refactor: do not import ansys.tools.visualizer by default when importing ansys.mapdl.core

[koubaa]

Attempting to import a module to check if it is installed is unnecessary. If there is a problem with a module that is installed, it will manifest on import and not on use, which can lead to errors even when that module is never used.

This doesn't fully fix the problem because _apply_default_theme is still used even when plotting isn't done

[ansys-reviewer-bot]

Thanks for opening a Pull Request. If you want to perform a review write a comment saying:

@ansys-reviewer-bot review

[codecov]

Codecov Report

Attention: Patch coverage is 96.55172% with 1 line in your changes missing coverage. Please review.

Project coverage is 88.28%. Comparing base (1ec7e5c) to head (40d6c9e).
Report is 2 commits behind head on main.

Additional details and impacted files

@@            Coverage Diff             @@
##             main    #3887      +/-   ##
==========================================
- Coverage   88.29%   88.28%   -0.02%     
==========================================
  Files         187      187              
  Lines       14819    14830      +11     
==========================================
+ Hits        13084    13092       +8     
- Misses       1735     1738       +3     

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.

[germa89]

@koubaa it seems you still need the try and except context manager:

https://github.com/ansys/pymapdl/actions/runs/14777942416/job/41491532825?pr=3887#step:10:92

GitHub

refactor: change how the install check works · ansys/pymapdl@edb6b18

A Python client library for Ansys MAPDL. Contribute to ansys/pymapdl development by creating an account on GitHub.

[germa89]

@koubaa

Thanks to your PR I realised that there a bit of code duplicity. I will take care of that in a follow up PR. Additionally what I meant with cache is what is shown here:

pymapdl/src/ansys/mapdl/core/misc.py

Line 420 in edb6b18

def is_package_installed_cached(package_name):

functools.cache can help us to cache functions so we only checks import (an their try/except once).

[germa89]

You might want to remove this:

pymapdl/src/ansys/mapdl/core/plotting/visualizer.py

Lines 46 to 51 in 8ca84ba

	if _HAS_VISUALIZER:
	import pyvista as pv
	from ansys.mapdl.core.plotting.plotting_defaults import DefaultSymbol
	BC_plot_settings = DefaultSymbol()

[koubaa]

This change has grown because the visualization interface is loaded on startup in multiple places, and I've modified it not to load on startup at all. This may break some scripts, but none of the examples we distribute are affected

[germa89]

Ping me when this is ready (CICD is failing at the moment)

[koubaa]

@germa89 I fixed the tests but there is one failure in autodoc that I don't understand:

WARNING: autodoc: failed to import class 'MapdlPlotter' from module 'ansys.mapdl.core.plotting'; the following exception was raised:

I made sure that all the references to MapdlPlotter use the full namespace in the pymapdl repo. Does autodoc pull from other repos too?

[germa89]

@koubaa the documentation failure was due to using previous build cache. I cleaned it up using this workflow and now it seems fine.

Additionally, codacy is showing this warning, but I guess it is fine... so I will disable it.

[germa89]

I am not particularly happy with having variables in a module to be initialized after the first plot.

from ansys.mapdl.core.plotting import MapdlPlotter
from ansys.mapdl.core.plotting import BC_plot_settings # Fails
# plot somethin
pl = MapdlPlotter()
from ansys.mapdl.core.plotting import BC_plot_settings # DOES NOT Fails

Which is why codacy raise the issue. This behaviour to me is an antipattern. However, if this is the only way to get rid of that bug we mentioned, let it be until fix that upstream.

== EDIT ==

Maybe we can just define BC_plot_settings in the module globally as None or a non-initialized class... So we get rid of the codacy warning, and we have a consistent module API.