In the past, whenever I had a script that I wanted to properly configure the settings for, I would use something like the following idiom at the top of the script:
import sys, os; dirname = os.path.dirname
# sys.path.insert(0, dirname(dirname(__file__)))
sys.path.insert(0, dirname(__file__))
os.environ['DJANGO_SETTINGS_MODULE'] = 'myapp.settings'
Notice that this is a relative setting to `__file__` variable in the script. The djangopath function is an attempt to do away with the above such that I can now write the following:
from lib import djangopath; djangopath(up=2, settings='myapp.settings')
This seems to work for me, but it assumes that you are packaging your script inside your projects/apps. If they are elsewhere then you may need to resort to another method (e.g. absolute paths, etc.)
AK
If you have a production, staging, testing and development environment, you might want to have a global checked in (in your version control system) settings.py for production + a local settings.py to override various settings (like database connection).
It's also good for development, since developers don't - by incident - commit to the production settings.py, since they can use their local settings, that should be ignored (.cvsignore, .svnignore or similar).