Unix/Linux

1. make sure you have the Software_Requirements libraries (Python, Numeric,PyVTK) installed.

2. Either:

   • Install the rpm file by running

     • "rpm -U ovf2vtk-0.x.x-x.noarch.rpm".

       This is the recommended option as it allows easy de-installation of the package (using "rpm -e ovf2vtk").

   Or

   • Install ovf2vtk from source:

     • untar the distribution file: "tar xfvz ovf2vtk-0.x.x.tar.gz"
     • cd into the new directory: "cd ovf2vtk"
     • install files: "python setup.py install"

     If you can't write to the site-packages directory on your machine (due to missing permissions), you can install the packages somewhere else using the --root switch (python setup.py install --root /home/myhome/local).

     If you do this, you need to let Python know that it should search for packages in that directory. This is achieved by setting the PYTHONPATH, for example:

     export PYTHONPATH=$PYTHONPATH:/home/myhome/local/usr/lib/python2.4/site-packages
     

     before you call the executable (which you should find in "/home/myhome/local/usr/bin/").

Windows

• If you don't have Python installed, the easiest way is to

  • download and install the Enthought Python distribution. This will do several things:

    • install Python
    • install numpy
    • add the python interpreter to the search path (usually in C:Python25) [This might only happen if you install Enthought Python as the administrator].
    • add the 'script directory' to the search path (usually in C:Python25Scripts) [This might only happen if you install Enthought Python as the administrator]
    • This will also install MayaVi which you might want to use to display the vtk files.

  • After you have installed Enthought Python, you should be able to open a Command Prompt and start the Python interpreter by typing "python". You should see something like this:

    C:>python
    Enthought Edition build 1057
    Python 2.3.3 (#51, FGeb 16 2004, ...)
    ...
    

    Press CTRL-Z and Return to exit this.

  • Install PyVTK

    This involves to download the source, unpack it, change directory into the source directory, execute "python.exe setup.py install".

  • Download the ovf2vtk windows installer and double click on it.

• If you have Python installed already, make sure you have numpy, PyVTK and run the Windows installer for ovf2vtk.

MacOS X

• Make sure you have python on the system (Fink can provide this.
• Install pyvtk (Fink provides this package)
• Install numpy (can use Fink)
• Install ovf2vtk from the source file (see installation instructions for Unix/Linux)

Enthought Python is now also available for Mac OS X, so in principle this can be used to provide python and numpy. It should be possible to install Pyvtk and ovf2vtk after having installed Enthought Python (but this is untested).

Uninstalling ovf2vtk

• Linux: if you have installed the rpm, just run "rpm -e ovf2vtk"
• Windows: go to add/remove software, select ovf2vtk
• If you have installed from source (all platforms):
  • remove the ovf2vtk directory below 'site-packages' of your python distribution
  • remove the 'ovf2vtk' and 'winovf2vtk.py' files from whereever they have been placed:
    • on windows usually c:Python25Scripts
    • on linux '/usr/local/bin', or '/usr/bin'
    • on Mac OS (with Fink): /sw/bin

Example

• Download this omf file and save it to a local directory

• Convert the data into a vtk file by issueing this command:

  ovf2vtk --add all cantedvortex.omf cantedvortex.vtk
  

  On windows, run:

  winovf2vtk.py --add all cantedvortex.omf cantedvortex.vtk
  

• This should produce a file with name "cantedvortex.vtk".

• This example file is included in the "Demo" subdirectory in the source file distribution.

How can I process more than one file?

OVF2VTK can only process one file at the time. However, the required functionality is usually provided by the operating system/tools.

Using the bash shell (i.e. on Linux, Mac OS X, Cygwin bash terminal on Windows), this for-loop:

for filename in *.omf; do ovf2vtk --add all $filename ${filename%.omf}.vtk; done

will go through all files with ending .omf in the current directory, and call ovf2vtk to convert them to vtk files. The slightly cryptic ${filename%.omf}.vtk will remove .omf from the filename that is stored in the variable $filename, and append .vtk.

OVF2VTK cannot read my ovf files

If you get the following error (this will be followed by more Traceback information, so look at the first few lines of the error message), you are likely to be trying to read the new OVF2.0 file format (instead of the pre-2010 ovf1.0 format):

$ ovf2vtk --add all sandwich-Oxs_TimeDriver-Magnetization-00-0001097.omf test.vtk
----------------------------------------------------------------------
ovf2vtk --- converting ovf files to vtk files
Hans Fangohr, Richard Boardman, University of Southampton
----------------------------------------------------------------------
The first item in a binary file is meant to be 123456789012345.0
but it is not. Instead, I read  31198.0488937 .
Cowardly stopping here.

In the long run, we will try to extend ovf2vtk so that it can read the new ovf2.0 file format. Until then, you can convert any ovf2.0 formatted data file (with name 'oldfile') into ovf1.0 formatted file (with name 'newfile') using this command:

oommf.tcl avf2ovf -format b8 oldfile newfile

It appears that one can also use the 'b4' switch rather than 'b8' to reduce diskspace usage (using only 4 byte per floating point number rather than 8).