Changeset 9680


Ignore:
Timestamp:
Feb 18, 2015, 5:23:14 PM (9 years ago)
Author:
steve
Message:

Updates to Installation readme

Location:
trunk/anuga_core
Files:
4 edited

Legend:

Unmodified
Added
Removed
  • trunk/anuga_core/.travis.yml

    r9677 r9680  
    3737  # On osx get GDAL from jjhelmus binstar channel
    3838  - if [[ "$PYTHON_VERSION" == "2.7" ]]; then
    39        conda install -c pingucarsti gdal-data;
     39       conda install -c pingucarsti gdal;
    4040       export GDAL_DATA=`gdal-config --datadir`;
    4141    fi
  • trunk/anuga_core/INSTALL.rst

    r9635 r9680  
    1010
    1111AnuGA is a python package with some C extensions (and an optional fortran
    12 extension). At present AnuGA has not been transported to python 3.
    13 We recommend using python 2.7.
     12extension). At present AnuGA has only been run and tested using python 2.x.
     13We recommend python 2.7.
    1414
    1515If you have a python 2.7 environment with gcc and gfortran support,
     
    1818
    1919
    20 Python Environment
    21 ------------------
     20Installing the dependencies
     21---------------------------
    2222
    23 A safe way to install AnuGA is to use a separate python environment
    24 to isolate it from the rest of your system. Miniconda provides such
    25 an environment.
     23AnuGA requires the following packages:
     24
     25* `numpy <http://numpy.scipy.org/>`_
     26* `scipy <http://scipy.org/>`_
     27* `matplotlib <http://matplotlib.sourceforge.net/>`_
     28* `gdal <http://gdal.org/>`_
     29* `nose <http://nose.readthedocs.org/en/latest/>`_
     30* A C compiler (preferably GCC or MinGW_ on Windows)
     31
     32The easiest and **preferred** way to get all dependencies in the latest
     33version is using the Anaconda_ or the Miniconda_ Python
     34distributions by `Continuum Analytics`_.
     35Miniconda_ allows you to create multiple python environments and is particularly
     36useful if you want to keep multiple versions of AnuGA.
     37It does not require administrative rights to your computer and doesn't
     38interfere with the Python installed in your system.
     39For Windows users Anaconda_ even comes with MinGW_ so you don't have to worry about
     40the many, many, many issues of compiling under Windows.
    2641
    2742
     43Anaconda
     44++++++++
    2845
    29 Easy Install
    30 ------------
     46Once you have downloaded and installed Anaconda_,
     47open a terminal (or ``cmd.exe`` on Windows) and run::
    3148
    32 Pip Install
    33 -----------
     49    conda install pip nose numpy scipy netcdf4 matplotlib
     50    conda install -c pingucarsti gdal
     51   
     52and setup GDAL_DATA environment variable:
    3453
    35 Install from Source
     54    export GDAL_DATA=`gdal-config --datadir`
     55   
     56   
     57Miniconda
     58+++++++++
     59
     60Once you have downloaded and installed Miniconda_,
     61open a terminal (or ``cmd.exe`` on Windows), create
     62a specific environment for AnuGA, by running::
     63
     64    conda create -n anuga_env python=2.7
     65    source activate anuga_env
     66   
     67    conda install pip nose numpy scipy netcdf4 matplotlib
     68    conda install -c pingucarsti gdal
     69   
     70and setup GDAL_DATA environment variable:
     71
     72    export GDAL_DATA=`gdal-config --datadir`
     73   
     74
     75
     76Extra dependencies for Windows users
     77++++++++++++++++++++++++++++++++++++
     78
     79Unfortunately, the ``gcc`` compiler included in Anaconda MinGW_
     80doesn't have OpenMP_ support. This is required to compile
     81some extension modules in AnuGA that have multi-threaded parallel code.
     82
     83You'll have download and install TDM-GCC_
     84**after** you've installed Anaconda and **before** you install AnuGA.
     85Don't forget to mark the ``openmp`` option in the "Choose Components" part of
     86the installation. See this `excellent documentation for Windows users`_
     87(they even have screenshots!).
     88
     89Installing AnuGA
     90----------------
     91
     92After you've installed the dependencies you can proceed to install AnuGA
     93using pip_.
     94Open a terminal (or ``cmd.exe`` on Windows) and run::
     95
     96    pip install anuga
     97
     98and that's it!
     99
     100If you already have AnuGA installed and want to **upgrade** to a newer
     101version, use::
     102
     103    pip install anuga --upgrade
     104
     105To uninstall simply run::
     106
     107    pip uninstall anuga
     108
     109
     110.. note::
     111
     112    The Windows installer from older versions is no longer supported.
     113
     114Installing the latest development version
     115-----------------------------------------
     116
     117If you want the very latest code and features,
     118you can install AnuGA directly from Github_.
     119We try to maintain the *master* branch stable and
     120`passing all tests <https://travis-ci.org/stoiver/anuga_core/branches>`__,
     121so it should be safe to use.
     122
     123First, you'll need to `install git`_.
     124Then, open a terminal and run::
     125
     126    git clone --depth=50 --branch=master git://github.com/stoiver/anuga_core.git
     127
     128This will fetch the source code from Github_
     129and place it in a folder called ``anuga_core`` in the directory where you ran the
     130command.
     131Then, just ``cd`` into the directory and run ``pip``::
     132
     133    cd anuga_core
     134    pip install --upgrade .
     135   
     136Testing the install
    36137-------------------
    37138
     139
     140From the source directory run ``python runtests.py``
     141
     142    python runtests.py
     143   
     144
     145If you get an error message or weird result,
     146please write to the `mailing list`_.
     147To make it easier for us to debug you problem, please include the following
     148information:
     149
     150* Operating system
     151* Python distribution (Anaconda_, PythonXY_, `ETS/Canopy`_, own install)
     152* Python version (2.6, 2.7, 3.3, 3.4, etc)
     153* The script you ran (and gave you an error/weird result)
     154* The error message (the part that says ``Traceback: ...``) or result (figure,
     155  numbers, etc)
     156   
     157   
     158.. _install git: http://git-scm.com/
     159.. _Github: https://github.com/stoiver/anuga_core/
     160.. _Python: http://www.python.org/
     161.. _pip: http://www.pip-installer.org
     162.. _MinGW: http://www.mingw.org/
     163.. _mailing list: anuga-user@lists.sourceforge.net
     164.. _Continuum Analytics: http://continuum.io/
     165.. _Anaconda: http://continuum.io/downloads
     166.. _Miniconda: http://conda.pydata.org/miniconda.html
     167.. _PythonXY: http://code.google.com/p/pythonxy/
     168.. _ETS/Canopy: http://code.enthought.com/projects/index.php
     169.. _OpenMP: http://openmp.org/
     170.. _TDM-GCC: http://tdm-gcc.tdragon.net/
     171.. _excellent documentation for Windows users: http://docs-windows.readthedocs.org/en/latest/devel.html#mingw-with-openmp-support
     172
  • trunk/anuga_core/anuga/__init__.py

