Releases: tudelft3d/val3dity
release v2.0.3
Each surface are now triangulated with a constrained Delaunay triangulation (CDT) instead of a non-Delaunay one. That improves stability, especially when there are nearly collinear vertices in a surfaces: slivers are avoided, which means less errors 204, and less wrong errors caused by numerical stability.
fixes issue with wrong errors 104 often raised
Explained here what is going on: CGAL/cgal#2733
This releases fixes it with a hack, but I think this is solid (touch wood).
fix bug with least-square fitting that caused wrong 203 errors
2.0.1 Merge branch 'hotfix/version'
Release version 2.0.0
Almost a complete re-write, with:
- support for all 3D primitives in ISO19107
- support for new formats: CityJSON, OFF
- completely new reports: in JSON and HTML (browsable version)
- complete unit test suite (with pytest)
Final version v1.x
the last commit before v2 became the master, with a few Git tricks to swap the branches (and avoid working days on merging...).
bug fix for parsing inner rings in surfaces referenced by an xlink
it now works. Thanks Claus Nagel.
Buildings now handled
Validation can now consider CityGML Buildings when validating, ie only the primitives used to represent Buildings are validated, and the report contains the Building ID for each primitive.
Plus, the summary shows how many Buildings are valid/invalid; keep in mind that one Building can be represented with both multiple Solids and/or MultiSurfaces.
v1.1.1: handling of collapsed triangles
prevents collapsed triangles given as input from crashing return error 104 instead, as it should.
v1.1: better handling of inner shells
and error 308 (global orientation of normals for a shell) is now error 405 (since related to a Solid).
v1.0 of val3dity is released
Today I'm happy to release officially the v1.0 of val3dity: my code to validate 3D geometries (solids, composite- and multi-surface).
I started in 2010 to work on it (thanks to a sponsorship from Safe Software), and I've been working on it part-time since then.
Through an OGC CityGML Quality Interoperability Experiment (QIE), I've tested and improved it a lot the last few months.
It's now mature, robust and complete enough to be used by anyone.