🧠 Product Training Manual Chatbot

Product Training Manual Chatbot is an intelligent, locally hosted question-answering system built with Python, Flask, and Sentence Transformers. It reads a training manual and a QA dataset to provide contextual and accurate responses. Ideal for onboarding, training, or support scenarios within organizations.

---

🚀 Features

• 🔍 Semantic search using sentence-transformers
• 📄 Supports .docx training manuals
• 📁 Uses a JSON QA dataset for predefined responses
• 🧠 Fallback to vector similarity-based retrieval if no direct match
• 🌐 Flask-powered web interface (HTML frontend)
• 🧪 Validation capability to test model accuracy

---

🗂️ Project Structure

.
├── AC_chatbot.py              # Main application script
├── qa_dataset.json            # Predefined QA pairs (train & validation)
├── training_manual.docx       # Source manual content
├── templates/
│   └── index.html             # Frontend UI (Flask template)
├── static/                    # (Optional) Styles or assets
├── requirements.txt           # Python dependencies
└── README.md                  # Project documentation

---

⚙️ Setup Instructions

1. Clone the repository

git clone https://github.com/yourusername/ac-chatbot.git
cd ac-chatbot

2. Create a virtual environment (optional but recommended)

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

3. Install dependencies

pip install -r requirements.txt

If requirements.txt doesn't exist, you can generate it:

pip freeze > requirements.txt

4. Add your content

• Replace training_manual.docx with your own document.
• Update qa_dataset.json with your own Q&A pairs.

5. Run the app

python AC_chatbot.py

Visit http://localhost:5000 in your browser.

---

🧪 Validation

When you start the app, it will automatically validate performance using the "validation" section in qa_dataset.json and print the accuracy to the console.

---

📦 Dependencies

• Flask
• sentence-transformers
• docx
• torch
• numpy
• (see requirements.txt for the full list)

---

📄 License

MIT License

---

🙌 Acknowledgements

• Hugging Face – Sentence Transformers
• Flask Web Framework
• Manjish Naik

AlertCreatorChatbot