kaos-graph 0.1.0

High-performance graph library for the Kelvin Agentic OS — petgraph-backed engine, 40+ algorithms, RDF/SPARQL, knowledge-graph tools

kaos-graph

Part of Kelvin Agentic OS (KAOS) — open agentic infrastructure for legal work, built by 273 Ventures. See the full KAOS package map for the rest of the stack.

kaos-graph is a high-performance graph library for KAOS — a typed Python API over a pure-Rust core backed by petgraph and the oxrdf RDF engine. It provides the graph-algorithm and knowledge-graph building blocks the rest of the stack relies on: 40+ algorithms (PageRank, shortest paths, centrality, community detection, DAG ops), RDF/SPARQL parsing, and typed Python wrappers throughout.

It is dependency-light: the BASE install pulls only kaos-graph itself plus defusedxml (~250 KB pure-Python). Optional extras layer in the rest of the KAOS ecosystem.

Install

uv add kaos-graph
# or
pip install kaos-graph

kaos-graph requires Python 3.13 or newer. The published wheels are cp313-abi3 — one wheel per OS/architecture covers every CPython 3.13+ minor (3.13, 3.14, 3.15, …). No re-release needed when 3.15 ships.

Platform coverage: Linux x86_64 (manylinux + musllinux), Linux aarch64 (manylinux + musllinux), macOS arm64, Windows x86_64, Windows arm64.

Quick start

from kaos_graph import Graph
from kaos_graph.algorithms import pagerank, shortest_paths, topological_sort

g = Graph()
g.add_node("alice", role="engineer")
g.add_node("bob", role="manager")
g.add_edge("alice", "bob", relationship="reports_to")

# Algorithms
ranks = pagerank(g)
costs = shortest_paths(g, "alice")
order = topological_sort(g)

Concepts

The package is built around a small set of typed primitives.

Concept	What it is
Graph	Property graph backed by petgraph::StableDiGraph. String-keyed nodes with JSON-valued properties; edges with arbitrary properties; multi-graph semantics opt-in. Stable node refs across mutations (additions/removals don't invalidate handles).
Node / Edge / Triple	Typed dataclass results emitted at the FFI boundary. Node(id, properties), Edge(source, target, properties), Triple(subject, predicate, object).
Algorithms	40+ functions in kaos_graph.algorithms covering traversal (BFS/DFS, topological sort), shortest paths (Dijkstra, A*, Bellman-Ford, Floyd-Warshall), centrality (PageRank, betweenness, closeness, eigenvector), community (Louvain, label propagation), connectivity (SCC, weakly-connected, articulation points, bridges), structure (cycles, cliques, isomorphism, max flow), and DAG ops (longest path, critical path).
Sparse PageRank	O((V+E)·n) instead of petgraph's O(V²·E·n). FOLIO ontology (~25K nodes, ~36K edges): 7 ms vs petgraph's 105 s. The marketing-grade benchmark exercised on every CI run.
RDF / SPARQL	oxrdf + oxrdfio parsers (Turtle, N-Triples, N-Quads, RDF/XML, TriG) with byte / triple / size caps applied at the FFI boundary. SPARQL via pyoxigraph opt-in through the [rdf] extra.
I/O & bridges	JSON (round-trippable), GraphML / GEXF (defusedxml-hardened), edgelist, adjacency-list, mermaid, dot. NetworkX bridge; Polars / DuckDB tabular bridge planned via the [tabular] extra in 0.1.0a2.
Knowledge graph	Schema, KnowledgeGraph, fact ingestion, reasoning rules, ontology bridges.
Storage	VFS-backed save_to_vfs / load_from_vfs with name validation against namespace escape.

CLI

kaos-graph ships a kaos-graph administrative CLI plus an optional kaos-graph-serve MCP server (loopback-only by default):

kaos-graph version          # version info
kaos-graph serve            # forwards to kaos-graph-serve (stdio MCP)
kaos-graph-serve            # MCP server, stdio transport
kaos-graph-serve --http     # MCP server, streamable HTTP (loopback only)

Note: 17 MCP tools are registered by register_graph_tools() and become available with pip install kaos-graph[mcp] once the kaos-mcp companion package publishes (planned for 0.1.0a2).

