# 🤖 AI Chatbot using FastAPI & Scikit-learn

This is a lightweight, intelligent chatbot built using **FastAPI** for the backend and **scikit-learn** for AI model integration. It supports:

- User input and bot replies  
- ML-powered response generation using cosine similarity  
- Name detection and session-based memory  
- SQLite database to store all messages  

---

## 🗂️ Project Structure



chatbot-backend/
├── app/
│ ├── api/endpoints/
│ │ └── chat.py # Main chatbot API logic
│ ├── core/ # (Optional for future configs)
│ ├── crud/
│ │ └── message.py # DB CRUD operations
│ ├── db/
│ │ ├── base.py # Base model import/export
│ │ ├── session.py # DB session creation
│ │ └── message.py # SQLAlchemy message model
│ │ └── training.py # Session memory model
│ ├── schemas/
│ │ └── message.py # Pydantic schemas for validation
│ └── services/ # (Optional business logic layer)
│
├── main.py # FastAPI app entry point
├── train_model.py # ML model trainer script
├── seed_data.py # Optional: Seed training data
├── answers.pkl # Pickled list of answers
├── matrix.pkl # TF-IDF transformed matrix
├── vectorizer.pkl # TF-IDF vectorizer
├── README.md
├── requirements.txt
├── public/ # (Static files directory)
├── venv/ # Python virtual environment

---

🚀 Getting Started

1. Clone the Repo

git clone https://github.com/vpmistry13/ai-chat-bot.git
cd ai-chat-bot

2. Install Requirements

pip install -r requirements.txt

3. Train Your Model

Update training questions and answers in train_model.py, then run:

python train_model.py

This generates:

• vectorizer.pkl
• matrix.pkl
• answers.pkl

4. Run the App

Thanks! Since your FastAPI app is structured inside the app folder and the main entry file is main.py (located at the root), you're right to run it like this:

uvicorn app.main:app --host 0.0.0.0 --port 8000

I'll update the README.md section accordingly.

---

✅ Updated "Run the Backend Server" Section for README.md

## ✅ Run the Backend Server

To start your FastAPI chatbot backend, use the following command:

```bash
uvicorn app.main:app --host 0.0.0.0 --port 8000

Once the server is running, open your browser and visit:

http://localhost:8000/docs

This will open the interactive Swagger UI where you can test your chatbot endpoints.

---

💬 Chat Functionality

• POST /chat/messages – Send user message
• GET /chat/messages – Get full message history

The bot can:

• Detect and remember names from messages like “My name is Vishal”
• Respond with name if asked: “Do you know my name?”
• Reply intelligently using your trained dataset

---

🧠 Model Training Logic

Located in train_model.py:

• Uses TfidfVectorizer
• Compares questions using cosine_similarity
• Returns the best match or fallback reply

---

✅ Example Interaction

POST /chat/messages

{
  "message": "My name is Vishal",
  "sender": "user123"
}

Bot Reply:

Hi Vishal! Thanks for letting me know your good name.
Yes Vishal, how can I help you?

---

🧾 Requirements

• Python 3.8+
• FastAPI
• Uvicorn
• SQLAlchemy
• Scikit-learn

---

📌 Future Enhancements

• Persist session memory in the database
• Add context-aware multi-turn conversations
• Integrate frontend (React, Vue, etc.)
• Cloud deployment (Railway, Render, AWS, etc.)

---

👨‍💻 Author

Vishal Mistry Full-stack Developer | ReactJS | NextJS| AWS Services | PHP Laravel | AI Integrator