onya

Onya provides a knowledge graph data model and expression format that combines a formal specification with a Python implementation. The model treats edges and properties as first‑class entities that can carry their own nested assertions, enabling qualified values, relationship metadata, and n‑ary relations without reification. The Onya Literate serialization is Markdown‑native, making it readable for humans and consumable for large language models, and it supports interpretations (data contracts) for value‑level typing while remaining format‑agnostic. The Python library offers parsing, graph construction, traversal, and round‑trip serialization, plus optional persistence via file, SQLite, or PostgreSQL stores and a CLI for converting graphs to Mermaid or Graphviz formats. An optional networkx projection enables analytics that can be written back as typed assertions. A skill file (onya‑graph.SKILL.md) is provided to assist AI coding agents in authoring Onya graphs. The project is open source under Apache 2.0 for code and CC BY 4.0 for the specification, encouraging broad adoption and community contributions.