CrocoIT RAG Chatbot

📑 Description

A powerful, Retrieval-Augmented Generation (RAG) powered customer support chatbot designed specifically for CrocoIT (Integrated Solutions Gate Inc.). This tool scrapes the official Croco IT website, processes the data into a local vector database, and uses Google's Gemini 2.5 Flash model to provide highly accurate, context-aware answers to user inquiries.

✨ Features

• Automated Web Scraping: Uses LangChain's RecursiveUrlLoader and BeautifulSoup to recursively crawl and extract text from https://crocoit.com.
• Intelligent Document Retrieval (RAG): Splits extracted website data into manageable chunks and converts them into searchable vector embeddings.
• Local Vector Database: Utilizes ChromaDB for fast, local similarity search without relying on external cloud vector stores.
• State-of-the-Art LLM: Powered by Google Gemini 2.5 Flash to generate professional, accurate responses based only on the retrieved company context.
• Interactive CLI Interface: Provides a seamless, terminal-based chatting experience for testing and interacting with the bot.

🧠 Architecture

1. Ingestion: Extracts HTML from the target website and cleans it using BeautifulSoup.
2. Chunking: Splits the text using RecursiveCharacterTextSplitter (chunk size: 1000, overlap: 200).
3. Embedding: Embeds the text chunks using HuggingFace's all-MiniLM-L6-v2 model.
4. Storage: Saves the embeddings locally in a chroma_db directory.
5. Generation: Retrieves the most relevant chunks using Maximum Marginal Relevance (MMR) and passes them to the Gemini LLM with a strict system prompt to formulate the final answer.

🚀 Getting Started

Prerequisites

Ensure you have Python 3.9+ installed on your machine. You will also need a Google Gemini API Key.

Installation

1. Clone the repository:

   git clone [https://github.com/yourusername/CrocoIT_RAG.git](https://github.com/yourusername/CrocoIT_RAG.git)
   cd CrocoIT_RAG

2. Create a Virtual Environment (Recommended):

   python -m venv venv
   source venv/bin/activate  # On Windows use `venv\Scripts\activate`

3. Install Dependencies: All required packages are listed in requirements.txt.

   pip install -r requirements.txt

Configuration

This project securely manages API keys using environment variables. Before running the notebook, you must set up your local environment file.

1. Create your .env file: Copy the provided .env.example template to create a new .env file in the root directory.

   cp .env.example .env

2. Add your API Key: Open the newly created .env file and insert your actual Google Gemini API key into the GEMINI_API_KEY variable:

   APP_NAME="CrocoIT Chatbot"
   APP_VERSION="0.1"
   GEMINI_API_KEY="your_actual_api_key_here"

   (Note: Your .env file is safely ignored by Git to prevent your API key from leaking).

(Optional Note: If you haven't already, you may want to install the python-dotenv package and add load_dotenv() to the top of your notebook so Jupyter can automatically read the variables from the .env file into os.environ!)

💻 Usage

1. Open the Jupyter Notebook:

   jupyter notebook Notebook.ipynb

2. Run all cells in the notebook sequentially.
   • The notebook will first scrape the CrocoIT website and generate the required embeddings (this may take a moment on the first run).
   • Once the database (chroma_db/) is built, it will initialize the bot interface in the final cell.
3. Chat with the bot: Interact with the bot directly in the cell output. Type your questions, and type exit or quit to gracefully shut down the chatbot.

📂 Project Structure

├── Deployment
│   ├── Dockerfile
│   ├── app.py
│   └── requirements.txt
├── assets
│   └── logo.jfif
├── .env.example
├── .gitignore
├── Notebook.ipynb
├── README.md
└── requirements.txt

⚠️ Important Notes

• Data Persistence: The vector database is saved locally in the chroma_db folder. To force a fresh scrape and re-embedding of the website, delete this folder before running the notebook.
• Fallback Protocol: The bot's system prompt is strictly configured. If it cannot find the answer within the scraped website context, it will not hallucinate; it will apologize and direct the user to contact info@crocoit.com.

💻 Testing RAG Chatbot

1. User Question: What services does CrocoIT offer?

   CrocoIT offers a variety of services, including:
   *   Business Coaching
   *   Design and development of custom e-commerce platforms (online stores)
   *   Web application development
   *   Commerce mobile apps
   *   Mobile applications for Android and iOS
   *   Building amazing web and mobile apps
   *   ERP systems
   

2. User Question: What is 2Loyal?

   2Loyal is an AI-powered loyalty management software designed to enhance customer retention. It offers customizable rewards, real-time advanced analytics, and a flexible points system to drive engagement across all digital channels. Its purpose is to foster lasting relationships and ensure customers feel valued and connected to a business.

---

Built for the CrocoIT Company Interview Task