Deep Research Tutorial

A comprehensive AI-powered chemistry research system that combines database querying, literature analysis, and automated report generation. This project demonstrates how to build a multi-agent system for scientific research using Google's Agent Development Kit (ADK).

🚀 Features

Core Capabilities

• Multi-Agent Architecture: Hierarchical agent system with specialized sub-agents
• Database Integration: SQLite-based local database with CSV data loading
• Literature Analysis: Automated paper reading and content extraction
• Report Generation: AI-powered scientific report creation
• RESTful API: Flask-based API for database operations

Agent System

• Chemistry Research Agent: Main orchestrator agent that coordinates research tasks
• Database Agent: Handles structured data queries and fact retrieval
• Deep Research Agent: Performs literature reviews and generates reports
• Paper Agent: Processes individual research papers
• Report Agent: Synthesizes findings into comprehensive reports

📁 Project Structure

deep_research_tutorial/
├── agent/                          # Multi-agent system
│   ├── agent.py                    # Main chemistry research agent
│   ├── callbacks.py                # Agent callback functions
│   ├── llm_config.py              # LLM configuration
│   ├── database_agent/            # Database query agent
│   ├── deep_research_agent/       # Literature research agent
│   │   ├── agent.py               # Deep research orchestrator
│   │   ├── paper_agent/           # Individual paper processor
│   │   └── report_agent/          # Report generation agent
│   └── tools/                     # Agent tools and utilities
├── database_server/               # Database and API services
│   ├── service.py                 # Flask API and database service
│   ├── polymer.csv               # Polymer data
│   ├── paper_metadata.csv        # Paper metadata
│   └── paper_text.csv            # Paper content data
├── my_data.db                    # SQLite database file
├── pyproject.toml                # Project dependencies
└── README.md                     # This file

🛠️ Installation

Prerequisites

• Python 3.12 or higher
• UV package manager (recommended)

Setup

1. Clone the repository

   git clone <repository-url>
   cd deep_research_tutorial

2. Install dependencies

   uv sync

3. Set up environment variables

   # Create a .env file with your API keys
   echo "OPENAI_API_KEY=your_openai_key_here" > .env
   echo "GOOGLE_API_KEY=your_google_key_here" > .env
   echo "DEEPSEEK_API_KEY=your_deepseek_key_here" > .env

🚀 Quick Start

1. Start the Database Server

python database_server/service.py

This will:

• Load CSV data into SQLite database
• Start Flask API server on http://localhost:5000
• Make database accessible via REST API

2. Use the Agent System in Google ADK Web UI

uv run adk web

🤖 Agent Workflow

1. User Query Processing

The main chemistry research agent:

• Analyzes user intent
• Creates a step-by-step research plan
• Proposes actions one at a time

2. Database Querying

The database agent:

• Searches structured data for facts
• Retrieves relevant paper metadata
• Returns formatted results

3. Literature Research

The deep research agent:

• Processes relevant papers in parallel
• Extracts key information
• Synthesizes findings

4. Report Generation

The report agent:

• Compiles all findings
• Generates comprehensive reports
• Formats output for user consumption

🔍 Example Use Cases

Chemical Property Queries

"What is the melting point of paracetamol?"
"Find solvents for recrystallizing benzoic acid"
"What are the properties of polyimide polymers?"

Literature Research

"Recent advancements in asymmetric catalysis for ibuprofen synthesis"
"Biocompatibility studies of PLA polymers"
"Novel methods for polymer characterization"

🛠️ Configuration

LLM Configuration

Edit agent/llm_config.py to customize:

• Model selection (GPT-4, Claude, etc.)
• API endpoints
• Temperature and other parameters

Database Configuration

Modify database_server/service.py to:

• Change database file path
• Add new data sources
• Customize API endpoints

📝 Dependencies

Key dependencies include:

• google-adk>=1.5.0 - Google Agent Development Kit
• litellm>=1.73.1 - LLM abstraction layer
• openai>=1.91.0 - OpenAI API client
• pandas>=2.3.0 - Data manipulation
• flask>=3.1.1 - Web framework
• pydantic>=2.11.7 - Data validation
• requests>=2.32.4 - HTTP client

🤝 Contributing

1. Fork the repository
2. Create a feature branch
3. Make your changes
4. Add tests if applicable
5. Submit a pull request

📄 License

This project is licensed under the MIT License - see the LICENSE file for details.

🙏 Acknowledgments

• Google Agent Development Kit (ADK) team
• OpenAI for LLM APIs
• The chemistry research community at DP Technology

📞 Support

For questions and support:

• Open an issue on GitHub
• Check the documentation
• Review the example usage

---

Note: This is a tutorial project demonstrating AI-powered research workflows. For production use, ensure proper security, validation, and error handling.