Installation

There are currently three ways of installing MayaVi.

• As a stand-alone binary.
• Using RPMs, debs or Enthon
• From source as a Python module.

Installation as a binary

Currently there are two binaries available. One for Win32 and the other for Linux. The Linux and Win32 binaries were made using Gordon McMillan's Installer. The Windows installer was made using InnoSetup. Installing the binaries is pretty straight forward. The download is pretty big since it comes bundled with everything necessary to run MayaVi. Please note that when you install the binary, you cannot use MayaVi as a Python module from an existing installation of Python. If you want that, you must install MayaVi from the other installation methods.

Both the binary installations ship with a doc/ directory containing a quick installation guide, README, etc. and a user guide in doc/guide/book1.html. Also included is an examples/ directory that contains a sample VTK data file of a heart CT scan and a visualization based on that data.

Binary installation under Windows

Just download the binary, run it to start the installation, answer the questions on where you want to install it etc. That's it! If you have problems, please let us know.

Binary installation under Linux

Installing the Linux binary is also fairly straightforward. The Linux binary is simply a bzipped tarball. Untar it and run the mayavi program in the untarred directory. The binary requires the following:

• glibc2.3.2. This should be standard on most recent Linux installations.
• Mesa/OpenGL: You will need to have libGL.so.1, libGLU.so.1 and libGLcore.so.1 installed.
• X11: You will need to have the X libraries installed. The binaries were built with XFree86 4.3. I guess XFree86 4.x should work okay.

Using RPMs, debs or Enthon

RPM

Installing the RPM requires that you have installed Python, Tkinter and the VTK RPMs for VTK and VTK-Python. Once you have these working you can simply install the appropriate (depending on your Python version) MayaVi RPM. If you have installed VTK from the sources and not via RPMS there should be no problem with installing the MayaVi RPM.

Debian package

To install the Debian package simply install the current Debian MayaVi package (using apt-get install mayavi). This will rope in all necessary dependencies. Then simply install the downloaded Debian package. Debian (woody) ships with VTK 3.2. More recent VTK 4.x binaries (not debs) that work under testing (for the i386) are available here. These binaries are built under testing.

Enthought Python for Windows

Enthought distributes an enhanced Python installer that ships with a huge number of very useful Python packages. If you install this, you get a fully functional Python installation with MayaVi installed. Since all MayaVi's requirements are installed as part of Enthon, if you need to upgrade the version of MayaVi that ships with Enthon, you can simply install MayaVi from sources by following the instructions in the installation of MayaVi section.

Installation from Sources

MayaVi uses distutils for the installation process. Distutils is part of the Python standard library from versions 2.0 onwards. So, you should normally have this available. MayaVi requires that VTK be installed and usable from Python. For more details please read the MayaVi requirements.

If you would like to install MayaVi from the sources under Debian GNU/Linux you may simply run

Installation of MayaVi

If the requirements are all met, installing MayaVi is simple. Download the tarball, untar/unzip it, change directory to the new directory and do the following:

      % python setup.py install
    

      % python setup.py install --help
    

Please note that if you simply want to run MayaVi without installing it as a module you can always run it straight from the untarred directory like so:

      % mayavi [options]
    

Once you have MayaVi installed and running, you can test it by using the example files in the examples/ directory of the sources. The directory contains a sample VTK data file of a heart CT scan and a visualization based on that data. To run this visualization use the File->Load->Visualization menu and choose the example in the examples directory of the MayaVi source tree. If this loads fine then you are all set! For more information on how to use MayaVi read the MayaVi users guide. This guide is also included along with the source and is in the doc/guide/ directory.

Requirements

• MayaVi requires:
  • Python version >= 2.0 and a working Tkinter.
  • A working distutils.
  • The VTK library with the Python bindings, the VTK-Python Tk widget, and the VTK contrib library compiled. Under VTK 4.x you also need to compile the Hybrid classes (there is no 'contrib' directory in VTK 4.x) to get all the MayaVi modules to work.
• MayaVi runs fine under Linux and the Windows (TM) operating system. There are reports that it works on other Unix systems too. Technically, it should have no problems running on any platform where VTK and Python are available.
• Both the latest stable releases and the nightly VTK 4.x CVS tree builds work fine.

Usually the hardest part of a MayaVi install when using the sources is getting the VTK Python bindings installed properly. The next section tries to address this.

Installation of VTK

Before you can run MayaVi you need to have VTK installed and working properly. There are two ways to install VTK on any platform. One way is to build the packages from the sources and the other is to download pre-built binary packages. VTK is a huge package and building from source is not for the weak hearted. You can download binaries and sources from the VTK download page.

Instructions on how to build VTK suitably under Windows for MayaVi are available here. You should also read the README.html that is included with the VTK sources. When building VTK from the sources make sure that you build the Python bindings. It is a good idea to build the Tcl bindings also since that ensures that the VTK-Python Tk widgets are also compiled.

Pre-built VTK packages are available for Linux and Win32 and these are mentioned below. If installing VTK from sources is a hassle you should consider trying these instead.

Linux binaries

Links to nightly RPMS are available from the VTK download page. A tarball containing all the necessary libraries/modules in order to use VTK from Python is available here. This build uses the latest CVS tree and has been built on the Debian GNU/Linux woody distribution. More details on the installation etc. are available in the README.txt included at the site.

Binaries for Windows

Windows binaries are also available at the VTK download page. Some old windows binaries for VTK-Python should be available from here. This package has all the necessary items for MayaVi compiled in. Instructions on how to install it are available at the site. You might also be interested in checking out Enthought's Enhanced Python distribution. MayaVi should be easy to install from source if you are using this.

The Linux and Win32 binaries pages should be updated reasonably often so do check them out.

Testing your VTK installation

Included along with the MayaVi source is a file called test_vtk.py in the doc directory. Run this to test your VTK installation like so:

      % python test_vtk.py
      

If this runs without problems then your VTK installation is ok. You should then install MayaVi from the source and try to run it. If you have trouble the script will tell you where it failed. At this point if you are unable to proceed email us. You might also want to check the MayaVi Wiki.