py4vasp¶

Installation¶

While this is not required to be able to run py4vasp, you may want to consider creating a separate environment for installation to avoid interference with other installed packages. [1] You can then install py4vasp from PyPI using the pip package installer

pip install py4vasp

This will automatically download py4vasp as well as most of the required dependencies. However, we do not install the mdtraj dependency by default because it does not reliably work with pip. We recommend to install mdtraj using conda

conda install -c conda-forge mdtraj

if you need the features that depend on it (plotting a trajectory of structures).

For a minimalistic setup where you use py4vasp as a library, you can install the core package

pip install py4vasp-core

The core package contains the same source code as the main package and does not impact the usage. However, it does not install any of the dependencies of py4vasp except for numpy and h5py. Hence, this core package is most suitable for script developers that do not need all the visualization features of py4vasp.

Alternatively, you can obtain the code from GitHub and install it. This will give you the most recent version with all bugfixes. However, some features may only work once the next VASP version is released.

git clone https://github.com/vasp-dev/py4vasp.git
cd py4vasp
pip install .
conda install -c conda-forge mdtraj

If these commands succeed, you should be able to use py4vasp. You can make a quick test of your installation running the following command

python -c "import py4vasp; print(py4vasp.__version__)"

This should print the version of py4vasp that you installed.

Quick start¶

The user interface of py4vasp is optimized for usage inside a Jupyter environment (Jupyter notebook or Jupyter lab), though it can be used in regular Python scripts as well. To give you an illustrative example of what py4vasp can do, we assume that you created a Jupyter notebook inside the directory of your VASP calculation. In the VASP calculation, you computed the density of states (DOS) with orbital projections (LORBIT = 11). You may now want to read the data from your VASP calculation to post-process it further with a script. This can be achieved with

>>> import py4vasp
>>> dos = py4vasp.calculation.dos.read()

Under the hood, this will access the vaspout.h5 file, because py4vasp knows where the output is stored after you ran VASP. It will read the relevant tags from the file and store them all in a Python dictionary. If you want to access particular orbital projections, let’s say the p orbitals, you can pass a selection = "p" as an argument to the routine. More generally, you can check how to use a function with

>>> help(py4vasp.calculation.dos.read)

The most common use case for the DOS data may be to prepare a plot to get some insight into the system of interest. Because of this, we provide an easy wrapper for this particular functionality

>>> py4vasp.calculation.dos.plot()

This will return an interactive figure that you can use to investigate the DOS. The plot command takes the same arguments as the read command. Note that this requires a browser to work; if you execute this from within a interactive environment, it may open a browser for you or you can enforce it by appending .show()

>>> py4vasp.calculation.dos.plot().show()

The interface for the other quantities is very similar. Every quantity provides a read function to get the raw data into Python and where it makes sense a plot function visualizes the data. However, note that in particular, all data visualized inside the structure require a Jupyter notebook to work. All plots can be converted to csv files to_csv of pandas dataframes to_frame for further refinement.

If your calculation is not in the root directory, you can create your own instance

>>> from py4vasp import Calculation
>>> calc = Calculation.from_path("/path/to/your/VASP/calcualtion")

The attributes of the calculation correspond to different physical quantities that you could have calculated with VASP. If you have an interactive session you can type calc. and then hit Tab to get a list of all possible quantities. However only the ones that you computed with VASP will give you any meaningful result.

If you want to experience more features of py4vasp, we highly recommend taking a look at the tutorials for VASP. Many of them use py4vasp to plot or analyze the data produced by VASP, so this may give you an excellent starting point to learn how you can apply py4vasp in your research.

Packages

Classes