Expand description
Knowledge Graph extraction for hardware verification.
Extracts a structured Knowledge Graph from parsed hardware specifications, providing LLMs with formally grounded context for SVA generation.
§Ontology (Sprint 0E)
The formal hardware ontology provides 28+ entity types and 24+ relation types, replacing AssertionForge’s 35 LLM-prompt labels and 59 string labels with formally grounded, parameterized, serializable types.
Structs§
- HwKnowledge
Graph - Hardware Knowledge Graph — extracted from Kripke-lowered FOL.
- KgEdge
- An edge in the Knowledge Graph.
- KgProperty
- A temporal property node.
- KgSignal
- A signal node in the Knowledge Graph.
Enums§
- Arbitration
Scheme - Counter
Direction - HwEntity
Type - Formal hardware entity types with parameterized attributes.
- HwRelation
- Formal hardware relation types with parameterized attributes.
- KgRelation
- Relation type for KG edges.
- Port
Direction - Reset
Polarity - Signal
Role - Role of a hardware signal.
- Signal
Type
Functions§
- extract_
from_ kripke_ ast - Extract a Knowledge Graph from a Kripke-lowered LogicExpr AST.