Easy Way

There is now an installation script for OSX that handles the installation of Python, all the Python dependencies, and HDF5. You can get it by obtaining it from SVN:

svn export http://svn.enzotools.org/yt/trunk/doc/install_script_osx.sh

At this point you should read it so that you can see where it will install things. By default, it will create a new directory at the current working directory, inside which it will put all the source files. However, while yt itself will be installed there, the Python dependencies will be installed system-wide. This script is still experimental but it has been tested on both PPC and i386 machines running OSX 10.5. If you'd like to use it, and you're okay with where it will put things, go ahead and execute:

bash install_script_osx.sh

Please let me know if it does or does not work for you!

EPD

If you're looking for a more complete installation, with many more packages, you can use the Enthought Python Distribution. Get the  Enthought Python Distribution (free for educational users), install it, and then install YT from subversion.

Once you have done that, you should be able to run the commands reason, yt and iyt, which will be located in the directory /Library/Frameworks/Python.framework/Versions/Current/bin/.

OS X Installation Instructions

OS X has some good points and some bad points to installation. Some of the packages are harder to get, but fortunately that's made up for by some binary installers. You'll need to have HDF5 already installed, and I will not be covering that here. This has been tested on PPC and i386 machines using OSX 10.5. I suspect it will work with 10.4, but almost certainly not with 10.3.

  1. First install Python2.5.
  2. You must now update your terminal path as Python instructs you to do so; typically this means prepending your PATH with /Library/Frameworks/Python/Versions/Current/bin/ .
  3. Next up is NumPy. Download and install the binary package.
  4. Now that we have NumPy, download and install wxPython.
  5. We can now install our plotting library, Matplotlib. Download and install Matplotlib.
  6. This step is optional, but get and install  h5py -- it should be as simple as
    sudo python2.5 setup.py install --hdf5=/path/to/hdf5
    
    Instead of the above, you can set the following environment variables:
    HDF5_DIR=path to hdf5
    HDF5_API=16
    Then do:
    sudo easy_install h5py
    
    yt will serialize some types of data about the hierarchy and projections between instances if it has access to PyTables?.
  7. Alright, we're almost done. Now, just install yt (for more info, see the installation instructions). This should follow something like:
    svn co http://svn.enzotools.org/yt/trunk/ ./yt-trunk/
    cd yt-trunk
    sudo python2.5 ez_setup.py
    echo /path/to/hdf5 > hdf5.cfg
    sudo python2.5 setup.py develop
    

yt is now installed to that directory, yt-trunk. Any changes you make to the source code in that directory will be propagated across the system -- this makes development much easier!

Once completed, you should be able to run the commands reason, yt and iyt, which will be located in the directory /Library/Frameworks/Python.framework/Versions/Current/bin/.