Problem with relative imports in project with lot of nested folders

Apologies in advance if my question seems basic. I'm relatively new to Python and I'm currently in the process of learning the language.
I want to use relative paths in my project to refer to files (e.g. config files, log files etc); browsing online i found that it is good practice to have a file in the root directory or in a first level directory that contains a sort of ROOT_DIR variable that you can import and use in the rest of the project. Given the following code structure, I'm having problems implementing this solution:

root_dir
├── config
│   ├── config.ini
│   ├── definitions.py
├── libs
│   ├── lib_1
│   │   ├── A.py
│   │   └── test
│   │       └── test_A.py
│   ├── lib_2
│   │   ├── B.py
│   │   └── test
│   │       └── test_B.py
├── main.py

The content of the definitions.py file is:

import os
ROOT_DIR = os.path.realpath(os.path.join(os.path.dirname(__file__), '..'))

This works fine if I use ROOT_DIR from within the main.py that is one level above definitions.py but does not work if I use ROOT_DIR from, say, A.py or B.py, doing from config.definitions import ROOT_DIR.
I tried adding __init__.py files to every directory but nothing. PyCharm seems to be fine and recognizes the import but if I run A.py or B.py from the terminal either from the root folder or from lib_1/lib_2, ModuleNotFoundError is fired.

Can anyone explain to me how can I achieve this behaviour using relative imports from any subfolder of the project? What is the pythonic way to do so?
Thank you all in advance

The issue you're experiencing comes from how you're running your Python scripts. When you run a Python script directly, Python's module search path (sys.path) begins from the script's directory. As a result, if you're inside lib_1 or lib_2 and try to run A.py or B.py directly, Python won't be able to find the config module, as it's neither in the script's directory nor in the PYTHONPATH.

Here is the solution: instead of running A.py or B.py directly, you should execute them as modules using Python's -m flag, starting from the root directory of your project. Here's how you can do that:

1. Open your terminal and navigate to the root directory of your project (root_dir).
2. Run your Python modules with the following commands:

python -m libs.lib_1.A
python -m libs.lib_2.B

Here, the -m flag tells Python to run the module as a script, and Python will be able to correctly resolve the relative import from config.definitions import ROOT_DIR.

This is a Pythonic way of dealing with module organization and it makes your code more maintainable and portable without having to worry about manipulating sys.path or about operating system specific paths.

Also, remember to add an empty __init__.py file in each directory that you want Python to recognize as a package. This includes root_dir/config, root_dir/libs, root_dir/libs/lib_1, root_dir/libs/lib_2, and the test subdirectories. These files can be empty, but they need to be present.