Metadata-Version: 1.1
Name: pyJHTDB
Version: 20141230
Summary: Python wrapper for the Johns Hopkins turbulence database library
Home-page: http://turbulence.pha.jhu.edu/
Author: Johns Hopkins Turbulence Database Group
Author-email: turbulence@pha.jhu.edu
License: Apache Version 2.0
Download-URL: https://github.com/idies/pyJHTDB
Description: 
        =======
        pyJHTDB
        =======
        
        Python wrapper for the JHU Turbulence Database Cluster library.
        More information can be found at http://turbulence.pha.jhu.edu/.
        
        Installing pypi version
        =======================
        
        If you have ``pip``, you can simply do this:
        
        .. code:: bash
        
            pip install pyJHTDB
        
        If you're running unix (i.e. some MacOS or GNU/Linux variant), you will
        probably need to have a ``sudo`` in front of the ``pip`` command.
        If you don't have ``pip`` on your system, it is quite easy to get it
        following the instructions at
        http://pip.readthedocs.org/en/latest/installing.html.
        
        **Cutout/local data functionality**
        
        If you want to use the cutout functionality, you will need to install
        ``h5py`` *before* you install ``pyJHTDB``, with:
        
        .. code:: bash
        
            pip install h5py
        
        If you would like to use the local data functionality, you will need a
        full installation of the HDF5 libraries, see
        http://www.hdfgroup.org/HDF5/ for instructions.
        
        Installing from source
        ======================
        
        **ubuntu 14.04**
        
        Bare-bone installation:
        
        .. code:: bash
        
            sudo apt-get install build-essential gfortran
            sudo apt-get install python-setuptools
            sudo apt-get install python-dev
            sudo easy_install numpy
            sudo python setup.py install
        
        Note that doing this should, in principle, also install ``sympy`` on
        your system, since it's used by ``pyJHTDB``.
        
        Happy fun installation:
        
        .. code:: bash
        
            sudo apt-get install build-essential gfortran
            sudo apt-get install python-setuptools
            sudo apt-get install python-dev
            sudo apt-get install libpng-dev libfreetype6-dev
            sudo apt-get install libhdf5-dev
            sudo easy_install numpy
            sudo easy_install h5py
            sudo easy_install matplotlib
            sudo python setup.py install
        
        I haven't tested the installation on any other system, but I think
        reasonable variations on the above should work for the minimal
        installation on all unix systems (i.e. for MacOS as well).
        If you manage to get it working (i.e. you import test_plain like the
        README says and you can run it), please let me know what steps you
        needed to take for your system, so I can append the instructions to
        this file.
        
        Basic usage
        ===========
        
        On first contact with this library, we recommend that you first run
        ``test_plain``. To be more specific:
        
        .. code:: python
        
            from pyJHTDB import test_plain
            test_plain()
        
        The code that is executed can be found in "pyJHTDB/test.py", and it's
        the simplest example of how to access the turbulence database.
        
        Configuration
        =============
        
        While our service is open to anyone, we would like to keep track of who
        is using the service, and how. To this end, we would like each user or
        site to obtain an authorization token from us:
        http://turbulence.pha.jhu.edu/help/authtoken.html
        For simple experimentation, the default token included in the package
        should be valid.
        
        If you do obtain an authorization token, please write it in the file
        ``auth_token.txt``, in the folder ``.config/JHTDB`` from your home
        folder. This folder should be generated automatically upon first
        importing the package.
        
        The ``.config/JHTDB`` folder is also used to store data used by the
        ``pyJHTDB.interpolator.spline_interpolator`` class, including shared
        libraries. If you do not plan on using the local interpolation
        functionality, no data files will be generated.
        
        
Platform: UNKNOWN
Classifier: Development Status :: 4 - Beta
Classifier: Environment :: Console
Classifier: Intended Audience :: Science/Research
Classifier: License :: OSI Approved :: Apache Software License
Classifier: Natural Language :: English
Classifier: Programming Language :: Python
Classifier: Topic :: Scientific/Engineering :: Mathematics
Classifier: Topic :: Scientific/Engineering :: Physics
