[fix][test] test script cannot create temp folder on Win11

Discussed in #57 (comment 100977)

mentioned in issue #57 (closed)

created branch 59-fix-test-test-script-cannot-create-temp-folder-on-win11 to address this issue

Using Powershell terminal, to activate environment resulted in error: venv\Scripts\activate : File C:\Users\jjd1c23\Documents\MELDB\concepts-processing\venv\Scripts\Activate.ps1 cannot be loaded because running scripts is disabled on this system. To Enable execution policy run Set-ExecutionPolicy RemoteSigned in PS with admin privileges. Now can activate venv within powershell.

Running hatch run pytest command still resulted in error: PermissionError: [WinError 5] Access is denied: 'C:\\Users\\jjd1c23\\Documents\\MELDB\\concepts-processing\\tests\\tmp\\phen\\.git\\objects\\b3\\cb7928d503b6fb0bb489780d7215a937814da6'

changed the description

Attempt to use raw string for path links (r"") did not succeed.

New temporary folders are created and tests are seem to run successfully. However shutil.rmtree() deletes whole repo after each test, which is denied permission when deleting "./git" folder.

Similar Issue on Stackoverfow#76546956, however changing file permissions of all files in temp_dir has not work.

Errors can be ignored with shutil.rmtree(temp_dir, ignore_errors=True) to allow tests to pass, however as the files: .git/objects/b3, .git/objects/bf, .git/objects/e6 are not deleted, may result in unexpected behaviour. Phen.py makes use of the same function, making this not a viable solution.

@jjd1c23 This is a Windows python environment problem. When I've developed on Windows I've done so through Anaconda, created a shell and then launched jupyter-lab from the working directory. So essentialy it's the same type of environment that we use on linux.

I would suggest switching to jupyter-lab on the windows if it was only a test problem but as we are using the shutil.rmtree function bin phen.py we are probably going to need to solve it.

We could add the function to util.py to do a depth first deletion using os. I've not tested this but it's an option...

import os
import stat

def delete_directory(directory):
    """Recursively deletes a directory and all its contents using os.walk() depth first."""

    if os.path.exists(directory):

        # Walk the directory tree bottom-up (depth-first)
        for root, dirs, files in os.walk(directory, topdown=False):
            # Delete all files
            for file in files:
                file_path = os.path.join(root, file)
                try:
                    os.chmod(file_path, stat.S_IWRITE)  # Ensure writable
                    os.remove(file_path)
                except Exception as e:
                    _logger.error(f"Failed to delete file {file_path}: {e}")

            # Delete all directories
            for dir in dirs:
                dir_path = os.path.join(root, dir)
                try:
                    os.rmdir(dir_path)
                except Exception as e:
                    _logger.error(f"Failed to delete directory {dir_path}: {e}")

        # Remove the root directory itself
        try:
            os.rmdir(directory)
            _logger.debug(f"Deleted: {directory}")
        except Exception as e:
            _logger.error(f"Failed to delete {directory}: {e}")
    else:
        print(f"Directory not found: {directory}")

I have tried setting up Anaconda and running JupyterLab, however this seems to just run a standard PS resulting in the same error.

The DFS delete function above as well as the exception handler below, both catch the Win Error 5 and all files in .git folder can be deleted.

def handleRemoveReadonly(func, path, exc):
    if not os.access(path, os.W_OK):
        print("HANDLING ERROR", path)
        os.chmod(path, stat.S_IWUSR)
        func(path)
    else:
        print("NOT HANDLING ERROR", path)
        raise

However the when trying to delete the empty \phen itself an new error is thrown: PermissionError: [WinError 32] The process cannot access the file because it is being used by another process

SOLUTION: install wsl and Ubuntu 22.04.5 LTS environment. Python and Python-venv need to be installed with sudo apt install, before running instructions from contributing.md. @mjbonifa

closed