Python: Copy visual transform from object with constraints

Question

i am working with an alembic file which contains points and movement information. for project reasons, i need the movement information in keyframes, so i am writing myself a script that grabs for a given frame the transformation and stores it in a separate object and inserts a keyframe.

problem is, the movement of the point is not represented in the transform window.

so my solution for that problem is the "Visual Transform" action, as shown manually below

though, when i try to perform the action target.visual_transform_apply(), i get the error message AttributeError: 'Object' object has no attribute 'visual_transform_apply'

why does it work doing it by hand, and it doesn't when performing it with code?

here is a reduced code snippet of my problem.

target = bpy.context.active_object

bpy.ops.mesh.primitive_plane_add()
new_obj = bpy.context.active_object

target.visual_transform_apply()     

new_obj.location[0] = target.location[0]
new_obj.location[1] = target.location[1]
new_obj.location[2] = target.location[2]

new_obj.keyframe_insert(data_path = "location", frame = 1)

my alembic testfile, i hope a filehoster link is fine https://rapidgator.net/file/d4d4f04912c8fb24a0f716276fa11ca9/Export_01a.abc.html

Answer 1

found a solution for my own problem.

the transformation tooltip for python didn't help me here.

somewhere else i found this line of code

new_obj.location[0] = target.matrix_world.translation[0]

this returned me the actual vector of the transformation. by accessing the items of the array one by one, you can work with each value individually.

here is the code snippet from the beginning fixed.

def set_active_object(obj):
    bpy.context.view_layer.objects.active = obj

target = bpy.context.active_object

bpy.ops.mesh.primitive_plane_add()
new_obj = bpy.context.active_object

set_active_object(target)
bpy.ops.object.visual_transform_apply()
set_active_object(new_obj)
    
new_obj.location[0] = target.matrix_world.translation[0]
new_obj.location[1] = target.matrix_world.translation[1]
new_obj.location[2] = target.matrix_world.translation[2]

new_obj.keyframe_insert(data_path = "location", frame = 1)

And if anybody stumbles across here, looking for how to get the rotation and scale value, here they are:

new_obj.rotation_euler[0] = target.matrix_world.to_euler('XYZ')[0]
new_obj.rotation_euler[1] = target.matrix_world.to_euler('XYZ')[1]
new_obj.rotation_euler[2] = target.matrix_world.to_euler('XYZ')[2]

new_obj.scale[0] = my_target.matrix_world.to_scale()[0]
new_obj.scale[1] = my_target.matrix_world.to_scale()[1]
new_obj.scale[2] = my_target.matrix_world.to_scale()[2]

This is by far not the cleanest code, i apologize for that. it's my first blender script.

Answer 2

I just had to do this for one of my scripts to duplicate an object and remove any constraints/armature settings.

My code is below (Blender 3.x) that does what I need it to for now: obj is the source mesh object to duplicate, o_tmp is the duplicated copy with the visual transformation of the source object applied to it.

        o_tmp = obj.copy()
        o_tmp.data = obj.data.copy()
        o_tmp.animation_data_clear()
        bpy.context.collection.objects.link(o_tmp)
        o_tmp.parent = None

        dg = bpy.context.evaluated_depsgraph_get()
        o = obj.evaluated_get(dg)
        m = o.matrix_world.copy()
        o_tmp.matrix_world = m