possible ctypes and numpy conflict?

[herbal_rage]

Greetings, i wonder if anyone could help me to solve the following problem with python and following libraries: numpy, cv2, ctypes, pywin32 as they seem to conflict.
The stroy goes as follows: i always had cv2 with numpy on my machine in a system wide python installation, and i'm pretty sure i had pywin32 as well. However after downloading ctypes all of the above libraries suddenly
started to error in their init files mostly telling that the import of "windll" kept on failing. After a closer look at the isue i came across this post on github stating that ctypes could need to be "activated" by executing the postinstall file
in ctypes directory. After that windll succesefully apeared, however the lower and uppercase letters seem to differ from what numpy tried to import. and following that the ctypes directory immideatly went to previos state with no windll
after console session was closed from the cmd by typing quit(). At this point keeping ctypes is not important as i need to recover my previos projects. The following actions were already taken by me: Full python & packages reinstall [pip cache cleared],
making fully independednt virtual enviroment, removing packages and reinstalling cv2 and numpy with pip --force-reinstall, windows recovery and the only code that was attempted with the usage of ctypes related libraries was the following


'''

import pythoncom
import win32com.client
print("dbg success")
#must be 0 deal with it
context = pythoncom.CreateBindCtx(0)
#getting the objects
running_coms = pythoncom.GetRunningObjectTable()
#making an enum
monikers = running_coms.EnumRunning()

for element in monikers:
	print(f"displayname: {element.GetDisplayName(context, element)}")

`

which interestingly enough never cleared the console and thus on every console output the following appears:

`
dbg success
displayname: clsid:A3189FC5-BA37-4EF5-BC41-102A87B07C00:
displayname: clsid:E0EC0F2B-773D-4DD7-BE6C-7D85D6AA6269:
displayname: !Personal-Monikers::FileSyncClient
displayname: !Personal-Monikers::ToastActivation
displayname: !Personal-Monikers::SyncEngineStorageProviderHandlerProxy
displayname: !Personal-Monikers::SyncEngineCOMServer
`
whether this actually affects anyhting im not sure, its usefull to mention that just for this script another editor was used called sublime text 3.

Besides, thats the error i keep on recieving:

`

Error:
Traceback (most recent call last):
  File "homework.py", line 1, in <module>
    import numpy as np
  File "C:\Users\grass\EE_CS\lib\site-packages\numpy\__init__.py", line 140, in <module>
    from . import _distributor_init
  File "C:\Users\grass\EE_CS\lib\site-packages\numpy\_distributor_init.py", line 9, in <module>
    from ctypes import WinDLL
ImportError: cannot import name 'WinDLL'

`

[JoelMendoza]

Hello,

It looks like you're dealing with a potential conflict between ctypes and numpy. This type of issue can happen when ctypes tries to interface with numpy arrays, as both libraries handle memory allocation in different ways.

Here are a few things to check or try:

1. Memory Buffer Ownership:

If you're passing a numpy array to ctypes, you should make sure that the numpy array is using a memory buffer that ctypes can safely reference. You can access the underlying buffer of a numpy array using .ctypes.data:

import numpy as np
import ctypes

# Create a numpy array
arr = np.array([1.0, 2.0, 3.0], dtype=np.float64)

# Pass the numpy array's data to ctypes
arr_ptr = ctypes.cast(arr.ctypes.data, ctypes.POINTER(ctypes.c_double))

This ensures that you're passing a pointer to the raw memory instead of creating a new ctypes object that may conflict with numpy’s internal memory management.

2. Ensure Compatibility Between Versions:

Make sure you are using compatible versions of numpy and ctypes. Sometimes, these conflicts arise because certain versions of one library may not work well with others. You can check and update the versions by running:

pip show numpy
pip show ctypes

You can also update numpy (since ctypes is part of Python’s standard library and doesn’t typically need updating):

pip install --upgrade numpy
3. Check for Cython Usage:

If you're using Cython or writing C extensions, conflicts can sometimes arise when mixing ctypes with numpy arrays. In this case, ensure that you're managing memory allocation carefully, especially when passing data between Python and C functions.

4. Investigate Error Messages:

If you're seeing a specific error message, that can help pinpoint the problem. It might indicate whether it's a memory alignment issue, type mismatch, or something else. Feel free to share the error message, and I can help you troubleshoot more effectively.

If none of these suggestions seem to solve the problem, please provide more details or share a code snippet so that we can dive deeper into the issue!