Replace GTS with CDT #617

mjcarroll · 2024-07-24T19:15:27Z

This removes the dependency on GTS.

It removes the Mesh boolean operations functions, which were part of the building editor in Gazebo-classic but are unused in modern gazebo.
It replaces the GTS delaunay triangulation with the header-only CDT library.

Signed-off-by: Michael Carroll <[email protected]>

scpeters

some small whitespace nits

and I've prepared a change for homebrew-simulation for after this lands: osrf/homebrew-simulation#2704

CMakeLists.txt

graphics/src/MeshManager.cc

azeey

The MeshCSG code was used in Gazebo-classic to implement the Building Editor. Since there is no plan to port this to modern Gazebo (we can potentially use rmf_site for this in the future), I'd say let's go ahead and remove it from gz-common.

azeey · 2024-08-05T21:56:21Z

graphics/src/CDT/CDT.h

Can we put CDT in a vendor subdirectory with the a README to indicate its source and version info?

iche033 · 2024-08-12T23:26:40Z

gts dep removal in the gz-common6-release repo when this lands: gazebo-release/gz-common6-release#4

iche033 · 2024-08-14T00:55:35Z

I tested this with the polylines.sdf world and I think some of the triangles may be oriented incorrectly, i.e. normals are in the opposite direction:

left: GTS vs right: CDT

graphics/include/gz/common/MeshManager.hh

scpeters · 2024-08-15T20:00:16Z

I tested this with the polylines.sdf world and I think some of the triangles may be oriented incorrectly, i.e. normals are in the opposite direction:

I added a printout to the DelaunayTriangulation test to show the vertices for the triangles:

patch

diff --git a/graphics/src/DelaunayTriangulation_TEST.cc b/graphics/src/DelaunayTriangulation_TEST.cc
index df919d3..c441257 100644
--- a/graphics/src/DelaunayTriangulation_TEST.cc
+++ b/graphics/src/DelaunayTriangulation_TEST.cc
@@ -69,6 +69,32 @@ TEST_F(DelaunayTriangulation, DelaunayTriangulation)
   // same as number of vertices in the path
   EXPECT_EQ(subMesh.VertexCount(), 8u);
 
+  std::cerr << subMesh.Max() << '\n';
+  std::cerr << subMesh.Min() << '\n';
+  // Print vertices
+  std::cerr << "vertices: \n";
+  for (int v = 0; v < subMesh.VertexCount(); ++v)
+  {
+    std::cerr << subMesh.Vertex(v) << '\n';
+  }
+  // Print normals
+  std::cerr << "normals: \n";
+  for (int v = 0; v < subMesh.NormalCount(); ++v)
+  {
+    std::cerr << subMesh.Normal(v) << '\n';
+  }
+
   // there should be 8 triangles.
   EXPECT_EQ(subMesh.IndexCount() / 3u, 8u);
+  // Print triangles
+  std::cerr << "triangles: \n";
+  for (int i = 0; i < subMesh.IndexCount() / 3u; ++i)
+  {
+    int i1 = subMesh.Index(i*3 + 0);
+    int i2 = subMesh.Index(i*3 + 1);
+    int i3 = subMesh.Index(i*3 + 2);
+    std::cerr << '[' << subMesh.Vertex(i1) << "]\n"
+              << '[' << subMesh.Vertex(i2) << "]\n"
+              << '[' << subMesh.Vertex(i3) << "]\n\n";
+  }
 }

output with GTS

[ RUN      ] GTSMeshUtils.DelaunayTriangulation
1 1 0
0 0 0
vertices: 
0.75 0.75 0
1 0 0
1 1 0
0.75 0.25 0
0.25 0.75 0
0 1 0
0.25 0.25 0
0 0 0
normals: 
triangles: 
[0.75 0.75 0]
[1 1 0]
[1 0 0]

[0.75 0.25 0]
[0.75 0.75 0]
[1 0 0]

[0.25 0.75 0]
[0.25 0.25 0]
[0 1 0]

[0.75 0.25 0]
[1 0 0]
[0.25 0.25 0]

[1 0 0]
[0 0 0]
[0.25 0.25 0]

[0.25 0.25 0]
[0 0 0]
[0 1 0]

[0.75 0.75 0]
[0.25 0.75 0]
[1 1 0]

[0.25 0.75 0]
[0 1 0]
[1 1 0]

