Skip to content

AnguseZhang/MatMaster

Repository files navigation

MatMaster Agent Project

A multi-agent based platform for materials science computation and data analysis, integrating various computational tools and database interfaces to provide intelligent solutions for materials research.

Project Structure

.
├── README.md # Project documentation
├── __init__.py # Python package initialization
├── agents/ # Agent modules directory
│ ├── matmaster_agent/ # Main agent module
│ │ ├── base_agents/ # Base agent classes
│ │ ├── ABACUS_agent/ # ABACUS calculation agent
│ │ ├── DPACalculator_agent/ # DPA calculation agent
│ │ ├── HEACalculator_agent/ # HEA calculation agent
│ │ ├── HEA_assistant_agent/ # HEA assistant agent
│ │ ├── INVAR_agent/ # INVAR materials agent
│ │ ├── MrDice_agent/ # Materials database agent
│ │ ├── chembrain_agent/ # ChemBrain analysis agent
│ │ ├── ssebrain_agent/ # Solid-state electrolyte agent
│ │ └── ... # Other specialized domain agents
│ └── __init__.py
├── evaluate/ # Evaluation module
│ ├── experiments/ # Experimental evaluations
│ ├── metric/ # Evaluation metrics
│ └── base/ # Evaluation base classes
├── pyproject.toml # Python project configuration
└── uv.lock # Dependency lock file

Features

Computational Tools Integration

  • ABACUS: First-principles calculations
  • DPA Calculator: Deep potential calculations
  • HEA Calculator: High-entropy alloy calculations

Specialized Domain Agents

  • Chemical Materials Analysis: Molecular structure, reaction path analysis
  • Solid-State Electrolyte Research: Ion conductor material screening and analysis
  • Thermoelectric Materials: Thermoelectric property calculation and optimization
  • Superconducting Materials: Superconducting characteristic analysis and prediction
  • Perovskite Materials: Photovoltaic material research and design

Database Interfaces

  • MrDice: Materials database query

Auxiliary Tools

  • Structure generation and optimization
  • Trajectory analysis
  • Document parsing
  • Reaction path planning

Installation & Usage

Requirements

  • Python 3.12+
  • google-adk
  • uv

Installation Steps

# Clone the project
git clone <repository-url>
cd MatMaster/

# Install dependencies
uv sync

Configuration

The project uses pyproject.toml for dependency management, including:

  • Model API key settings

  • Computational resource allocation

  • Database connection configuration

  • Log level settings

Evaluation System

The project includes a comprehensive evaluation module supporting:

  • Single-turn dialogue evaluation

  • Multi-turn dialogue evaluation

  • Task transfer quality assessment

  • Human simulation evaluation

Contributing

Welcome to submit Issues and Pull Requests to help improve the project.

Support

Please submit Issues or contact the development team if you have any questions.