ScribeHandler is a simple proxy layer that works with the python standard
logging module (http://docs.python.org/library/logging.html).  ScribeHandler
acts a a handler object that gets added to a logger in the standard way.

Important bits:

When instantiating the handler, you will want to specify the connection
setting to scribed.  The following keyword arguments are reconized (and
defaults):

Argument            Default
---------------------------
host                127.0.0.1
port                1463
category            %(hostname)s-%(loggername)s
transport           ScribeHandler.FRAMED
uri                 None

- Host and port, should be obvious

- If transport is ScribeHandler.FRAMED, then
    thrift.transport.TTransport.TFramedTransport() is used.
- If transport is ScribeHandler.BUFFERED, then
    thrift.transport.TTransport.TBufferedTransport() is used.
- If transport is ScribeHandler.HTTP then
    thrift.transport.THttpClient.THttpClient() is used and uri MUST NOT be
    None
- If transport is None then you can assign any acceptable Thrift/Scribe
    transport via 

        ScirbeHandler.transport = <transport>

    where <transport> is a valid Scibe/Thrift transport (this is an attempt at
    future proofing).

- category is translated into the category parameter to scribe.LogEntry()
    object.  It should be a standard format string using mapping keys.
    The following are the reconized mapping keys:

    Key Name        Source
    ----------------------
    module          logging.LogRecord.module
    levelname       logging.LogRecord.levelname
    loggername      logging.LogRecord.name
    processname     logging.LogRecord.processName
    hostname        socket.gethostname()


Here's a simple example of how to use it:

import logging
import ScribeHandler
mylogger = logging.getLogger('ScribeLogger')
mylogger.setLevel(logging.DEBUG)
handler = ScribeHandler.ScribeHandler(category='%(hostname)s-%(levelname)s', file_buffer='/tmp/logbuffer')
mylogger.addHandler(handler)

mylogger.debug('stuff happens')