Chat Bot

This chatbot answers questions related to Computer Engineering using a custom Q&A system.

This is a Q&A-based chatbot, not a Large Language Model (LLM).

---

🚀 Features

• Sentence embedding
• Question answering
• Customizable Q&A database

---

🛠️ Tech Stack

• Language: Python
• Frameworks / Libraries: Streamlit https://docs.streamlit.io/develop/tutorials/chat-and-llm-apps/build-conversational-apps
• Model / Tool: all-MiniLM-L6-v2 (Sentence Transformer) https://huggingface.co/sentence-transformers/all-MiniLM-L6-v2

---

▶️ Usage

Run the chatbot using Streamlit:

# example
streamlit run ChatBot.py

---

📖 Example

---

🧠 How It Works

1. When you ask a question, the sentence is embedded first.

2. The answer() function compares your question with all embedded questions in the question bank and finds the most similar one.

3. The similarity score ranges from 0 (least similar) to 1 (most similar).
   If the highest similarity score is below 0.3, the function returns "I don't know"

---

🧪 Tests

This project includes two test files:

1. EmbeddingTesting.py
   • Compares two sentences and shows how similar they are (from 0 to 1).
2. AnsweringTesting.py
   • Similar to ChatBot.py, but without the Streamlit GUI.
   • Useful if you encounter issues with Streamlit

---

❓Is it just for Computer Engineering?

No! you can customize your own question bank. just use this pattern for questionBank.json :

{
  "questions": [
    {
      "id": "optional",
      "question": "your question",
      "answer": "the answer returned by the chatbot",
      "embedding": []
    }
  ]
}

---

👤 Author

Sadra Mir Mohammad Rezaei Email: sadrarezaei4@gmail.com