Thank you for your quick response.
From your explanation, I now understand that if I can select the architectural elements through a CoordinateInput object, I can retrieve the transformation matrix using:

coordinate_input.GetSelectedElementAssocView().GetTransformationMatrix() -> Matrix3D

So, if I understand correctly, with this Matrix3D, calculating the 2D representation of any 3D point should be straightforward

However, a few points are still unclear:

1. Element selection
So far, I have been using the MultiElementSelectInteractor class to handle element selection:

self.selection_result = MultiElementSelectInteractorResult()
self.script_object_interactor = MultiElementSelectInteractor(
    self.selection_result,
    [AllplanEleAdapter.WallTier_TypeUUID],
    "Select walls"
) 

This approach provides the selection behavior I need, but it does not appear to expose its internal CoordinateInput object (Calling "self.script_object_interactor.__coord_input" draws an error). I attempted to subclass it; however, the API documentation only lists function names and arguments, making it difficult to understand how the selection is handled internally. More generally, it remains unclear how CoordinateInput operates within MultiElementSelectInteractor.
Is the source code for the Python API available by any chance?

2. Multiple associative views
Assuming I can select the desired architectural elements via a CoordinateInput class, how can I control which associative view is used?

For example, if the same wall appears in multiple sections (e.g., four different views), I would like to choose among those views. However, GetSelectedElementAssocView() only returns a single AssocViewElementAdapter, and it is not clear how that specific view is determined or how to influence the selection.

Thank you in advance for your help.

Thank you for the informative response.

Thanks to your help I was able to get the script working by implementing a custom interactor. By clicking on the architectural element inside the associated view, I can access the BaseElementAdapter of the section (with type UnifiedSection_TypeUUID) and perform the 3D-to-2D mapping using the TransformationMatrix property. This works well.

However, this led to another question. I would like to identify which architectural elements lie within the selected section (inside the box defined by the section bounds, see figure). This information does not appear to be directly accessible from the BaseElementAdapter that I previously accessed.
Is the geometry of the section's bounding box defined by a ClippingPathBody_TypeUUID object? If so, how can I access the ClippingPathBody_TypeUUID associated with a given UnifiedSection_TypeUUID?

Thank you for your help in advance.