chatbot = ChatBot()
   chatbot.load_data(path='Intents.json')
   chatbot.train()

• Run the backend and frontend as mentioned in the Running Locally section
• You can now use the chatbot as a FAQ/Support/Helpdesk chatbot for your website