Connect Clusters

PCGEx | Cluster : Connect

Connects clusters together.

The Connect Clusters node creates bridge edges between clusters using methods like Delaunay, Least Edges, or Most Edges, always connecting the closest points between clusters to form a larger, unified cluster.

Vtx

Endpoints of the input Edges

Edges

Edges associated with the input Vtxs

Vtx

Endpoints of the output Edges

Edges

Edges associated with the output Vtxs

Working with Clusters

Table of content

Properties

The Connect Clusters creates “bridge” edges between all input clusters and outputs a single bigger one as a result.

details/clusters-connect/lead.png

Properties

Property	Description
Bridge Method	The method that will be used to identify and create bridges between clusters.

Note that no matter what method is selected, a bridge will always connect the two closest points between two clusters. The chosen method only drives which cluster is connected to which other cluster.

Bridge Methods

Method	Description
	Delaunay When using this method, each cluster is abstracted into a single bounding box that encapsulates all its vertices. A 3D Delaunay is generated using each bounding box center as an input, and the resulting delaunay edges are used as bridges.
	Least Edges When using this method, the algorithm will generate the least possible amount of bridge in order to connect all the clusters together. Careful because it can easily look like a minimum spanning tree, but it’s not.
	Most Edges When using this method, the algorithm will create a bridge from each cluster to every other cluster.

Projection Settings

The projection settings control how the point position is translated to a 2D space before the graph is computed; and how this projection will translate back to the original space, if relevant.

Property	Description
Projection Normal	Normal vector of the plane used for projection. By default, the projection plan normal is `Up`; so the graph is computed over the `X Y` plane.
Local Projection Normal	If enabled, uses a per-point projection vector.
Local Normal	Attribute ti read normal from,

Local projection normal is very powerful but can also be very clunky to use – it’s very easy to end up with singularities that will prevent the graph from being properly computed.

Carry Over Settings

Since this node merges multiple inputs together, “Carry over” settings lets you pick-and-choose which attributes & tags carry over to the new data.

Property	Description
Preserve Attributes Default Value	If enabled, the node will attempt to create attributes on the data in a way that preserve the original, underlying default value of the attribute. This can be critical in order to identify which data originally belonged to which, as well as properly initializing flags to a desirable default.
Attributes	Lets you pick and choose which attributes to keep or dismiss.
Tags	Lets you pick and choose which attributes to keep or dismiss.

Both Attributes & Tags share the same string-based filters.

Note that the filters look for a single valid match amongst the list; you cannot create and/or conditions.

Filter Details

Property	Description
Filter Mode	Chooses how the filter operates. - `All` let everything pass. - `Exclude` filters out the result of the filter. - `Include` only allows the items that pass the filters.
Matches	Lets you define a list of checks pairs: a string value, and a `Match Mode`.
Comma Separated Names	Easy to override, lets you specify a list of comma-separated names. The only caveat is that you can only pick a unique match mode used for each entry.
Comma Separated Names Filter	Which filter will be used along the Comma Separated Names.
Preserve PCGEx Data	Most of the time you’ll want to leave it to its default value. It ensures `PCGEx/` prefixed data are not captured by the filter.

Filter Modes

Mode	Description
Equals	Checks for strict equality of the filtered value and the associated string.
Contains	Checks if the filtered value contains the associated string.
Starts With	Checks if the filtered value is prefixed with the associated string.
Ends With	Checks if the filtered value is suffixed with the associated string.

Cluster Output Settings

See Working with Clusters.

Property	Description
Edge Position	If enabled, edge point’ position will be the result of that value used as a lerp between its start and end `Vtx` point.

Pruning
Min Vtx Count	If enabled, only ouputs clusters that have more ( > ) `Vtx` than the specified number.
Max Vtx Count	If enabled, only ouputs clusters that have less ( < ) `Vtx` than the specified number.
Min Edge Count	If enabled, only ouputs clusters that have more ( > ) `Edge` than the specified number.
Max Edge Count	If enabled, only ouputs clusters that have less ( < ) `Edge` than the specified number.

Misc
Refresh Edge Seed	If enabled, `Edge` points gets a fresh seed.
Build and Cache Cluster	If enabled, pre-build and cache cluster along with the point data. This has a slight memory cost associated to it, but can offer tremendous performance improvement. If disabled, cluster processors that comes down the line have to rebuild clusters from point data, which is very costly as they are also tested for errors and possible disconnections in the process.
Expand Clusters	If enabled, also build & cache another layer of cache data. This can have a significant memory cost, as well as a minimal performance overhead, but can greatly improve certain specific operations down the line.