`suppress_output` isn't compatible with Windows, causing install-at-import to fail

[dap-biospec]

Fresh installation using:

pip install -U autoeis
... [success] 

followed by:

import numpy as np
import autoeis as ae

Traceback (most recent call last):
  File "<stdin>", line 1, in <module>
  File "D:\Python\Python312\Lib\site-packages\autoeis\__init__.py", line 1, in <module>
    from . import core, io, metrics, parser, utils, visualization  # noqa: F401, E402
    ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  ....
  File "D:\Python\Python312\Lib\site-packages\juliapkg\install_julia.py", line 23, in log
    print(prefix, *args)
OSError: [WinError 1] Incorrect function
[juliapkg] Locating Julia ^1.6.1

 # Load test dataset shipped with AutoEIS
 Z, freq = ae.io.load_test_dataset()
Traceback (most recent call last):
  File "<stdin>", line 1, in <module>
NameError: name 'ae' is not defined

Same behavior with standalone or Microsoft store installation; with or without Julia pre-installed.

Recommendations?

[ma-sadeghi]

@dap-biospec Sorry for the late reply. Not sure about the root cause, but I believe this should now be fixed. Could you please give it another try? (ideally in a fresh virtualenv just in case)

[ma-sadeghi]

Since our automated tests are currently passing for the combination of Windows and Python 3.12, I'll close this issue, but feel free to re-open if you still can't import autoeis.

[dap-biospec]

@ma-sadeghi Package us is now locating Julia versions "~1.6.1, ~1.7, ~1.8, ~1.9, =1.10.0" but returns the same "OSError: [WinError 1] Incorrect function." Subsequent code returns "name 'ae' is not defined"

[ma-sadeghi]

@dap-biospec What Python distribution do you use? Anaconda, Pyenv, etc., or did you download it directly from Python.org?

[ma-sadeghi]

@dap-biospec Sorry for the frustration. I got my hands on a Windows machine this morning, and figured out the root cause. It's now fixed, please upgrade autoeis, and should be good to go.

PS. The reason our Windows CI was not failing was because we were using the CLI approach to install Julia dependencies (python -m autoeis install), which worked fine, but the install-at-first-import didn't, which wasn't tested. Now, all CIs install Julia dependencies at import, so it's now fully covered in case something goes wrong in the future.