# Openverse's IIIT Kottayam Query Bot

![Openverse Logo](openverse2.svg)

## Overview

Openverse's IIIT Kottayam Query Bot is an advanced, AI-powered chatbot designed to answer questions about IIIT Kottayam using a curated knowledge base. It features:
- **Retrieval-Augmented Generation (RAG):** Answers are strictly based on a CSV knowledge base, ensuring factual and contextually relevant responses.
- **Groq Model Rotation:** Automatically rotates between multiple Groq LLMs to avoid rate limits and maximize reliability.
- **Admin Panel:** Allows admins to review and resolve unanswered questions, adding new knowledge to the database.
- **Google Sheets Integration:** Unanswered questions are logged to a shared Google Sheet for collaborative review.
- **Modern UI:** Streamlit-based interface with branding, watermark, and suggested question bubbles for a user-friendly experience.

---

## Features

- **Strict Contextual Answers:** The bot only answers using the provided CSV context. If the answer is not found, it responds with a fallback phrase.
- **Model Rotation:** Avoids Groq API rate limits by rotating models every 10 minutes.
- **Admin Workflow:** Unanswered questions are logged to Google Sheets and can be resolved via the admin panel, which updates the CSV knowledge base.
- **Branding:** Includes Openverse branding and a watermark image in the chat area.
- **Suggested Questions:** Displays clickable question bubbles to help users get started.
- **Vectorstore Inspector:** Inspect and test the FAISS vectorstore directly from the sidebar.

---

## Quickstart

### 1. Clone the Repository
```bash
git clone <your-repo-url>
cd RagBot
```

### 2. Install Requirements
```bash
pip install -r requirements.txt
```

### 3. Environment Setup
- Create a `.env` file with your Groq API key:
  ```env
  GROQ_API_KEY=your_groq_api_key_here
  ```
- Place your Google service account JSON (for Sheets logging) in the project root.
- Share your Google Sheet with the service account email.

### 4. Prepare the Knowledge Base
- Edit `data.csv` to add or update Q&A pairs. Format: `TOPIC,Question,answer`

### 5. Run the App
```bash
streamlit run index.py
```

---

## File Structure

```
RagBot/
├── admin_panel.py           # Streamlit admin panel for unresolved queries
├── data.csv                 # Main knowledge base (Q&A)
├── gsheet_logger.py         # Google Sheets logging utility
├── index.py                 # Main Streamlit app
├── modules/
│   ├── chat.py              # Chat UI and logic
│   ├── chroma_inspector.py  # Vectorstore inspector
│   ├── llm.py               # LLM chain and model rotation
│   ├── pdf_handler.py       # (Legacy) PDF support
│   └── vectorstore.py       # Vectorstore loading
├── requirements.txt         # Python dependencies
├── openverse2.svg           # Branding/logo
└── ...
```

---

## Google Sheets Integration
- Unanswered questions are logged to a Google Sheet (see `gsheet_logger.py`).
- The admin panel reads from this sheet and allows admins to resolve and add new answers to the CSV.
- To set up:
  1. Create a Google Sheet and share it with your service account email.
  2. Update the Sheet ID in `gsheet_logger.py`.

---

## Customization
- **Branding:** Replace `openverse2.svg` with your own logo if desired.
- **Watermark:** The chat section displays a faded watermark image.
- **Suggested Questions:** Edit or expand the list by updating `data.csv`.

---

## Troubleshooting
- **OpenSSL/Cryptography Errors:**
  - If you see errors about OpenSSL legacy provider, set:
    ```bash
    export CRYPTOGRAPHY_OPENSSL_NO_LEGACY=1
    ```
- **Google Sheets Logging Not Working:**
  - Ensure your service account JSON is correct and the sheet is shared with the service account email.
- **Groq API Issues:**
  - Check your API key and model names in `.env` and `llm.py`.

---

## Credits
- Built by Openverse for IIIT Kottayam.
- Powered by Streamlit, Langchain, Groq, FAISS, and Google Sheets.

---

## License
MIT License

---

## Contact
For questions, suggestions, or contributions, open an issue or contact the Openverse team.