Compatibility & status

Aspect
Python	3.13, 3.14 (informational matrix entries for 3.14t free-threaded and 3.15-dev). One cp313-abi3 wheel per OS/arch covers all 3.13+ minors.
OS	Linux (manylinux + musllinux, x86_64 + aarch64), macOS arm64, Windows x86_64, Windows arm64. macOS x86_64 deliberately skipped (Apple ended Intel sales in 2023).
Maturity	Alpha. The public API is documented in kaos_graph.__all__.
Stability policy	Pre-1.0: minor bumps may change behaviour. Every change is documented in CHANGELOG.md.
Test coverage	173 Rust unit tests + 402 Python tests passing (575 total). 38 dedicated security regressions covering the audit-pass A2 + two follow-up rounds.
Type checker	Validated with ty, Astral's Python type checker.

Companion packages

kaos-graph is one of the packages in the Kelvin Agentic OS. The broader stack:

Package	Layer	What it does
kaos-core	Core	Foundational runtime, MCP-native types, registries, execution engine, VFS
kaos-content	Core	Typed document AST: Block/Inline, provenance, views
kaos-mcp	Bridge	FastMCP server, kaos management CLI, MCP resource templates
kaos-pdf	Extraction	PDF → AST with provenance
kaos-web	Extraction	Web extraction, browser automation, search, domain intelligence
kaos-office	Extraction	DOCX / PPTX / XLSX readers + writers to AST
kaos-tabular	Extraction	DuckDB-powered SQL analytics
kaos-source	Data	Government + financial data connectors (Federal Register, eCFR, EDGAR, GovInfo, PACER, GLEIF)
kaos-llm-client	LLM	Multi-provider LLM transport
kaos-llm-core	LLM	Typed LLM programming (Signatures, Programs, Optimizers)
kaos-nlp-core	Primitives (Rust)	High-performance NLP primitives
kaos-nlp-transformers	ML	Dense embeddings + retrieval
kaos-graph	Primitives (Rust)	Graph algorithms + RDF/SPARQL
kaos-ml-core	Primitives (Rust)	Classical ML on the document AST
kaos-citations	Legal	Legal citation extraction, resolution, verification
kaos-agents	Agentic	Agent runtime, memory, recipes
kaos-reference	Sample	Reference module for module authors

Packages depend on kaos-core; everything else is opt-in. Mix and match the ones you need.

Development

git clone https://github.com/273v/kaos-graph
cd kaos-graph
uv sync --group dev --extra rdf
uv run maturin develop --release

Install pre-commit hooks (recommended — they run the same checks as CI on every commit, scoped to staged files):

uvx pre-commit install
uvx pre-commit run --all-files     # one-time full sweep

Manual QA commands (the same set CI runs):

cargo fmt --check
cargo clippy --no-default-features --all-targets -- -D warnings
cargo test --no-default-features --lib
uv run ruff format --check python/kaos_graph tests
uv run ruff check python/kaos_graph tests
uv run ty check python/kaos_graph tests
uv run pytest tests/

Build from source

uv build
uv pip install dist/*.whl

Contributing

Issues and pull requests are welcome. See CONTRIBUTING.md for setup, quality gates, pull request expectations, and engineering standards. By contributing you agree to follow the project conduct expectations and certify the Developer Certificate of Origin v1.1 — sign every commit with git commit -s. Please open an issue before starting on a non-trivial change so we can align on scope.

Security

For security issues, please do not file a public issue. Report privately via GitHub Private Vulnerability Reporting or email [e-mail]. See SECURITY.md for the full disclosure policy.

License

Apache License 2.0 — see LICENSE and NOTICE.

Copyright 2026 273 Ventures LLC. Built for kelvin.legal.