AnalysisSitus¶
Home / External Tools & Partners / AnalysisSitus
AnalysisSitus is an open-source CAD analysis framework built on OpenCascade, developed by Quaoar Studio (Yerevan, Armenia). It provides geometric inspection, feature recognition, and Attributed Adjacency Graph (AAG) construction from B-Rep data. Commercial extensions add CNC-specific recognizers for machined features, face accessibility analysis, and DFM checks.
Overview¶
| Website | analysissitus.org |
| Repository | gitlab.com/ssv/AnalysisSitus |
| Developer | Quaoar Studio LLC |
| Type | Open-source framework (BSD 3-Clause) + commercial CNC SDK |
| Domain | CAD geometry analysis, feature recognition, AAG construction |
| Status | Active — early R&D |
How We Use It¶
AnalysisSitus is the primary feature recognition tool in our early AutoCAM R&D. We started with the open-source repository and now use the commercial CNC SDK for our main feature detection pipeline.
- Feature recognition — Classical recognizers identify holes, threads, fillets, chamfers, slots, pockets, and other machinable features from B-Rep geometry. This gives us working feature detection on basic parts, unblocking AutoCAM experimentation without waiting for ML models to be trained.
- AAG construction — Builds Attributed Adjacency Graphs from B-Rep data (faces as nodes, shared edges as arcs with convexity and dihedral angle attributes). This enables AAG-based experimentation and gives us a concrete understanding of the graph data structure that ML models like AAGNet operate on.
Transition Path¶
AnalysisSitus's classical recognizers are well-suited for the basic parts we work with in early R&D. As we collect more training data and tackle more complex parts — where CAM decision-making depends more heavily on detected machinable features — we expect to transition toward model-based feature recognition methods like AAGNet and BRepMFR. The AAG representation that AnalysisSitus produces is the same graph structure these ML models consume, making it a natural bridge from classical to learned methods.
References¶
- analysissitus.org — Official site with documentation and downloads
- GitLab repository — Open-source codebase (BSD 3-Clause)
- Quaoar Studio blog — Developer blog with technical deep-dives on feature recognition, AAG, and geometry algorithms