AutosarVersion case is inconsistent

[EdVanDance]

Hi 😃

I found an inconsistency with the AutosarVersion enum.

The docs and type annotations are using the same case as in rust, e.g. Autosar_00048 but it's actually all upper case:

autosar-data-py/src/version.rs

Line 38 in a546d05

#[pyo3(name = "AUTOSAR_00048")]

I thought about submitting a PR initially, but I wasn't sure how to handle this part:

autosar-data-py/src/version.rs

Lines 67 to 69 in a546d05

	fn __str__(&self) -> String {
	let ver: autosar_data_specification::AutosarVersion = (*self).into();
	ver.to_string()

This needs a conversion to upper case as well. Would .to_uppercase() do the trick?

Another small thing: 00053 and LATEST are missing in the type stubs.

[DanielT]

Hi,

You're right, the current state is rather bad.
Using .to_uppercase() in the __str__() method is correct.

I'd be happy to accept a PR from you if you want to fix these points.

[EdVanDance]

I had a look and now I'm confused ...

AutosarVersion.__str__() seems to be fine, since .to_string() ends up using describe(). I noticed that there is also a test for that:

autosar-data-py/test/version_test.py

Line 11 in a546d05

assert ver.__str__() == "AUTOSAR 4.0.1"

I assumed __str__() isn't working as expected as I saw some strange things while working on #8.

So I added some debug prints at the end of test_specification_sub_elements_spec():

    print(f"str:  {AutosarVersion.AUTOSAR_00050}")
    print(f"repr: {repr(AutosarVersion.AUTOSAR_00050)}")
    print()
    print(sub_element_spec)
    print()
    print(sub_element_spec.allowed_versions)
    assert False

Output:

str:  AUTOSAR R21-11
repr: AutosarVersion.AUTOSAR_00050

SubElementSpec { 
    element_name: "AR-PACKAGE",
    element_type: ElementType(ElementType { def: 349, typ: 207 }),
    allowed_versions: [
        Autosar_4_0_1, Autosar_4_0_2, Autosar_4_0_3, Autosar_4_1_1, Autosar_4_1_2,
        Autosar_4_1_3, Autosar_4_2_1, Autosar_4_2_2, Autosar_4_3_0, Autosar_00042,
        Autosar_00043, Autosar_00044, Autosar_00045, Autosar_00046, Autosar_00047,
        Autosar_00048, Autosar_00049, Autosar_00050, Autosar_00051, Autosar_00052,
        Autosar_00053
    ]
}

[AutosarVersion.AUTOSAR_4_0_1, AutosarVersion.AUTOSAR_4_0_2, AutosarVersion.AUTOSAR_4_0_3,
AutosarVersion.AUTOSAR_4_1_1, AutosarVersion.AUTOSAR_4_1_2, AutosarVersion.AUTOSAR_4_1_3,
AutosarVersion.AUTOSAR_4_2_1, AutosarVersion.AUTOSAR_4_2_2, AutosarVersion.AUTOSAR_4_3_0,
AutosarVersion.AUTOSAR_00042, AutosarVersion.AUTOSAR_00043, AutosarVersion.AUTOSAR_00044,
AutosarVersion.AUTOSAR_00045, AutosarVersion.AUTOSAR_00046, AutosarVersion.AUTOSAR_00047,
AutosarVersion.AUTOSAR_00048, AutosarVersion.AUTOSAR_00049, AutosarVersion.AUTOSAR_00050,
AutosarVersion.AUTOSAR_00051, AutosarVersion.AUTOSAR_00052, AutosarVersion.AUTOSAR_00053]

AutosarVersion string conversion and repr look good. But the string conversion of SubElementSpec is strange.
I don't get it 😄
Printing a SubElementSpec object seems to use the rust representation of the enum, but printing the allowed_versions property seems to use __repr__().

---

Well ... After writing all of this I noticed:

autosar-data-py/src/specification.rs

Lines 153 to 157 in a546d05

	impl SubElementSpec {
	fn __repr__(&self) -> String {
	format!("{self:?}")
	}
	}

That's probably why it looks like the rust enum 😄

Shall we simply accept the rusty SubElementSpec repr or do you want this to be consistent as well?

[DanielT]

Improving the display of all of the objects is an open point.
The current form had the benefit of being easy to implement, but is not great otherwise. This applies to both the rust crate and the python wrapper. For example, I think it's not useful in either case to completely flood the output with recursively printed sub-elements when displaying an Element.
Getting the case the SubElementSpec right was not on my todo list yet, but it is now.