The entities of two goups are not completely overlapping, only partially face overlapping, but the subtract operation will also result in nil. The subtract results presented in the Sketchup are different every time. 无标题.skp (874.3 KB)
Maybe this should include an instance, not a definition.
Entities is not something unique in your model… you should specify a instance from the definition
Here, … entities is the reference to the "Test" definition’s Entities collection, which at this point has no Drawingelement < Entity members.
Directly after these statements the code added a Group instance to this "Test" definition’s entities collection (which also transparently caused a parent definition for the group instance to be created in the model’s DefinitionList collection.)
Then the wall_group instance was passed to the build_group() method, which did bunch of calculating before using the group that was passed into the method.
@carol I suggest you immediately add a cpoint to your groups at their origin to prevent SketchUp from deleting your empty groups. So, in to_su_wall() you should do …
def to_su_wall(wall)
begin
comp_def = Sketchup.active_model.definitions.add "Test"
entities = comp_def.entities
wall_group = entities.add_group
cpt = wall_group.definition.entities.add_cpoint(ORIGIN)
build_group(wall_group, wall.build_element)
openings = wall.openings
wall_group = build_openings_group(openings, entities, wall_group)
cpt.erase!
wall_group
rescue => e
e.message
end
end
Your code can erase the cpoint when it is no longer needed.
Yes. Finally, I will generate an instance based on this definition.
I have added the code like this：
trans = Geom::Transformation.new(ORIGIN)
instance = Sketchup.active_model.active_entities.add_instance(comp_def, trans)
But my question is,
Why does the subtract() have nil or unexpected results. For example, sometimes the result is nil, sometimes the result obtained is one or two faces, and sometimes the result obtained in A-B looks like B-A. This may be because the entities in my two groups may have overlapping faces, but how can I avoid this situation.
The second question is,
When I try to use copy() to backup wall_group, and then reassign the wall_group_copy to wall_group when I obtain the result of NIL. This is to avoid the exception of group being deleted and resulting in nil after subtract(), but it was found that both wall_group and wall_group_copy were deleted at the same time.
Thanks for your guidance,
But after I added the line of cpt=…, there will still be calculation errors of subtract() or the wall_group_copy that copy from wall_group being deleted
Sorry, I objectively considered it to be a non-manifold entity. Thank you for your guidance. I will add some checks and try again.
However, I noticed that the documentation mentions that ‘copy’ would return a new group. However, why is it that when ‘subtract’ encounters a ‘nil’ value on the original group, both the original group and its copy are deleted?
At the same time, as shown in the .skp document and screenshot, I expected to achieve the effect of subtracting the smaller entity from the larger entity using the ‘subtract’ method. But the results are always inconsistent, with different outcomes for the remaining shapes in the file, all of which were processed using the same data and method.
Thank you for your answer.
Firstly, I have two non-manifold groups, wall_group and opening_group. Whenever I attempt to subtract opening_group from wall_group using opening_group.subtract(wall_group), the results vary unpredictably. I suspect this is because of precision issues with overlapping faces, which sometimes leaves behind a single face after clipping.
Secondly, in the SU 2020, let’s assume I have a manifold group called a_group and a non-manifold group named b_group. Before performing the subtraction, I create a backup of a_group called a_group_copy. However, I’ve noticed that when I execute b_group.subtract(a_group), if the result is nil, both a_group and its backup a_group_copy are unexpectedly deleted.
