Closest Point Calculation with RTree
Demonstrates how to perform a closest point calculation using an RTree data structure.
partial class Examples
{
static void SearchCallback(object sender, RTreeEventArgs e)
{
SearchData data = e.Tag as SearchData;
if (data == null)
return;
data.HitCount = data.HitCount + 1;
Point3f vertex = data.Mesh.Vertices[e.Id];
double distance = data.Point.DistanceTo(vertex);
if (data.Index == -1 || data.Distance > distance)
{
// shrink the sphere to help improve the test
e.SearchSphere = new Sphere(data.Point, distance);
data.Index = e.Id;
data.Distance = distance;
}
}
class SearchData
{
public SearchData(Mesh mesh, Point3d point)
{
Point = point;
Mesh = mesh;
HitCount = 0;
Index = -1;
Distance = 0;
}
public int HitCount { get; set; }
public Point3d Point { get; private set; }
public Mesh Mesh { get; private set; }
public int Index { get; set; }
public double Distance { get; set; }
}
public static Rhino.Commands.Result RTreeClosestPoint(RhinoDoc doc)
{
Rhino.DocObjects.ObjRef objref;
var rc = Rhino.Input.RhinoGet.GetOneObject("select mesh", false, Rhino.DocObjects.ObjectType.Mesh, out objref);
if (rc != Rhino.Commands.Result.Success)
return rc;
Mesh mesh = objref.Mesh();
objref.Object().Select(false);
doc.Views.Redraw();
using (RTree tree = new RTree())
{
for (int i = 0; i < mesh.Vertices.Count; i++)
{
// we can make a C++ function that just builds an rtree from the
// vertices in one quick shot, but for now...
tree.Insert(mesh.Vertices[i], i);
}
while (true)
{
Point3d point;
rc = Rhino.Input.RhinoGet.GetPoint("test point", false, out point);
if (rc != Rhino.Commands.Result.Success)
break;
SearchData data = new SearchData(mesh, point);
// Use the first vertex in the mesh to define a start sphere
double distance = point.DistanceTo(mesh.Vertices[0]);
Sphere sphere = new Sphere(point, distance * 1.1);
if (tree.Search(sphere, SearchCallback, data))
{
doc.Objects.AddPoint(mesh.Vertices[data.Index]);
doc.Views.Redraw();
RhinoApp.WriteLine("Found point in {0} tests", data.HitCount);
}
}
}
return Rhino.Commands.Result.Success;
}
}
Partial Friend Class Examples
Private Shared Sub SearchCallback(ByVal sender As Object, ByVal e As RTreeEventArgs)
Dim data As SearchData = TryCast(e.Tag, SearchData)
If data Is Nothing Then
Return
End If
data.HitCount = data.HitCount + 1
Dim vertex As Point3f = data.Mesh.Vertices(e.Id)
Dim distance As Double = data.Point.DistanceTo(vertex)
If data.Index = -1 OrElse data.Distance > distance Then
' shrink the sphere to help improve the test
e.SearchSphere = New Sphere(data.Point, distance)
data.Index = e.Id
data.Distance = distance
End If
End Sub
Private Class SearchData
Public Sub New(ByVal mesh As Mesh, ByVal point As Point3d)
Me.Point = point
Me.Mesh = mesh
HitCount = 0
Index = -1
Distance = 0
End Sub
Public Property HitCount() As Integer
Private privatePoint As Point3d
Public Property Point() As Point3d
Get
Return privatePoint
End Get
Private Set(ByVal value As Point3d)
privatePoint = value
End Set
End Property
Private privateMesh As Mesh
Public Property Mesh() As Mesh
Get
Return privateMesh
End Get
Private Set(ByVal value As Mesh)
privateMesh = value
End Set
End Property
Public Property Index() As Integer
Public Property Distance() As Double
End Class
Public Shared Function RTreeClosestPoint(ByVal doc As RhinoDoc) As Rhino.Commands.Result
Dim objref As Rhino.DocObjects.ObjRef = Nothing
Dim rc = Rhino.Input.RhinoGet.GetOneObject("select mesh", False, Rhino.DocObjects.ObjectType.Mesh, objref)
If rc IsNot Rhino.Commands.Result.Success Then
Return rc
End If
Dim mesh As Mesh = objref.Mesh()
objref.Object().Select(False)
doc.Views.Redraw()
Using tree As New RTree()
For i As Integer = 0 To mesh.Vertices.Count - 1
' we can make a C++ function that just builds an rtree from the
' vertices in one quick shot, but for now...
tree.Insert(mesh.Vertices(i), i)
Next i
Do
Dim point As Point3d = Nothing
rc = Rhino.Input.RhinoGet.GetPoint("test point", False, point)
If rc IsNot Rhino.Commands.Result.Success Then
Exit Do
End If
Dim data As New SearchData(mesh, point)
' Use the first vertex in the mesh to define a start sphere
Dim distance As Double = point.DistanceTo(mesh.Vertices(0))
Dim sphere As New Sphere(point, distance * 1.1)
If tree.Search(sphere, AddressOf SearchCallback, data) Then
doc.Objects.AddPoint(mesh.Vertices(data.Index))
doc.Views.Redraw()
RhinoApp.WriteLine("Found point in {0} tests", data.HitCount)
End If
Loop
End Using
Return Rhino.Commands.Result.Success
End Function
End Class
import Rhino
import rhinoscriptsyntax as rs
# data passed to the RTree's SearchCallback function that
# we can use for recording what is going on
class SearchData:
def __init__(self, mesh, point):
self.HitCount = 0
self.Mesh = mesh
self.Point = point
self.Index = -1
self.Distance = 0
def SearchCallback(sender, e):
data = e.Tag
data.HitCount += 1
vertex = data.Mesh.Vertices[e.Id]
distance = data.Point.DistanceTo(vertex)
if data.Index == -1 or data.Distance > distance:
# shrink the sphere to help improve the test
e.SearchSphere = Rhino.Geometry.Sphere(data.Point, distance)
data.Index = e.Id
data.Distance = distance
def RunSearch():
id = rs.GetObject("select mesh", rs.filter.mesh)
mesh = rs.coercemesh(id)
if mesh:
rs.UnselectObject(id)
tree = Rhino.Geometry.RTree()
# I can add a RhinoCommon function that just builds an rtree from the
# vertices in one quick shot, but for now...
for i,vertex in enumerate(mesh.Vertices): tree.Insert(vertex, i)
while(True):
point = rs.GetPoint("test point")
if not point: break
data = SearchData(mesh, point)
# Use the first vertex in the mesh to define a start sphere
distance = point.DistanceTo(mesh.Vertices[0])
sphere = Rhino.Geometry.Sphere(point, distance * 1.1)
if tree.Search(sphere, SearchCallback, data):
rs.AddPoint(mesh.Vertices[data.Index])
print "Found point in {0} tests".format(data.HitCount)
if __name__=="__main__":
RunSearch()