Logger issue when running test_sim by itself

[ebrahimebrahim]

When running pytest -k test_sim rather than just pytest, we get a logging related error

def getMessage(self):
        """
        Return the message for this LogRecord.
    
        Return the message for this LogRecord after merging any user-supplied
        arguments with the message.
        """
        msg = str(self.msg)
        if self.args:
>           msg = msg % self.args
E           TypeError: not all arguments converted during string formatting

msg        = '  reference sound speed: '
self       = <LogRecord: root, 20, /home/ebrahim/openwater/.venv3.11/lib/python3.11/site-packages/kwave/kWaveSimulation_helper/set_sound_speed_ref.py, 89, "  reference sound speed: ">

/usr/local/lib/python3.11/logging/__init__.py:377: TypeError

This is a k-wave bug that was solved in waltsims/k-wave-python#458, however sicne k-wave dropped python3.9 support before they fixed this bug we cannot upgrade just yet (see #167).

We don't see the issue in the CI because when all tests are run at once the logging handlers are probably set to some kind of null handler.

We can wait until the constraint #167 is relaxed and this will be fixed. Or we can fix it right away by silencing logs in an appropriate way.

[peterhollender]

We can wait a bit. How far out are we from Slicer bumping it's Python version? 3.9 is getting long in the tooth... even 3.11 is in security releases only at this point.

[sjh26]

@jcfr is currently working on updating to Python 3.12, just a few more things to sort out.

[vaishnavidesai09]

Reproduction attempt:

• Commit/branch tested: main (886d683)
• Environment: Windows 10, Python 3.14.2, pytest 9.0.3
• Steps taken:
  1. Cloned repo, created venv, ran pip install -e '.[dev]'
  2. Ran pytest -k test_sim
• Outcome: Partial reproduction — test fails, but with a different error than reported. On Python 3.14, k-wave-python cannot be installed because scipy==1.15.3 requires a Fortran compiler not present on Windows. The originally reported TypeError logger bug is therefore blocked by this install failure on Python 3.14.
• Logs: ModuleNotFoundError: No module named 'kwave'

ERROR: scipy build failed — no Fortran compiler found

• Note: This is consistent with the issue's root cause — kwave version constraints tied to Python version. Applying needs-info to request confirmation of whether this is also reproducible on Python 3.11/3.12 with kwave successfully installed.

(I don't have write access — could a maintainer apply needs-triage:reproduced / needs-info?)

[ebrahimebrahim]

I think you ran into different errors here because you did not install the library correctly. We recently overhauled how to install it, but the doc update hasn't come through yet (#469). You could look at #469 for the needed install instructions. I suspect this logger issue is gone but it would be nice to have confirmation

[vaishnavidesai09]

Updated reproduction attempt (after following new install instructions from #469):

• Commit/branch tested: main (45c0104)
• Environment: Windows 10, Python 3.11.5, pytest 9.0.3, k-wave-python 0.4.0
• Steps taken:
  1. Cloned repo, created venv with Python 3.11, ran pip install -e '.[dev]'
  2. Installed sim extras: pip install "openlifu[app]" + pip install vtk scikit-image trimesh nibabel pydicom
  3. Ran pytest -k test_sim
• Outcome: Reproduced — exact TypeError from the issue confirmed on Windows 10 / Python 3.11.5
• Logs:

TypeError: not all arguments converted during string formatting
msg = '  computational grid size:'
self = <LogRecord: root, 20, ...expand_grid_matrices.py, 268, "  computational grid size:">
logging/__init__.py:377: TypeError

• Note: Bug occurs in expand_grid_matrices.py:268 (and likely other kwave logger calls). Only triggers when running pytest -k test_sim alone, not when running full pytest suite, consistent with the original report.

[dblizinski]

@vaishnavidesai09, thank you. This is a genuinely thorough first pass. Reproducing it on a pinned environment (Win 10 / Python 3.11.5 / k-wave-python 0.4.0) with the exact error at expand_grid_matrices.py:268 is exactly the kind of triage that turns a bug from "reported" into "fixable." Adding the needs-triage:reproduced label now.

This one looks like a good candidate to take further if you're up for it. It seems self-contained (a logging call that only surfaces when
the test runs in isolation), so it'd make a clean first PR. Want to take a crack at the fix?