[       OK ] GTSMeshUtils.DelaunayTriangulation (1 ms)

output with CDT

[ RUN      ] DelaunayTriangulation.DelaunayTriangulation
1 1 0
0 0 0
vertices: 
0 0 0
1 0 0
1 1 0
0 1 0
0.25 0.25 0
0.25 0.75 0
0.75 0.75 0
0.75 0.25 0
normals: 
triangles: 
[0.75 0.75 0]
[1 1 0]
[0 1 0]

[0.25 0.75 0]
[0.75 0.75 0]
[0 1 0]

[0.25 0.25 0]
[0 1 0]
[0 0 0]

[0.75 0.25 0]
[0.25 0.25 0]
[1 0 0]

[0.25 0.75 0]
[0 1 0]
[0.25 0.25 0]

[0.75 0.25 0]
[1 0 0]
[0.75 0.75 0]

[1 0 0]
[1 1 0]
[0.75 0.75 0]

[1 0 0]
[0.25 0.25 0]
[0 0 0]

[       OK ] DelaunayTriangulation.DelaunayTriangulation (1 ms)

iche033 · 2024-08-15T20:09:12Z

I tested this with the polylines.sdf world and I think some of the triangles may be oriented incorrectly, i.e. normals are in the opposite direction:

I added a printout to the DelaunayTriangulation test to show the vertices for the triangles:

patch

output with GTS

output with CDT

Thanks, I created a fix for this in #623. It also addresses other feedback in this PR.

* Fix triangle dir, lint, style * update migration guide * Test that Delaunay triangles wind clockwise (#624) --------- Signed-off-by: Ian Chen <[email protected]> Signed-off-by: Steve Peters <[email protected]> Co-authored-by: Steve Peters <[email protected]>

codecov · 2024-08-15T22:52:51Z

Codecov Report

Attention: Patch coverage is 60.56884% with 610 lines in your changes missing coverage. Please review.

Project coverage is 79.32%. Comparing base (316ea33) to head (b852307).
Report is 6 commits behind head on main.

Files	Patch %	Lines
graphics/src/CDT/Triangulation.hpp	63.66%	274 Missing ⚠️
graphics/src/CDT/KDTree.h	28.49%	128 Missing ⚠️
graphics/src/CDT/portable_nth_element.hpp	53.71%	56 Missing ⚠️
graphics/src/CDT/Triangulation.h	46.73%	49 Missing ⚠️
graphics/src/CDT/CDTUtils.hpp	61.20%	45 Missing ⚠️
graphics/src/CDT/predicates.h	76.68%	45 Missing ⚠️
graphics/src/DelaunayTriangulation.cc	82.14%	5 Missing ⚠️
graphics/src/CDT/CDTUtils.h	91.30%	4 Missing ⚠️
graphics/src/CDT/LocatorKDTree.h	76.47%	4 Missing ⚠️

Additional details and impacted files

@@            Coverage Diff             @@
##             main     #617      +/-   ##
==========================================
- Coverage   80.54%   79.32%   -1.22%     
==========================================
  Files          80       87       +7     
  Lines       13588    14877    +1289     
==========================================
+ Hits        10944    11801     +857     
- Misses       2644     3076     +432

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

scpeters · 2024-08-15T23:01:38Z

graphics/src/DelaunayTriangulation_TEST.cc

  EXPECT_TRUE(result);

  // same as number of vertices in the path
  EXPECT_EQ(subMesh.VertexCount(), 8u);

  // there should be 8 triangles.
  EXPECT_EQ(subMesh.IndexCount() / 3u, 8u);
+
+  // verify that triangles have clockwise winding
+  for (int t = 0; t < subMesh.IndexCount() / 3u; ++t)


Suggested change

for (int t = 0; t < subMesh.IndexCount() / 3u; ++t)

for (unsigned int t = 0; t < subMesh.IndexCount() / 3u; ++t)

apologies I missed the compiler warning in #624

scpeters · 2024-08-15T23:03:56Z

I think we could ignore graphics/src/CDT in codecov.yml

Signed-off-by: Ian Chen <[email protected]>

iche033 · 2024-08-16T00:12:40Z

I think we could ignore graphics/src/CDT in codecov.yml

I updated the ignore files in 7ddc32a

scpeters

forgot one thing: we can remove the following line

https://github.com/gazebosim/gz-common/blob/mjcarroll/cdt/.github/ci/packages.apt#L8

