I have the same issue as you did. I think the problem is that you used relative import in in-package import. There is no __init__.py in your directory. So just import as Moses answered above.
The core issue I think is when you import with a dot:
from .p_02_paying_debt_off_in_a_year import compute_balance_after
It is equivalent to:
from __main__.p_02_paying_debt_off_in_a_year import compute_balance_after
where __main__ refers to your current module p_03_using_bisection_search.py.
Briefly, the interpreter does not know your directory architecture.
When the interpreter get in p_03.py, the script equals:
from p_03_using_bisection_search.p_02_paying_debt_off_in_a_year import compute_balance_after
and p_03_using_bisection_search does not contain any modules or instances called p_02_paying_debt_off_in_a_year.
So I came up with a cleaner solution without changing python environment valuables (after looking up how requests do in relative import):
The main architecture of the directory is:
main.py
setup.py
problem_set_02/
__init__.py
p01.py
p02.py
p03.py
Then write in __init__.py:
from .p_02_paying_debt_off_in_a_year import compute_balance_after
Here __main__ is __init__ , it exactly refers to the module problem_set_02.
Then go to main.py:
import problem_set_02
You can also write a setup.py to add specific module to the environment.