VisTrails Home

Building From Source

From VisTrailsWiki

(Difference between revisions)
Jump to: navigation, search
m
Line 1: Line 1:
'''Requirements'''
'''Requirements'''
-
Base requirements:
+
 
* Python 2.4 or 2.5 http://www.python.org
* Python 2.4 or 2.5 http://www.python.org
* Graphviz http://www.graphviz.org
* Graphviz http://www.graphviz.org
Line 13: Line 13:
* MySQLdb http://sourceforge.net/projects/mysql-python (execution logging)
* MySQLdb http://sourceforge.net/projects/mysql-python (execution logging)
* VTK http://www.vtk.org (vtk)
* VTK http://www.vtk.org (vtk)
 +
* numpy http://www.numpy.org (scipy)
* scipy http://www.scipy.org (scipy)
* scipy http://www.scipy.org (scipy)
* matplotlib http://matplotlib.sourceforge.net/ (pylab)
* matplotlib http://matplotlib.sourceforge.net/ (pylab)

Revision as of 09:24, 30 August 2007

Requirements

For running various not-required packages and features (packages are listed last, in parentheses):

Instructions

Click here for Specific Instructions for Mac OS X

  • Install Python
    • On most Linux distributions it comes by default.
    • On Windows: Get the msi release here
    • Add the installation directory to the PATH variable
  • Install Python Extensions
    • PyXml is needed for web services and is available here. There's a binary package for Mac here
      • To install it from source, download the .tar.gz file and uncompress it
      • Then inside the uncompressed folder, type: python setup.py build and later sudo python setup.py install
    • fpconst and SOAPpy are also required for using web services. fpconst is available here. SOAPpy is available here
    • For WINDOWS & LINUX: we need ctypes available here
  • Install VTK CVS version (if you want to use VTK)
    • First you'll need CMake: get the latest release here
    • Then you can get vtk: follow the instructions to download the CVS source code on this page
    • (Skip this step if you don't need to run CORIE workflows) Follow these instructions to add CORIE and other needed files to VTK.
    • To compile VTK follow the instructions in the file README.html in the root of VTK directory tree
    • On configuring CMake, make sure that the variables VTK_WRAP_PYTHON, VTK_USE_GUISUPPORT, VTK_USE_DISPLAY are ON and VTK_USE_QVTK is OFF
    • It is recommended to create a environment variable called PYTHONPATH
      • PYTHONPATH=<path/to/vtk_build>/Wrapping/Python:<path/to/vtk_build>/bin:.
  • Install Qt 4.2.0 or above (Qt 4.2.3 recommended)
    • You can get the open source version for all platforms here
    • When installing on windows, Qt requires the mingw compiler. It will also offer to download and install automatically for you
    • After installing you need to build the Qt library:
    • (Skip this step if you're not a windows user) Run the Qt 4.2.x Command Prompt located in the Qt Program Folder (it will set all the environment variables to build the library for you).
    • Run configure then make
  • Install SIP
    • Get the stable release version available here. (The latest snapshot version is here)
    • Uncompress the file, and inside the directory, type python configure.py. On WIndows type python configure.py -p win32-g++. On 64-bit Linux type python configure.py -d /usr/lib64/python2.4/site-packages/. Then make and make install
  • Install PyQt4
    • Get the stable version, available here. (The latest snapshot version is here)
    • Uncompress the file, and inside the directory, type python configure.py. On 64-bit Linux type python configure.py -d /usr/lib64/python2.4/site-packages/. Then make and make install
  • Install MySQLdb (Python library allowing MySQL databases access)
    • The file is available here
  • Install SciPy (Necessary for the SciPy package)
    • Windows
      • Download and install the binary version of NumPy and SciPy
    • Linux
      • Download and install BLAS [1] - Note: Most package managers allow easy download and installation through the native interface however; you must also get the source code and set the environment variable BLAS_SRC
      • Download and install LAPACK [2] - Note: Most package managers allow easy download and installation through the native interface however; you must also get the source code and set the environment variable LAPACK_SRC
      • Download and install numpy [3]
      • Download and install scipy [4]
  • Install matplotlib
    • Windows
    • Linux
      • Files are available here
  • Install ctypes
    • The source is avalable here
Personal tools