🤖 Agentic RAG Assistant

A powerful, multimodal AI assistant designed to interact with your local documents and datasets. This project combines Retrieval-Augmented Generation (RAG) for unstructured text and Agentic Data Analysis for structured tabular data.

Built with Streamlit, LangChain, and Google Gemini.

---

🚀 Features

• Unstructured Search (RAG): Upload PDFs, TXT, or DOCX files. The agent chunks and indexes them using FAISS and Google's latest embedding models.
• Structured Data Analysis: Upload CSV or Excel files. A smart Pandas Agent can perform calculations, create summaries, and answer questions about your data.
• Intelligent Routing: The agent automatically decides whether to search through documents or analyze data based on your query.
• Manual Sync & Refresh: Complete control over when files are processed. Sync only when you're ready.
• Quick Actions: One-click buttons for common tasks like summarizing documents, extracting key info, or analyzing resumes.
• Persistent Memory: Maintains chat history for context-aware conversations.

---

🛠️ Tech Stack

• Frontend: Streamlit
• Orchestration: LangChain
• Embeddings: GoogleGenerativeAIEmbeddings
• Vector Database: FAISS
• Models: Google Gemini 2.5 Flash
• Data Handling: Pandas, RecursiveCharacterTextSplitter, PyPDFLoader

---

📦 Installation

1. Clone the repository:

   git clone https://github.com/rajrounak21/Agentic-Rag-Agent.git
   cd Agentic-Agent-Rag

2. Create a virtual environment:

   python -m venv venv
   On Windows: venv\Scripts\activate

3. Install dependencies:

   pip install -r requirements.txt

4. Environment Setup: Create a .env file in the root directory and add your Google API Key:

   GOOGLE_API_KEY=your_gen_ai_api_key_here

---

🎮 How to Use

1. Run the App:

   streamlit run app.py

2. Upload Files: Open the sidebar and use the file uploaders to select your documents (PDF/TXT/DOCX) or spreadsheets (CSV/XLSX).

3. Sync Documents: Click the "🔄 Sync & Process Files" button. This will process the files and prepare the knowledge base.

4. Chat: Ask questions in the chat box or use the ⚡ Quick Action buttons to get immediate results!

---

📁 Directory Structure

├── app.py              # Main Streamlit application
├── uploads/            # Temporary directory for uploaded files
├── faiss_*/            # Local FAISS indices for processed files
├── .env                # API Key configuration
└── requirements.txt    # Project dependencies

---

📜 License

MIT License - feel free to use and modify for your own projects!