Skip to content

zold Building UVCDAT

Jump to bottom
mattben edited this page Sep 8, 2014 · 1 revision

####Prerequisites and Install Tips You need a recent version of CMake (2.8.8 or later) in order to install UV-CDAT on Max OSX and Linux platforms. To build UV-CDAT on your system you will need common devel packages for your OS distribution along with fortran (gfortran) and Qt (4.8.4 but not 5.x). Also, you will need CMake 2.8.8 or later. You can get CMake from http://cmake.org/cmake/resources/software.html.

For detailed information on required pacakges and versions, please see UV-CDAT Requirements

####To build CMake with the Qt-GUI follow the instructions below

tar xvf cmake-2.8.9.tar.gz
cd cmake-2.8.9
./configure --prefix=/install/directory/cmake-2.8.9 --qt-gui
gmake install

####Getting Started If you do not install QT in a standard location, don't forget to setup the "QT_QMAKE_EXECUTABLE" in the ccmake step bellow. Once you have CMake installed on your system you can begin configuring UVCDAT with CMake. To do this you first need to get the source using git by running the following:

git clone git://github.com/UV-CDAT/uvcdat.git    
or    
git clone https://github.com/UV-CDAT/uvcdat.git    
or if you have a github account    
git clone [email protected]:UV-CDAT/uvcdat.git

Alternatively you can download a zip file of the repo from:    
https://github.com/UV-CDAT/uvcdat

To build a specific branch or UV-CDAT from devel repository follow these instructions (after you have checked out UV-CDAT from the repositories above)

git remote add devel git://github.com/UV-CDAT/uvcdat-devel.git
or 
git remote add devel https://github.com/UV-CDAT/uvcdat-devel.git
git checkout -b devel-master --track devel/master

If you have already done the checkout and added the devel repository and want to build a specific branch then do the following:

cd your_top_level_uvcdat_source_directory
git fetch --all -p
git checkout -b branch_name --track devel/branch_name

If you have already built the UV-CDAT and want to use a different branch of VisTrails then do the following:

cd your_top_level_uvcdat_install_directory
cd vistrails
git fetch --all -p
git checkout -b branch_name --track devel/branch_name

Adjacent to your source checkout you should make a build directory. From within the build directory we invoke either ccmake which uses an ncurses interface (great when you only have terminal acces) or cmake-gui which launches CMake's Qt gui interface:

mkdir build-uvcdat
cd build-uvcdat
cmake-gui ../uvcdat

Alternatively you can use the command line only configuration:

ccmake ../uvcdat

The cmake-gui window (or a text version of it if you used ccmake) will appear. Press the Configure button (or C if you're using ccmake). If you are behind a firewall in which the git:// protocol is blocked, you should set GIT_PROTOCOL CMake variable to use http:// instead. CMake will checkout all dependencies during this step and this might take a while. Make sure you have an internet connection during the Configure and Building process. The window will look like this:

alt text

The CMake interface provides you with a number of options, particularly telling the build system to build a particular package or to use a system versions. The safest bet it to not adjust any option. CMake will install uv-cdat in an install directory adjacent to the source and build directories.

Press Configure again (or c if you're using ccmake) until you don't see anymore "red highlighte dlines" (or "starred lines" if using ccmake) and then Generate (or g if you're using ccmake). Close the cmake-gui window (or press Q if you're using ccmake). Go back to your terminal and run make. It should start building.

make

After the process finishes, you have to setup your environment. Under bash:

source /path/to/cmake/install/bin/setup_runtime.sh

Or if you are using (t)csh:

source /path/to/cmake/install/bin/setup_runtime.csh

The installed version of UV-CDAT will be by default in the 'install' subdirectory of build.

####Special Builds

During the ccmake step there are a few options you might want to tune for easier/faster build.

  • CDAT_BUILD_ESGF: turning this ON will be a SUPER light weight version of UV-CDAT w/o GUI or Graphics, and almost no packages, it is the equivalent of the old --cdms-only, or the (independently maintained) CDATLite
  • CDAT_BUILD_GRAPHICS: turning this OFF will disable building anything graphics related (including GUI)
  • CDAT_BUILD_GUI: turning this off will turn off anything gui related (TKinter, Vistrails, UVCDAT GUI, Paraview, VisIT)
  • CDAT_BUILD_PARALLEL: turning this off will disable openmpi/mpi4py features

####Testing matplotlib

Make sure matplotlib was installed by running python (make sure you have set the PATH variable to point to UV-CDAT python) and import matplotlib:

python
import matplotlib
import pylab

If the comands above didn't raise any errors, proceed to the "Running UV-CDAT" section. Else, install matplotlib:

mkdir matplotlib_git
cd matplotlib_git
git clone git://github.com/matplotlib/matplotlib.git #or https://github.com/matplotlib/matplotlib.git
cd matplotlib
python setup.py install

####Running UV-CDAT

After sourcing the setup_runtime.sh as mentioned above:

uvcdat

All packages should be loaded automatically and you should see the window below:

alt text

Clone this wiki locally