How does the layout of a Python application look? [closed]

Question

In most programming environments it's clear how the code is distributed into several parts and how everything interacts. In Python I seem to be completely lost.

• How should the layout of a Python application look?

  Currently I have:

  setup.py
  application_name/
      __main__.py
      __init__.py
      views/
      controllers/
      model/
      resources/   <- images, videos, ...
  

• How does one execute the application?

  I've got a runner script with the following content

  #!/usr/bin/env python -m "application_name"
  

  Should one even use __main__.py for this purpose? Is a runner script necessary?

• How should one import parts of the application? (Python 2.6)

  For example in application_name/__main__.py

  from . import controllers.MainWindow
  

How do you layout your applications?

Answer 1

There are several parts to this question so I'll try to answer them in turn:

1: Its really up to you, there are no hard-and-fast rules beyond those for establishing that a directory should be treated as a package and so on. Some frameworks will prescribe a directory structure using a script to generate scaffolding (a bit like Rails does in the Ruby world) but this is purely a convenience or convention of the given framework. Organise your code and modules so they make sense logically as you would in any other language.

2: What you have there is absolutely fine. Alternatively you can use an installed script if you are using distutils, a console_script as part of a .egg install, or as a last resort just call the main.py (or whatever you name it) script directly. The console_script is quite common though and is used by tools such as the nose testing framework for example.

3: There is a PEP for this specific topic. In my experience though you should really prefer absolute imports to relative ones. To force this behaviour you can do:

 from __future__ import absolute_import