Python EVTK Library

1.6.0 · active · verified Fri Apr 17

pyevtk is a Python library for exporting numerical data (like structured grids, unstructured grids, and points) into binary VTK (Visualization Toolkit) files, enabling visualization in tools like ParaView or VisIt. As of version 1.6.0, it supports various data types and has an active, albeit slow, release cadence.

Common errors

```
ModuleNotFoundError: No module named 'evtk'
```
cause Attempting to import from the deprecated `evtk` package, which might not be installed or available as an alias in the current environment/version.

fix Change all imports from `evtk` to `pyevtk`, e.g., `from pyevtk.hl import gridToVTK`.
```
TypeError: Expected a numpy array for 'data' but got <class 'list'>
```
cause Input data (e.g., field values, coordinate arrays) was provided as a Python list or non-NumPy array.

fix Convert the input data to a NumPy array before passing it to `pyevtk` functions, e.g., `import numpy as np; data = np.asarray(my_list)`.
```
ValueError: Data dimensions do not match grid dimensions.
```
cause The shape of the data array (e.g., `cellData` or `pointData`) does not correspond to the dimensions of the grid (e.g., `(nx, ny, nz)` for structured grids).

fix Carefully check the expected dimensions for the specific VTK export function being used. For cell data on an `(nx, ny, nz)` grid, the data array should typically have shape `(nx-1, ny-1, nz-1)`. For point data, it should be `(nx, ny, nz)`.

Warnings

deprecated Using `from evtk.hl import ...` is deprecated. The `evtk` package is an alias for `pyevtk` for backward compatibility but will issue a `DeprecationWarning` and may be removed in future versions.
Fix: Always use `from pyevtk.hl import ...` for all imports, e.g., `from pyevtk.hl import gridToVTK`.
gotcha All input data arrays (e.g., `data`, `x`, `y`, `z` coordinates) must be NumPy arrays. Passing standard Python lists or other array-like objects will result in `TypeError` or unexpected behavior.
Fix: Ensure all array inputs are `numpy.ndarray` objects. Convert them using `numpy.asarray()` if necessary.
gotcha VTK files generated by `pyevtk` often assume cell-centered data for structured grids. Misinterpreting this can lead to data being shifted by half a cell when visualized in tools like ParaView if your data is node-centered.
Fix: Be mindful of whether your data is node-centered or cell-centered. If your data is node-centered on an `(Nx, Ny, Nz)` grid, the data array passed to `gridToVTK` for cell data should typically have dimensions `(Nx-1, Ny-1, Nz-1)` to align correctly. Alternatively, explicitly define point data if appropriate.

Install

pip install pyevtk Install stable version

Imports

gridToVTK
wrong:
```
from evtk.hl import gridToVTK
```
correct:
```
from pyevtk.hl import gridToVTK
```
The `evtk` package is an alias for `pyevtk` for backward compatibility but is deprecated and will issue a warning. Always use `pyevtk`.
pointsToVTK
```
from pyevtk.hl import pointsToVTK
```
imageToVTK
```
from pyevtk.hl import imageToVTK
```

Quickstart

This quickstart demonstrates how to export a structured grid with both cell-centered and point-centered data to a VTK Structured Grid (.vts) file using `gridToVTK`.

import numpy as np
from pyevtk.hl import gridToVTK

# Define grid dimensions
nx, ny, nz = 64, 64, 64

# Coordinates
x = np.arange(0, nx, 1, dtype='float64')
y = np.arange(0, ny, 1, dtype='float64')
z = np.arange(0, nz, 1, dtype='float64')

# Cell data (e.g., temperature)
temp = np.random.rand(nx - 1, ny - 1, nz - 1)

# Point data (e.g., pressure)
press = np.random.rand(nx, ny, nz)

# Export to VTK file
gridToVTK(
    "output_grid",
    x, y, z,
    cellData={'temperature': temp},
    pointData={'pressure': press}
)

print("VTK file 'output_grid.vts' generated successfully.")

view raw JSON →