Installing madIS
================

The requirements for using madIS are:

1.  **Python 2.6.x** or **2.7.x**: You can download a windows distribution of Python from
    http://www.python.org/download/releases/ .You need to download the latest
    Python in the 2.6 or 2.7 series. madIS doesn't currently work with Python 3.0.

.. note::
    Most Linux distributions come with Python 2.6 or Python 2.7 preinstalled.

2.  **APSW**:

    Windows:

    Download the windows distribution of `APSW` from http://code.google.com/p/apsw/downloads/list .

    Linux:

    Most Linux distributions contain in their repositories a sufficiently recent version of `APSW`.
    If you use Ubuntu 10.04 and later, install the `python-apsw` package::

        sudo apt-get install python-apsw

    In case `APSW` is not provided with your Linux distribution of choise, `APSW` should be build from source.
    The detailed instructions for building `APSW` can be found at:

    http://apidoc.apsw.googlecode.com/hg/build.html#recommended

    A quick command line to build and install `APSW` is::

        sudo python setup.py fetch --all --missing-checksum-ok build install

    If you are using Ubuntu and wish to build `APSW` from source, you also need the packages::

        build-essential libreadline5-dev zlib1g-dev python-dev

    MacOS:

    On *MacOS 10.6* install `APSW` and `readline` via `easy_install`::

        easy_install apsw
        easy_install readline

    On *MacOS 10.7* you have to build `APSW` from source (as showed above), for which you'll
    also need *Xcode*.

    After building `APSW`, install `readline` via `easy_install`::

        sudo easy_install readline

3.  Download the madIS from:

    https://github.com/madgik/madis

4.  Use madis terminal by entering into the new "madis/" path and executing "python mterm.py [dbfile.db]". When a database file is not given madis opens a database in memory.

5.  Use "cat data | python mexec.py -f script.sql [-d/-w dbfile.db]" to run a madis sql script on data from standard input as a stream. The output is also streamed in the standard output. The mexec script may optionally read (-d) or write (-w) to a database file or run completely in memory.

6.  Use madis as a python library and run SQL queries with UDFs importing "madis.py"::

        import madis
        
    `madis` package is used in the same way as sqlite3 package does. 
    The only difference is that the executed queries may contain madIS UDFs and may be expressed in madIS extended SQL syntax.    
        
7.  madIS also runs with PyPy (https://pypy.org) much faster. Just download the last version or a nightly build of PyPy (compatible with Python 2.7) and run `pypy mterm.py` without downloading and installing APSW. Several terminal features (history, auto-complete etc.) are not yet available with PyPy but it is still operational. madIS as a library and `mexec` script can be used without any limitation.