Import module in another directory from a "parallel" sub-directory

Question

I want to have a hierarchy that looks like this (and it has to look like this)

main_folder\
    main.py
    domain_sub_directory\
        __init__.py
        domain.py
    ui_sub_direcotory\
        __init__.py
        menu.py

I need to activate ui.py frome main.py but then acces domain.py from menu.py. How can I do that ?

I did this in main:

    import ui_sub_directory.ui

This in ui:

    import domain_sub_directory.domain

But the UI module does not see the domain module.

What am I doing wrong ?

BTW do I need to also import the class I'm working with ? and what is the difference between this and:

from x import y 

?

* Edit * for those who don't understand I want to import from:

folder1 /folder2 /folder3 /module1 

I want to import this:

folder1 /folder2 /module2

Answer 1

You asked the difference in the import statements. Its partially a matter of the namespace for which the object will be imported under, and also a way to limit the exact amount of code that is imported.

import os
from os import path

Both os and os.path are modules. The first imports the entire os module and all its submodules. This could be more than you need, and for big libraries might be unneeded overhead. Though you can still access path via os.path

The second form is a way to selectively only import the path module. Additionally, instead of coming into your code under the os namespace, it now lives at the root level as just path.

While this link Import Script from a Parent Directory does tell you everything you need to know, here is some more specific info:

# this will make your package available on your pythonpath
sys.path.append("/path/to/main_folder")

Then your various scripts can reference other module all relative to under main_folder, such as:

from ui_sub_direcotory import menu

from domain_sub_directory import domain

import main

These are all valid imports inside your package.