Metadata-Version: 2.1
Name: agat
Version: 8.0.3
Summary: Atomic Graph ATtention networks for predicting atomic energies and forces.
Home-page: https://github.com/jzhang-github/AGAT
Author: ZHANG Jun; ZHAO Shijun
Author-email: j.zhang@my.cityu.edu.hk
License: GPL
Platform: all
Classifier: Development Status :: 3 - Alpha
Classifier: License :: OSI Approved :: GNU General Public License (GPL)
Classifier: Intended Audience :: Developers
Classifier: Operating System :: OS Independent
Classifier: Natural Language :: English
Classifier: Programming Language :: Python
Classifier: Programming Language :: Python :: 3.10
Classifier: Topic :: Software Development :: Libraries
Requires-Python: >=3.10
Description-Content-Type: text/markdown
License-File: LICENSE
Requires-Dist: numpy
Requires-Dist: ase
Requires-Dist: tqdm


# AGAT (Atomic Graph ATtention networks)
[![GitHub](https://img.shields.io/github/license/jzhang-github/AGAT)](https://github.com/jzhang-github/AGAT/blob/main/LICENSE)
[![Pypi](https://img.shields.io/pypi/v/agat.svg)](https://pypi.org/project/agat/)
[![PyPI - Downloads](https://img.shields.io/pypi/dm/agat)](https://pypi.org/project/agat/)
[![PyPI - Wheel](https://img.shields.io/pypi/wheel/agat)](https://pypi.org/project/agat/)
[![Documentation Status](https://readthedocs.org/projects/agat/badge/?version=latest)](https://jzhang-github.github.io/AGAT/)
 
 ### The [PyTorch](https://pytorch.org/) backend AGAT is available now, try with `pip install agat==8.*`. For previous version, install with `pip install agat==7.*`.
  <br>  <br>  ![Model architecture](files/architecture.svg)

# Using AGAT
The [documentation](https://jzhang-github.github.io/AGAT/) of AGAT API is available.

# Installation

### Install with [conda](https://conda.io/projects/conda/en/latest/user-guide/install/index.html) environment
- Create a new environment   
  ```console
  conda create -n agat python==3.10
  ```

- Activate the environment  
  ```console
  conda activate agat
  ```

- Install [PyTorch](https://pytorch.org/),   
  Navigate to the [installation page](https://pytorch.org/get-started/locally/#start-locally) and choose you platform.
  For example (GPU):
  ```console
  conda install pytorch torchvision torchaudio pytorch-cuda=11.8 -c pytorch -c nvidia
  ```

- Install [dgl](https://www.dgl.ai/).   
Please navigate to the [Get Started](https://www.dgl.ai/pages/start.html) page of [dgl](https://www.dgl.ai/). 
For example (GPU):  
  ```console
  conda install -c dglteam/label/cu118 dgl
  ```

- Install AGAT package  
  ```console
  pip install agat
  ```

- Install CUDA and CUDNN [**Optional**].
	- For HPC, you may load CUDA by checking `module av`, or you can contact your administrator for help.
	- [CUDA Toolkit](https://developer.nvidia.com/cuda-downloads)
	- [cuDNN](https://developer.nvidia.com/cudnn)

# Quick start
### Prepare VASP calculations
Run VASP calculations at this step.

### Collect paths of VASP calculations
- We provided examples of VASP outputs at [VASP_calculations_example](https://github.com/jzhang-github/AGAT/tree/v1.0.0/files/VASP_calculations_example).   
- Find all directories containing `OUTCAR` file:   
  ```
  find . -name OUTCAR > paths.log
  ```    
- Remove the string 'OUTCAR' in `paths.log`.   
  ```
  sed -i 's/OUTCAR$//g' paths.log
  ```   
- Specify the absolute paths in `paths.log`.   
  ```
  sed -i "s#^.#${PWD}#g" paths.log
  ``` 

### Build database
```python
from agat.data import BuildDatabase
if __name__ == '__main__':
    database = BuildDatabase(mode_of_NN='ase_dist', num_of_cores=16)
    database.build()
```

### Train AGAT model
```python
from agat.model import Fit
f = Fit()
f.fit()
```

### Application (geometry optimization)
```python
from ase.optimize import BFGS
from agat.app import AgatCalculator
from ase.io import read
from ase import Atoms

model_save_dir = 'agat_model'
graph_build_scheme_dir = 'dataset'

atoms = read('POSCAR')
calculator=AgatCalculator(model_save_dir,
                          graph_build_scheme_dir)
atoms = Atoms(atoms, calculator=calculator)
dyn = BFGS(atoms, trajectory='test.traj')
dyn.run(fmax=0.05)
```
### Application (high-throughput prediction)
```python
from agat.app.cata import HpAds

model_save_dir = 'agat_model'
graph_build_scheme_dir = 'dataset'
formula='NiCoFePdPt'

ha = HpAds(model_save_dir=model_save_dir, graph_build_scheme_dir=graph_build_scheme_dir)
ha.run(formula=formula)
```

### For more custom manipulations, see our [documentation](https://jzhang-github.github.io/AGAT/) page.


[![Documentation Status](https://readthedocs.org/projects/agat/badge/?version=latest)](https://jzhang-github.github.io/AGAT/)



### Some default parameters
[agat/default_parameters.py](agat/default_parameters.py); Explanations: [docs/sphinx/source/Default parameters.md](https://github.com/jzhang-github/AGAT/blob/main/docs/sphinx/source/Default%20parameters.md).


# Change log  
Please check [Change_log.md](https://github.com/jzhang-github/AGAT/blob/main/Change_log.md)
