Skip to content

GEMINI.md

Latest commit

 

History

History
84 lines (62 loc) · 3.09 KB

GEMINI.md

File metadata and controls

84 lines (62 loc) · 3.09 KB

GEMINI.md

Project Overview

This project, DeePTB, is a Python package that uses deep learning to accelerate ab initio electronic structure simulations. It provides a command-line interface to train, test, and run deep learning-based tight-binding models. The project is built using Python, PyTorch, and other scientific computing libraries. It uses uv for dependency management.

The project has two main components:

  1. DeePTB-SK: deep learning based local environment dependent Slater-Koster TB.
  2. DeePTB-E3: E3-equivariant neural networks for representing quantum operators.

Building and Running

Installation

The project uses uv for dependency management. To install the dependencies, you can follow the instructions in the README.md.

  1. Clone the repository:

    git clone https://github.com/deepmodeling/DeePTB.git
cd DeePTB

  2. Install UV (if not already installed):

    pip install uv

  3. Install all dependencies:

    uv sync

    This automatically creates a virtual environment and installs all dependencies including PyTorch and torch_scatter.

Running the application

The project provides a command-line interface through the dptb command. UV automatically activates the environment when using uv run:

  • uv run dptb config: Generate a config template.
  • uv run dptb bond: Bond distance analysis.
  • uv run dptb train: Train a model.
  • uv run dptb test: Test a model.
  • uv run dptb run: Run a TB model.
  • uv run dptb n2j: Convert NRL file to json.
  • uv run dptb p2j: Convert PTH ckpt to json.
  • dptb data: Preprocess data.
  • dptb cskf: Collect SK parameters from sk files.
  • dptb skf2nn: Convert sk files to nn-sk TB model.
  • dptb esk: Generate initial empirical SK parameters.

For more details on each command, you can use the -h or --help flag, for example uv run dptb train -h.

Testing

To run the tests, use the following command:

uv run pytest ./dptb/tests/

Development Conventions

  • The project uses uv for dependency management.
  • The code is structured into a main dptb package with subpackages for different functionalities like data, nn, negf, etc.
  • Entry points for the command-line interface are located in the dptb/entrypoints directory.
  • Tests are located in the dptb/tests directory.

Directory Structure

├───dptb/                  # Main source code
│   ├───__main__.py        # Main entry point for the CLI
│   ├───entrypoints/       # Subcommands for the CLI
│   ├───data/              # Data loading and processing
│   ├───nn/                # Neural network models
│   ├───negf/              # NEGF calculations
│   └───tests/             # Unit tests
├───docs/                  # Documentation
├───examples/              # Example usage
├───pyproject.toml         # Project metadata and dependencies
├───README.md              # Project overview
└───ut.sh                  # Test script