🧠 ANARIX AI Agent (FastAPI + Gemini + MySQL)

---

📘 Project Overview

This project focuses on building an AI agent designed to answer e-commerce data questions. The agent is capable of interpreting natural language questions, generating and executing SQL queries on provided datasets, and responding with accurate answers.

---

🚀 Features

This ANARIX AI Agent provides a powerful and intuitive way to interact with your e-commerce data:

• Natural Language Querying: The AI agent can understand and answer any question related to the provided e-commerce data.

• API-Driven Interaction: Questions are received via API endpoints, and the agent processes them to return accurate responses.

• SQL Conversion & Execution: The agent includes logic to convert natural language questions into SQL queries, fetch answers from the database, and return them in a human-readable format.

---

🧠 How It Works

1. User Input — Ask a question at the /ask API endpoint.
2. Gemini Processing — Generates a SQL query or chart instruction.
3. SQL Execution — FastAPI runs the SQL on the connected MySQL database.
4. Chart Generation — If requested, Matplotlib generates a chart and saves it as an image file in assets/screenshots/.
   The API response includes the path to the chart image (e.g., "chart_path": "assets/screenshots/chart.png"), which can be viewed or served via your web server.
5. Response — The agent returns the answer, SQL query, query result, and chart path (if applicable).

---

📦 Project Structure

📂ANARIX-AI-AGENT
├── 📂pycache
├── 📂assets
│ └── 📂screenshots (has imgs)
├── 📂datasets
│ └── 📄(Excel files or raw datasets)
├── 📂db
│ └── 📄(MySQL config or helpers)
├── 📂env (excluded by .gitignore)
├── 📂llm
│ └── 📄agent.py (Gemini + SQL + chart logic)
├── 📄.env (holds GEMINI_API_KEY)
├── 📄.gitignore (ignores env(virtual environment)/, pycache/, .env, api_check.py, cmds.txt.)
├── 📄main.py (FastAPI entry point with /ask endpoint)
├── 📄requirements.txt (all required pip packages)
├── 📄TASK DESCRIPTION.md (provided use-case or brief)
└── 📄README.md _(project documentation you're reading)_

---

1. Create a Virtual Environment

python -m venv env

source env/Scripts/activate   

2. Install Dependencies

pip install -r requirements.txt

3. Add Your Gemini API Key

Create a .env file:

GEMINI_API_KEY=your_gemini_api_key_here

Get your key from: https://makersuite.google.com/app/apikey

4. Configure MySQL DB

Ensure MySQL is running, and update the DB connection in db/session.py if needed:

DATABASE_URL = f"mysql+mysqlconnector://{DB_USER}:{DB_PASSWORD}@{DB_HOST}:{DB_PORT}/{DB_NAME}"

Make sure your database and tables are already created and populated.

5. Run the API Server

uvicorn main:app --reload

Access Swagger UI at: http://127.0.0.1:8000/docs

💬 Example Questions

📊 Data Queries "What is the total ad spend?"

"How many clicks did item 101 get?"

"Show me all metrics for item 21"

📈 Business Metrics "What is the Return on Ad Spend (RoAS)?"

"Which item had the highest CPC?"

📊 Chart Requests "Generate a pie chart of sales by item"

"Show a bar chart of revenue per product"

🛠️ Data Updates "Delete entry where item_id is 99"

---

📦 Requirements

Python 3.8+

FastAPI

SQLAlchemy

pandas

matplotlib

python-dotenv

sse-starlette

MySQL (local or hosted)

Google Generative AI client (google-generativeai)

---

🧠 How It Works

User Input — Ask a question at the /ask API endpoint.

Gemini Processing — Generates a SQL query or chart instruction.

SQL Execution — FastAPI runs the SQL on the connected MySQL database.

Chart (Optional) — If requested, Matplotlib opens a popup with a chart.

---

📌 Notes

.env is excluded by .gitignore for security

Matplotlib is used to pop up chart windows (no frontend)

Error-handling and Gemini fallback are supported

---

📊 Chart Output

• Chart images are saved automatically in assets/screenshots/ whenever a chart is requested.
• The API response includes the chart image path for easy access and integration.
• Example response:

  {
    "question": "Show me a bar chart of total sales by product",
    "sql_generated": "SELECT product, SUM(total_sales) FROM total_sales_metrics GROUP BY product",
    "query_result": [...],
    "answer": "Query executed successfully with chart",
    "chart_path": "assets/screenshots/chart.png",
    "type": "select"
  }

---

📷 Screenshots

DB & Tables in MySQL

Swagger UI(/docs)

Chart Output Example (Saved Image)

Terminal Output (Generated by Gemini API)

MySQL Output

---

🙌 Credits

FastAPI

Google Gemini

SQLAlchemy

Matplotlib

Pandas

---

🙋‍♂️ Author

Nitish B
Final Year Student | Software Developer & AI Enthusiast
📫 GitHub • LinkedIn

---

⭐️ Star the Repository

If you found this project helpful, consider starring ⭐ it on GitHub to support future development!