Changes between Version 35 and Version 36 of CdatLite


Ignore:
Timestamp:
24/06/11 10:26:55 (9 years ago)
Author:
spascoe
Comment:

--

Legend:

Unmodified
Added
Removed
Modified
  • CdatLite

    v35 v36  
    11= NEWS = 
    22 
     3 '''2011-06-24''':: 
     4  [http://ndg.nerc.ac.uk/dist/cdat_lite-6.0rc1.tar.gz cdat_lite-6.0rc1.tar.gz] is available.  This release supports {{{nc-config}}} as a mechanism 
     5  for detecting your NetCDF installation.  It has code merged from the latest CDAT master branch.  The version scheme is diverging from CDAT since  
     6  users were confused by the "alpha" tag.  Although CDAT is still alpha the cdms2 package is generally stable. 
    37 '''2011-03-18''':: 
    48  [http://ndg.nerc.ac.uk/dist/cdat_lite-6.0.alpha-4.tar.gz cdat_lite-6.0.alpha-4.tar.gz] is available.  This release fixes a problem 
     
    3438.. sectnum:: 
    3539 
    36 CDAT-Lite is a Python package for managing and analysing climate 
     40Cdat-lite is a Python package for managing and analysing climate 
    3741science data.  It is a subset of the Climate Data Analysis Tools 
    3842(CDAT_) developed by PCMDI_ at Lawrence Livermore National Laboratory. 
    3943 
    40 CDAT-lite aims to compliment CDAT by focussing on it's core data 
     44Cdat-lite aims to compliment CDAT by focussing on it's core data 
    4145management and analysis components and by offering a radically 
    4246different installation system to CDAT.  As a result it is much more 
    4347lightweight (hence the name): CDAT's source distribution is the order 
    44 of 1Gb whereas CDAT-lite is under 5Mb. 
    45  
    46 As of version 5.2-2, CDAT-lite also includes the CF Checker developed by `Rosalyn Hatcher`_.  The latest version of this code is available at the `CF SVN`_ repository. 
     48of 1Gb whereas cdat-lite is under 5Mb. 
     49 
     50Cdat-lite is designed to work with the `CF checker`_ package. 
    4751 
    4852.. _CDAT: http://www2-pcmdi.llnl.gov/cdat 
    4953.. _PCMDI: http://www2-pcmdi.llnl.gov/ 
    50 .. _`Rosalyn Hatcher`: mailto:r.s.hatcher@reading.ac.uk 
    51 .. _`CF SVN`: http://cf-pcmdi.llnl.gov/svn/repository-cf/cf-checker 
    52  
    53  
     54.. _`CF checker`: http://pypi.python.org/pypi/cfchecker 
     55 
     56 
     57cdat-lite versioning 
     58==================== 
     59 
     60Cdat-lite is a project that tracks versions of 2 other projects (CDAT 
     61and cdunifpp).  From version 6.0rc1 the cdat-lite version will not be 
     62based directly on the CDAT version.  This is because CDAT updates it's 
     63version very seldomly and stays as an "alpha" distribution for long 
     64periods when the parts included in cdat-lite are generally stable. 
     65 
     66Full details of which versions of CDAT and cdunifpp a cdat-lite 
     67distribution includes is available in the setup.py file and the 
     68PKG_INFO metadata. 
    5469Installing cdat-lite 
    5570==================== 
     
    6479  $ export HDF5_HOME=/usr/local/hdf5 
    6580  $ easy_install cdat_lite  
     81 
    6682 
    6783Dependencies 
     
    94110cdat-lite will work with NetCDF3 or NetCDF4 but because it is 
    95111referenced by shared libraries (the python C extension modules) it 
    96 must be compiled as position independent code.  It is probably easiest 
    97 to install NetCDF as a shared library (use ``--enable-shared`` in the 
    98 NetCDF ``configure`` script).  Alternatively, you can configure NetCDF with:: 
     112must be compiled as position independent code.  If you have a NetCDF4 
     113installation you almost certainly are using shared libraries and even 
     114if you wish to use NetCDF3 it is probably easiest to install NetCDF as 
     115a shared library (use ``--enable-shared`` in the NetCDF ``configure`` 
     116script).  Alternatively, you can configure NetCDF with:: 
    99117 
    100118  $ ./configure --with-pic ... 
     