Signed-off-by: Steve Peters <[email protected]>

Signed-off-by: Ian Chen <[email protected]>

iche033 · 2024-08-16T20:17:35Z

removed gts dep from package.xml in a159b79

Replace GTS with CDT

93f0118

Signed-off-by: Michael Carroll <[email protected]>

mjcarroll self-assigned this Jul 24, 2024

github-actions bot added the 🏛️ ionic label Jul 24, 2024

Restrict warnings to clang/gcc

bb52cce

Signed-off-by: Michael Carroll <[email protected]>

azeey added the beta Targeting beta release of upcoming collection label Jul 29, 2024

scpeters mentioned this pull request Jul 29, 2024

gz-common6: remove gts dependency osrf/homebrew-simulation#2704

Merged

scpeters reviewed Jul 29, 2024

View reviewed changes

CMakeLists.txt Show resolved Hide resolved

CMakeLists.txt Show resolved Hide resolved

graphics/src/MeshManager.cc Show resolved Hide resolved

graphics/src/MeshManager.cc Show resolved Hide resolved

azeey reviewed Aug 5, 2024

View reviewed changes

azeey mentioned this pull request Aug 6, 2024

Added MeshCSG test #469

Closed

8 tasks

iche033 mentioned this pull request Aug 12, 2024

Remove gts dependency gazebo-release/gz-common6-release#4

Merged

scpeters reviewed Aug 15, 2024

View reviewed changes

graphics/include/gz/common/MeshManager.hh Show resolved Hide resolved

iche033 and others added 2 commits August 15, 2024 15:44

Merge branch 'main' into mjcarroll/cdt

b852307

scpeters reviewed Aug 15, 2024

View reviewed changes

update ignore files

7ddc32a

Signed-off-by: Ian Chen <[email protected]>

scpeters approved these changes Aug 16, 2024

View reviewed changes

scpeters requested changes Aug 16, 2024

View reviewed changes

scpeters marked this pull request as ready for review August 16, 2024 17:25

scpeters requested a review from marcoag as a code owner August 16, 2024 17:25

packages.apt: remove its

41b58cb

Signed-off-by: Steve Peters <[email protected]>

scpeters approved these changes Aug 16, 2024

View reviewed changes

iche033 approved these changes Aug 16, 2024

View reviewed changes

remove gts from package.xml

a159b79

Signed-off-by: Ian Chen <[email protected]>

iche033 merged commit 1f52701 into main Aug 16, 2024
10 checks passed

iche033 deleted the mjcarroll/cdt branch August 16, 2024 21:26

traversaro mentioned this pull request Sep 29, 2024

Remove gts references from documentation #645

Merged

8 tasks

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Replace GTS with CDT #617

Replace GTS with CDT #617

mjcarroll commented Jul 24, 2024

scpeters left a comment

azeey left a comment

azeey Aug 5, 2024

scpeters Aug 15, 2024

iche033 commented Aug 12, 2024

iche033 commented Aug 14, 2024

scpeters commented Aug 15, 2024

iche033 commented Aug 15, 2024

patch

output with GTS

output with CDT

codecov bot commented Aug 15, 2024

scpeters Aug 15, 2024

scpeters commented Aug 15, 2024

iche033 commented Aug 16, 2024

scpeters left a comment

iche033 commented Aug 16, 2024

	for (int t = 0; t < subMesh.IndexCount() / 3u; ++t)
	for (unsigned int t = 0; t < subMesh.IndexCount() / 3u; ++t)

Replace GTS with CDT #617

Replace GTS with CDT #617

Conversation

mjcarroll commented Jul 24, 2024

scpeters left a comment

Choose a reason for hiding this comment

azeey left a comment

Choose a reason for hiding this comment

azeey Aug 5, 2024

Choose a reason for hiding this comment

scpeters Aug 15, 2024

Choose a reason for hiding this comment

iche033 commented Aug 12, 2024

iche033 commented Aug 14, 2024

scpeters commented Aug 15, 2024

patch

output with GTS

output with CDT

iche033 commented Aug 15, 2024

patch

output with GTS

output with CDT

codecov bot commented Aug 15, 2024

Codecov Report

scpeters Aug 15, 2024

Choose a reason for hiding this comment

scpeters commented Aug 15, 2024

iche033 commented Aug 16, 2024

scpeters left a comment

Choose a reason for hiding this comment

iche033 commented Aug 16, 2024