[Supplies] Custom Scripts and Other Stuff - Requests and Downloads

[Mr. Bluesummers-3DT]

What does the error code in the little pink box in the lower left say?

[poopipe]

dear Santa,
I would like a script that splits edges according to it's UV pelt boundaries/those blue seams you can put in manually (whatever they're called) for christmas please.

I've been good honest

love from poo

[Mr. Bluesummers-3DT]

Juicy one, my little Poopie. Does it have to be the pelt seams, or can it be just UVW clusters? Let me look into this one and get back to you...

[Mr. Bluesummers-3DT]

Originally Posted by Pickman

Well thanks again for trying. Ok lets try this again.

Script idea:
Ultimate Detach-detached objects maintain the parents stack.
For example if I detach the horns from my model they still have symmetry and turbosmooth(at 2 iterations with isoline). It'll save time if i don't have to re-apply those. And if it would work with skin, physique, or both then it would be very usefull I'm sure.

I'm thinking about this one too. How would this go beyond replicating the parent's modifier stack in the detatched objects? Are there special procedures that need to happen? How much time will this save?

As far as skin/physique goes, getting things to sync back up with detached objects could be pretty wicked depending on how dirty the detachment process is. This requires a little MXS researching...

[poopipe]

for the UVs either works for me, they amount to the same thing anyway


for that detach one...
From my experience skin will probably work out fine even if you dont bake the vertex weights before reattaching - detaching won't cause any problems, Physique will almost certainly give you a lot of grief because it's rubbish, unwraps are generally fine with this sort of behaviour and any topologically independent modifier should be fine.

The thing is, while it's an interesting technical challenge it does beg the question "why bother?". almost everything you'd want it to do can be achieved manually in a couple of clicks - if the script were running off a toolbar it probably wouldn't save you anything.

no offence meant pickman - it just seems a little unnecessary

[Mr. Bluesummers-3DT]

Ugh; the pelt seam conversion to edge selection is all kinds of broken. It appears in the viewport, but for all scripting purposes, this function is a total lie. I'll get back to you Poopipe.

[poopipe]

that probably explains why the seams don't persist through unwrap modifiers - it could well be a dead end.

I'd go with the UV clusters then

[Mr. Bluesummers-3DT]

Here're the two ways I see your proposal working, Poopster:
1. Get the edge selection from the pelt seams via

$.modifiers[#unwrap_uvw].getPeltSelectedSeams()

and then convert that into a geometry edge selection, and then use the split tool.
>> Problem: The edge seam bit array returned by this function is wrong. The edge IDs returned aren't like what the rest of the UVW modifier or object are expecting. It comes off garbled.

2. Use the built-in conversion from pelt seam to edge selection:

$.modifiers[#unwrap_uvw].unwrap5.peltSeamToEdgeSel true

and then store the UVW edge selection to a bit array that can be brought into the base PolyObject and converted into an edge selection. Then we can use the split tool.
>> Problem: This function doesn't create a real edge selection. It's the weirdest thing ever. The viewport shows an edge selection, but any other commands act on whatever edge selection was in place BEFORE you tried this command.

This is total BS. I'm starting to think this feature doesn't really work at all.

[Mr. Bluesummers-3DT]

Yah, you're probably right. Selecting UVW cluster edges is probably a more feasible option; but this one was definitely easier if these functions worked.

[Mr. Bluesummers-3DT]

Filed a report on the Area forums. Planning on getting ignored by Autodesk again.

[Mr. Bluesummers-3DT]

I've written out the entire script, the only component missing is getting edges from the UVW selection window to a useful bit array. Below is the code as it stands...

-- Split by UVW cluster by Mr. Bluesummers
-- Not sea-worthy.  Don't use this.

undo on
(
	local NumFaces = PolyOp.getNumFaces $
	local AllFaces = #{1..NumFaces}
	local FacesCompleted = #{}
	local Finished = false

	while not Finished do
	(
		-- Get the next face to test
		NextFace = 1

		while FacesCompleted[NextFace] do 
			NextFace += 1
		
		-- Get a face in an element.
		$.modifiers[#unwrap_uvw].unwrap2.selectFaces #{NextFace}
		
		-- Select the whole element, and convert to edges.
		$.modifiers[#unwrap_uvw].unwrap2.selectElement()
		$.modifiers[#unwrap_uvw].unwrap2.faceToEdgeSelect()
		
		-- Get the current selection.  Get a new selection ready.
		CurrentSelection = $.modifiers[#unwrap_uvw].unwrap2.getSelectedEdges()
		NewSelection = #{}
		
		-- Select half the edges
		for b in CurrentSelection do
			if (mod b 2) == 0.0 do
				NewSelection[b] = true
		
		-- Assign the new, half-selection of edges.
		$.modifiers[#unwrap_uvw].unwrap2.selectEdges NewSelection
		
		-- Select open edges
		$.modifiers[#unwrap_uvw].unwrap2.openEdgeSelect()
		
		-- Get the current selection again, prep another new selection.
		CurrentSelection = $.modifiers[#unwrap_uvw].unwrap2.getSelectedEdges()
		NextNewSelection = #{}
		
		-- Subtract the original half-selection from this current selection.
		for b in CurrentSelection do
			if not NewSelection[b] do
				NextNewSelection[b] = true
			
		-- Set the selection with the result, and redo the open edge selection.
		$.modifiers[#unwrap_uvw].unwrap2.selectEdges NextNewSelection
		$.modifiers[#unwrap_uvw].unwrap2.openEdgeSelect()
		
		-- We now have the cluster edges.
		-- Get edges with UVW-based IDs.
		ClusterEdges = $.modifiers[#unwrap_uvw].unwrap2.getSelectedEdges()
			
		-- Get the edges with GEOMETRY-based IDs. FAILS!!
		-->> ClusterEdges = $.modifiers[#unwrap_uvw].unwrap5.getSelectedGeomEdges()
			
		-- Convert this to a geometry edge selection and split.
		-->> PolyOp.splitEdges $ ClusterEdges
		
		-- Add the current cluster to the completion list.
		FacesCompleted += $.modifiers[#unwrap_uvw].unwrap2.getSelectedFaces()
			
		-- If the completion list is equal to all faces, we're done!
		if FacesCompleted.numberset == AllFaces.numberset do Finished = true
	)
)

I'm sorry, but without a way to convert UVW edge IDs to geometry edge IDs, there's no way to make this work. I already tried the vertex workaround, and it fails for triangles.

[poopipe]

hmm. thats a bit of a bugger.

getting the edge selection is the time consuming part and your code does that bit so it's not a dead loss

there could be a hack here somewhere though since manually adding an edit poly on top of an unwrapUVW (with an active selection) carries the selection up the stack.

[discmage]

Is there a way to select all edges that have different smoothing groups on either side? if not, would that be hard to script?

[Astaroth]

there's someway to add a modifier for the editable poly(even in the quad menu) that works as a push modifier on the fly, without adding the modifier itself?
it should push the selected vertex in the direction of their normal, is it hard to do???

[poopipe]

just use normal constraint in edit poly - its there in max 2008 or above