Hacktoberfest 2026: le issue che i maintainer hanno segnato per ottobre, aperte e adatte ai principianti. Sfoglia le issue Hacktoberfest

[Proposal] Ontogeny — a self-improving operational ontology platform built on HugeGraph

Aperta
  • Qualcosa non va?
#376 4 commenti 2 reazioni 1 assegnatario Vedi su GitHub

I maintainer di solito rispondono entro 1 giorno

@simon824 ci sta già lavorando.

Dal 28/9/2026.

Valutazione

Questa issue non è ancora stata valutata.

Descrizione

Search before submitting
  • I had searched in the feature and found no similar feature requirement.
Feature description — Ontogeny, a self-improving operational ontology platform

Ontogeny (renamed from the earlier working title KINA, "Knowledge In Action") is an open-source operational ontology platform we'd like to contribute to hugegraph-ai. The name carries the thesis: Greek onto- (being, the root of ontology) + genesis (birth, development) — in biology, ontogeny is the full-life-cycle development of an individual organism; here, an ontology is not a static schema but a living thing, declared in Git and grown through use under governance. It is a governed semantic + kinetic layer between raw data and the existing AI features: business objects, links, actions and permissions that GraphRAG, agents and applications can share.

What it does

  • Ontology as code — objects / links / actions / functions / policies declared in YAML, PR-reviewed, compiled into table schemas and a HugeGraph graph projection (SQL-only fallback without HugeGraph).
  • Action runtime — Cedar-style deny-by-default authorization, markings, immutable audit, idempotency; every write goes through one five-stage gate.
  • Agent surface — MCP server + a pluggable LLM agent engine; agents live under the same policy surface as humans, and writes require human approval.
  • Governed self-improvement — telemetry → LLM-proposed DSL diffs → deterministic eval suites → tiered release (T0 auto-merge to T3 human-only).

Status: v0.3, ~8.6k lines of Python plus a React console; 456 backend / 241 frontend / 8 E2E tests green; Apache-2.0 headers in place.

Merge plan with hugegraph-ai (three steps)

The full codebase is now available on this repository's ontogeny branch (single squashed commit). The complete merge plan, including acceptance criteria per step, lives at docs/hugegraph-ai-merge-plan.md on that branch.

Step Scope
1. Structure merge (M0) hugegraph-ai merged into third_party/hugegraph-ai/ via subtree (pinned to a release tag), zero code changes, zero import coupling, license/NOTICE provenance registered
2. Functional merge (M1/M2) [graphai] optional extra wiring hugegraph-llm through the platform's LLM gateway: GraphRAG Q&A (ask_graph MCP tool, masked & cited) and text-to-ontology proposals entering the T1 human-review track
3. Deep merge (M3/M4) ML semantic search ([ml] extra, embeddings on the projection), graph-compute endpoints promoted from stubs, and new telemetry signals closing the RSI ↔ graph loop

Three ground rules hold across all steps: no second governance channel (every LLM call goes through the platform gateway, every write through the Action runtime), an anti-corruption layer (server/ never imports upstream packages directly), and graph queries return IDs assembled back through the engine so markings always apply.

Suggestion: once Step 1 (structure merge) passes its acceptance checks (baseline all green, zero coupling, provenance in place), we recommend merging the ontogeny branch into main first, then landing Steps 2–3 as normal feature PRs on top.

Known gaps

  • validating / tuning self-evolution on real scenarios;

  • third-party integrations (graph engines, LLMs, storage, BI/ERP connectors);

  • ontology modeling usability.

  • Yes, I am willing to submit a PR!

Code of Conduct

Lingua principale
Python
Stelle
144
Fork
91
Merge medio
7h 27m
PR unite (30g)
7

Preparare l'ambiente

Questo progetto non fornisce container di sviluppo, Dockerfile né guida per i contributori, quindi l'ambiente è a tuo carico: parti dal suo README e consulta la nostra guida al primo contributo per i passaggi generali.

Come iniziare

  1. Leggi tutta la issue e poi la guida ai contributi del progetto.
  2. Commenta sulla issue per dire che te ne occupi tu — evita che due persone facciano lo stesso lavoro.
  3. Fai un fork del repository e lavora su un branch.
  4. Apri una pull request che faccia riferimento al numero della issue.

Altre issue di apache/hugegraph-ai

Tutte le issue di apache/hugegraph-ai

Issue simili

Altre issue su Python

Ricevi le nuove issue nella tua casella

Un breve riepilogo di issue GitHub adatte ai principianti.