As per the discussion we think that the dependencies are not versioned which is also problematic. I think we should remove this change from all example scripts.

cc: @seanpearsonuk @mkundu1 @prmukherj

@hpohekar It looks like this is fairly common practice. I think that the pyfluent dependency should be given as ansys-fluent-core though.

@Gobot1234

@seanpearsonuk @Gobot1234

We can specify the dependencies required to run all examples in a section of pyproject.toml as follows:

[project.optional-dependencies]
doc = [
...
]
tests = [
...
]
additional = [
...
]
examples = [
...
]

We can install them as follows:

pip install ansys-fluent-core[examples]

This will remove duplicate boilerplate from all example scripts and helpful from the maintenance point of view.

We can add an instruction in docs to execute pip install ansys-fluent-core[examples] before running any example scripts locally.

cc: @mkundu1

@hpohekar

I believe @Gobot1234’s intent is to consider the real-world user experience: someone who already has PyFluent installed, comes across this script in isolation, and simply wants to know what additional packages are required to run it successfully.

If this script were circulating outside the repository, having the dependency list inside the script itself would be helpful and self-explanatory.

The question then is: at what point does it stop being useful to include that information? If we remove it purely because the script is now “in the repo,” we risk making it less accessible when encountered independently.

I’d argue that keeping lightweight dependency metadata in examples is a net positive for usability unless we have a stronger alternative convention.

@seanpearsonuk

Yes, it makes sense, considering the points of circulating script outside the repository and its independent accessibility.

I have also verified that Sphinx recognizes this format and does not render it in the generated examples gallery.

cc: @Gobot1234

This looked sus (moving the code around)

👍

👍