    102120If you are using NetCDF4 you will also need to configure HDF5 with ``--enable-shared`` or ``--with-pic``. 
    103121 
    104 cdat-lite will look for a NetCDF installation in several places and 
    105 prompt you if it can't find it.  If your NetCDF is installed somewhere 
    106 unusual, or if you want to select a specific installation, set the 
    107 NETCDF_HOME variable.  E.g.:: 
     122If you have the command ``nc-config`` in your path cdat-lite will 
     123detect all library and include dependencies.  Otherwise cdat-lite will 
     124look for a NetCDF installation in several places. 
     125 
     126If your NetCDF is installed somewhere unusual, or if you want to 
     127select a specific installation, set the NETCDF_HOME variable.  E.g.:: 
    108128 
    109129  # sh users 
     
    119139  # csh users 
    120140  $ setenv HDF5_HOME /usr/local/hdf5 
     141 
     142For compatibility with the ``netcdf4-python`` package cdat-lite also accepts ``NETCDF4_DIR`` AND ``HDF5_DIR`` as synonims for these environment variables. 
    121143 
    122144Note, you don't need these environment variables set to run cdat_lite, 
     
    167189.. _virtualenv: http://pypi.python.org/pypi/virtualenv 
    168190 
     191Platform-specific installation notes 
     192------------------------------------ 
     193 
     194OS X 
     195'''' 
     196 
     197Christopher Lee contributed the following experiences installing on OS X 10.6.7. 
     198 
     199My particular Macbook has an Intel CPU, and the default on the Mac is 
     200to compile for the architecture x86_64. In order to override this 
     201(because python is 32 bit, and the netcdf libraries I use are also 32 
     202bit) I needed to pass in "-arch i386" to the compiler. I also needed 
     203the little endian flag '-DBYTESWAP' when compiling the netcdf 
     204interface (inside libcdms). The -DBYTESWAP flag should be included by 
     205the libcdms configure script, where there is a section for 'darwin' 
     206(OS X), but it's currently configured without BYTESWAP (line 
     2076182). The problem here is that OS X used to run on PowerPC CPUs, 
     208which don't need the BYTESWAP flag. I'm not sure if this is your 
     209configure script or if it's from the cdat package. 
     210 
     211I included the -arch i386 and -DBYTESWAP in the setup.py in the 
     212libcdms section, and the setup works fine. 
     213 
     214After running python setup.py build ; python setup.py install ; I 
     215still get an error when importing cdms2. This problem is caused by the 
     216way that libcdms is linked to the netcdf libraries. The 'normal' Mac 
     217method is to link with absolute paths, but libcdms is linked with 
     218relative paths (the libraries are references with @rpath). The result 
     219is that LD_LIBRARY_PATH environment variable is often empty. I'm not 
     220sure how to fix this in the 'Mac' way with absolute paths, but I added 
     221my $NETCDF_HOME/lib directory to the variable and cdms2 now imports 
     222without error. 
     223 
     224 
    169225Testing the installation 
    170226======================== 
     
    186242 
    187243 
    188 CF Checker 
    189 ========== 
    190  
    191 See ``README_cfchecks`` for full description of the CF checker. 
    192 However, note the script will be installed under the name 
    193 ``cfchecks`` rather than ``cfchecks.py`` as described in ``README_cfchecks``. 
    194  
    195 This script requires udunits2 to be installed as a dynamic library and 
    196 included in your library search path. 
    197  
    198244 
    199245FAQ 
     
    210256-------------------------------------------------- 
    211257 
    212 Differences between CDAT and CDAT-lite can be classified as 
     258Differences between CDAT and cdat-lite can be classified as 
    213259differences in scope, i.e. which packages are included, and installation system. 
    214260 
     
    295341TODO 
    296342 
     343 
     344 
    297345}}}