    r9674 r9680  
    3939# Dev branch marker is: 'X.Y.dev' or 'X.Y.devN' where N is an integer.
    4040#
    41 __version__ = '1.3.7'
     41__version__ = '1.3.8'
    4242
    4343__svn_revision__ = filter(str.isdigit, "$Revision$")
  • trunk/anuga_core/setup.py

    r9674 r9680  
    7474                install_requires=['numpy',
    7575                          'scipy',
    76                           'netcdf4',
    77                           'nose',
     76                          'netcdf4',
    7877                          'matplotlib',
    79                           'gdal']
     78                          'gdal',
     79                          'nose']
    8080    )
    8181else:
     
    149149                    **extra_setuptools_args)
    150150
    151     if (len(sys.argv) >= 2
    152             and ('--help' in sys.argv[1:] or sys.argv[1]
    153                  in ('--help-commands', 'egg_info', '--version', 'clean'))):
     151#     if (len(sys.argv) >= 2
     152#             and ('--help' in sys.argv[1:] or sys.argv[1]
     153#                  in ('--help-commands', 'egg_info', '--version', 'clean'))):
     154#
     155#         # For these actions, NumPy is not required.
     156#         #
     157#         # They are required to succeed without Numpy for example when
     158#         # pip is used to install anuga when Numpy is not yet present in
     159#         # the system.
     160#         try:
     161#             from setuptools import setup
     162#         except ImportError:
     163#             from distutils.core import setup
     164#
     165#         metadata['version'] = VERSION
     166#     else:
     167#         from numpy.distutils.core import setup
     168#
     169#         metadata['configuration'] = configuration
     170       
    154171
    155         # For these actions, NumPy is not required.
    156         #
    157         # They are required to succeed without Numpy for example when
    158         # pip is used to install anuga when Numpy is not yet present in
    159         # the system.
    160         try:
    161             from setuptools import setup
    162         except ImportError:
    163             from distutils.core import setup
    164172
    165         metadata['version'] = VERSION
    166     else:
    167         from numpy.distutils.core import setup
     173    metadata['version'] = VERSION
     174    metadata['configuration'] = configuration
    168175
    169         metadata['configuration'] = configuration
    170 
     176    from numpy.distutils.core import setup
     177   
    171178    setup(**metadata)
    172179
Note: See TracChangeset for help on using the changeset viewer.