How to source a .py script that resides in a folder starting with a digit? - SOLUTION POSTED AT BOTTOM

Question

SOLUTION POSTED AT BOTTOM

Thanks to the commentators for pointing me in the right direction. Between their help and this other thread I was able to come to a robust solution which is posted at the bottom.

Situation

I have a .py script called job.py that relies on data in another .py script called data.py. Both scripts reside in the same folder, 05_scripts. The 05_scripts folder sits under Git version control in a parent folder called project.

Thus, the complete paths to the two scripts are:
project/05_scripts/job.py and
project/05_scripts/data.py.

Problem: I cannot import data.py without renaming the folder 05_scripts.

In job.py I want to source the data.py file. This is problematic because the data.py file sits in a folder whose name begins with a digit, 05_scripts/. If I change the folder name to _05_scripts/ or something else so that the first character is non-numeric, then I achieve my goal and import the data.py script without a problem. However, this is unacceptable since this is part of a much larger project and it is not acceptable to change all references to this folder.

What an acceptable solution must look like

The solution I need is code in job.py that can successfully import data.py without renaming the 05_scripts/ folder.

What I have tried

Attempt 1

Input:

from 05_scripts.data import *

Output:

  File "<stdin>", line 1
    from 05_scripts.data import *
         ^
SyntaxError: invalid token

Attempt 2

Input:

from '05_scripts'.data import *

Output:

  File "<stdin>", line 1
    from '05_scripts'.data import *
                    ^
SyntaxError: invalid syntax

Attempt 3

Input:

from '05_scripts.data' import *

Output:

  File "<stdin>", line 1
    from '05_scripts.data' import *
                         ^
SyntaxError: invalid syntax

Attempt 4

Input:

from data import *

Output:

ModuleNotFoundError: No module named 'data'

Attempt 5

Input:

import data

Output:

ModuleNotFoundError: No module named 'data'

Other attempts

I also tried variations of the solutions posted here. This included different variations of using __import__ and importlib.import_module() but could not get a working solution that did not include renaming the parent folder.

Recall, this works but is not an acceptable solution for my use case:

After renaming 05_scripts/ to _05_scripts,

Input:

from _05_scripts.data import *

SOLUTION

# Getting the path for the project and doing string manipulation.
import os
import re
working_dir = re.sub("\\\\", "/", os.getcwd())
scripts_dir = working_dir + "/05_scripts"

# Importing all from the data.py script
import site
site.addsitedir(scripts_dir)
from data import *

Answer 1

import os
import sys
syspath_05scripts = os.path.join(os.path.dirname(os.path.abspath(__file__)), "05_scripts")
sys.path.insert(0, syspath_05scripts)

import something_from_05

Answer 2

You can use site package to add custom locations to be included to for module search path

>>> import site
>>> site.addsitedir('05_scripts')
>>> import data
>>> data
<module 'data' from '<removed>/05_scripts/data.py'>

Answer 3

Assuming that you have a folder structure like this the following code should work for you.

project/main.py
project/05_scripts/
project/05_scripts/__init__.py
project/05_scripts/job.py
project/05_scripts/data.py

project/05_scripts/data.py

what_i_want_to_get = {'hello': 'there'}

project/main.py

from importlib import import_module

# this is the same as calling "import 05_scripts.data as data"
data = import_module('05_scripts.data')

print(data.what_i_want_to_get)

# this should print {'hello': 'there'}