Reading a .vtk file with python

Question

I've been given a legacy format vtk file (I think its an unstructured grid) and I'd like to read it in with python and output a .npy file instead, since I know how to deal with that.

The file is a dump from ATHENA and so has density, velocity, magnetic field along with the coordinates.

I'm very much a procedural programmer, so all these objects are confusing...

Oh, from jterrace's link I see that it's a binary format. Bleh. — Steven Rumbalski
– Steven Rumbalski, Commented Jul 30, 2012 at 19:23
You may be able to read it with vtk.vtkPolyDataReader(), but writing to .npy, I have no clue. — Bernhard
– Bernhard, Commented Aug 8, 2012 at 18:58

Guillaume Jacquenot · Accepted Answer · 2013-05-13 13:42:14Z

22

Here is the solution that I came up with, the trick was turning on ReadAllVectorsOn().

import numpy
from vtk import vtkStructuredPointsReader
from vtk.util import numpy_support as VN

reader = vtkStructuredPointsReader()
reader.SetFileName(filename)
reader.ReadAllVectorsOn()
reader.ReadAllScalarsOn()
reader.Update()

data = reader.GetOutput()

dim = data.GetDimensions()
vec = list(dim)
vec = [i-1 for i in dim]
vec.append(3)

u = VN.vtk_to_numpy(data.GetCellData().GetArray('velocity'))
b = VN.vtk_to_numpy(data.GetCellData().GetArray('cell_centered_B'))

u = u.reshape(vec,order='F')
b = b.reshape(vec,order='F')

x = zeros(data.GetNumberOfPoints())
y = zeros(data.GetNumberOfPoints())
z = zeros(data.GetNumberOfPoints())

for i in range(data.GetNumberOfPoints()):
        x[i],y[i],z[i] = data.GetPoint(i)

x = x.reshape(dim,order='F')
y = y.reshape(dim,order='F')
z = z.reshape(dim,order='F')

edited May 13, 2013 at 13:42

Guillaume Jacquenot

11.8k6 gold badges45 silver badges50 bronze badges

answered Aug 13, 2012 at 21:15

Ben Jackel

5491 gold badge3 silver badges6 bronze badges

Sign up to request clarification or add additional context in comments.

Comments

Nico Schlömer · Accepted Answer · 2018-06-11 19:14:24Z

13

meshio (a project of mine) knows the VTK format, so you could simply

pip install meshio

and then

import meshio
mesh = meshio.read('file.vtk')
# mesh.points, mesh.cells, ...

edited Jun 11, 2018 at 19:14

answered May 18, 2017 at 17:52

Nico Schlömer

59.7k35 gold badges216 silver badges291 bronze badges

3 Comments

npm Over a year ago

can we save to .csv or .txt using meshio?

Nico Schlömer Over a year ago

csv and txt are no mesh formats, so no, not out of the box. You can always read the data in and just dump it though.

Dai Over a year ago

@npm The coordinates and field data can be readily determined from the mesh: coords, data = mesh.points, mesh.point_data['<variable_name>']. It is then easy to write these to file in whatever format you wish

jterrace · Accepted Answer · 2012-08-10 03:54:12Z

10

Here's a script that reads polygon data into numpy arrays from a VTK file using the VTK Python SDK:

import sys

import numpy
import vtk

reader = vtk.vtkPolyDataReader()
reader.SetFileName(sys.argv[1])
reader.Update()

polydata = reader.GetOutput()

for i in range(polydata.GetNumberOfCells()):
   pts = polydata.GetCell(i).GetPoints()    
   np_pts = numpy.array([pts.GetPoint(i) for i in range(pts.GetNumberOfPoints())])
   print np_pts

edited Aug 10, 2012 at 3:54

answered Aug 10, 2012 at 1:31

jterrace

67.5k24 gold badges164 silver badges208 bronze badges

4 Comments

Ben Jackel Over a year ago

I have been able to get to the GetOutput() stage using vtkDataSetReader() instead of vtkPolyDataReader, but I am still confused about how to get the information out. GetNumberOfPoints and GetNumberOfCells seems to give me sensible numbers for how big I think the arrays should be, but I still don't know how to pull all the variables out. Is there a way to get information on what exactly the vtk has in it, and in what form? In an understandable way?

jterrace Over a year ago

If you pick one of the files here people.sc.fsu.edu/~jburkardt/data/vtk/vtk.html I can try and help extract the format. There are so many different formats.

Ben Jackel Over a year ago

reader.IsFileStructuredPoints() returns 1, the other options return 0, so I'm going out on a limb and saying its a legacy vtk of Structured Points.

jterrace Over a year ago

You probably want to use vtkStructuredPointsReader then?

ToddP · Accepted Answer · 2018-09-14 09:39:09Z

6

Here's a short snippet for reading points from legacy VTK files:

import numpy as np
import vtk

filename = 'model.vtk'

reader = vtk.vtkGenericDataObjectReader()
reader.SetFileName(filename)
reader.Update()

points = np.array( reader.GetOutput().GetPoints().GetData() )

The variable points is an (N,2) or (N,3) array, where N is the number of points.

answered Sep 14, 2018 at 9:39

ToddP

74413 silver badges20 bronze badges

Comments

g.stevo · Accepted Answer · 2012-07-31 23:01:16Z

4

Have you tried using paraview? (http://www.paraview.org/) It can give you a visual idea of what is going on behind the scenes and can output the file in a number of different ways. I would suggest this as I don't have a clue what your data is like. http://www.vtk.org/Wiki/VTK/Examples/Python may also have an example that may fit the bill for you. Personally, I'd have a play with paraview and go from there.

answered Jul 31, 2012 at 23:01

g.stevo

7403 silver badges10 bronze badges

1 Comment

Ben Jackel Over a year ago

I've used paraview before, or rather an add on to it called VISIT. However, I need to analyse what's in the file and do things like fft etc. and so simply visualizing it is not enough.

Tim · Accepted Answer · 2013-08-15 00:54:32Z

4

It should be mentioned that in its latest release, the yt project http://yt-project.org/ includes support for ATHENA, meaning that by all means this is way to analyze the simulation data using python.

answered Aug 15, 2013 at 0:54

Tim

491 silver badge1 bronze badge

Collectives™ on Stack Overflow

Reading a .vtk file with python

6 Answers 6

Comments

3 Comments

4 Comments

Comments

1 Comment

Comments

Your Answer

Linked

Hot Network Questions

Collectives™ on Stack Overflow

6 Answers 6

Comments

3 Comments

4 Comments

Comments

1 Comment

Comments

Your Answer

Sign up or log in

Post as a guest

Linked

Related