xgen-omnifuse 0.1.0

Backend-agnostic one-shot GraphRAG: fuse vector + graph (label/class/relation) seeds with MMR diversity into a single synthesis. Zero-infra default.

OmniFuse

Backend-agnostic, one-shot GraphRAG. Fire several retrieval strategies at once — vector/lexical passages + graph label-linking + class enumeration + relation expansion — and fuse them with MMR diversity into a single LLM synthesis. No iterative ReAct tool loop. Zero infra, zero lock-in: the full algorithm runs on a pure-Python in-memory backend (dict + BM25), and swaps to Fuseki / Qdrant / any LLM by passing objects that match three small protocols.

from omnifuse import build_inmemory, Node, Triple, Chunk

of = build_inmemory(nodes, triples, chunks)        # in-memory, no DB, no API key
print(of.search("불법도박 관련 규정").answer)

Why graph fusion (not just vectors)

Pure vector RAG answers from the top-k passages it happens to embed near the query. A graph store also gives you operations cosine similarity can't:

• Complete enumeration — all instances of a class ("list every regulation"), exact counts.
• Relations / multi-hop — what an entity is connected to, 1-hop neighbors, paths.
• Minority evidence survives — MMR diversity keeps the decisive exception/warning that near-duplicate passages would otherwise crowd out of a fixed top-k.

OmniFuse fuses both: the vector seed for content, the graph seeds for structure.

Design — algorithm as a library

The algorithm only talks to three typing.Protocols, never to a database:

class GraphStore(Protocol):
    def search_labels(self, query, *, limit=30) -> list[tuple[Node, float]]: ...   # full-text label search
    def class_instances(self, class_id, *, limit=1000) -> list[Node]: ...          # enumeration
    def neighbors(self, node_id, *, hops=1, limit=100) -> list[tuple[str,str,str]]: ...  # traversal
    def count_class(self, class_id) -> int: ...
    def get_node(self, node_id) -> Node | None: ...

class VectorStore(Protocol):
    def search(self, query, *, limit=20) -> list[tuple[Chunk, float]]: ...
    def fetch(self, ids) -> list[Chunk]: ...

class LLM(Protocol):
    def generate(self, prompt, *, system="", timeout=None) -> str: ...

• Zero-infra default — InMemoryGraph indexes node labels with BM25 (CJK character n-grams, so Korean/CJK search works with no morphological analyzer), and InMemoryVector uses cosine when embeddings are present, else BM25 lexical.
• dependencies = [] — the core needs nothing but the standard library. Real backends are optional extras (pip install "xgen-omnifuse[fuseki,qdrant]").
• Bring your own LLM — pass anything with generate(...); the bundled EchoLLM returns the fused evidence so the pipeline runs end-to-end with no API key.

The pipeline (OmniFuse.search)

1. vector/lexical seed → adaptive top-k (score-distribution cut, not fixed k)
2. graph label-linking → 1-hop relations
3. class enumeration (complete list/count)
4. HippoRAG — entities of the retrieved chunks → 1-hop expansion
5. evidence assembled with MMR diversity (Jaccard, no embeddings needed)
6. one LLM synthesis over the fused evidence
7. honest evidence_nodes — only the nodes the answer actually cites

Install

pip install xgen-omnifuse            # core (zero deps)
pip install "xgen-omnifuse[dev]"     # + pytest, ruff

Run the demo with no install:

python examples/quickstart.py

Layout

src/omnifuse/
  protocols.py     # GraphStore / VectorStore / LLM  (the swap points)
  models.py        # Node, Triple, Chunk, SearchResult
  text.py          # tokenizer + BM25 (CJK n-grams)
  fusion.py        # MMR, adaptive top-k, relation ranking
  oneshot.py       # OmniFuse.search — the fusion algorithm
  backends/memory.py  # InMemoryGraph + InMemoryVector (zero infra)
  llm.py           # EchoLLM, CallableLLM
  facade.py        # build_inmemory(...)
examples/  tests/

Two interchangeable modes (same algorithm)

# (a) self-contained — zero infra
from omnifuse import build_inmemory
of = build_inmemory(nodes, triples, chunks)

# (b) backed by Apache Jena Fuseki (or any SPARQL endpoint) — graph-only or with a vector store
from omnifuse import OmniFuse, InMemoryVector
from omnifuse.backends.fuseki import FusekiGraph
graph = FusekiGraph("http://localhost:3030/ds/query", graph_uri="urn:my-graph", user="admin", password="…")
of = OmniFuse(graph, InMemoryVector([]))   # search() unchanged

FusekiGraph is stdlib-only (urllib) and uses portable FILTER(CONTAINS(...)), so it works on any SPARQL 1.1 store — not just jena-text.

Roadmap

• backends/qdrant.py vector adapter; jena-text fast path for FusekiGraph
• async pipeline (parallel seeds via asyncio.gather)
• reranker / cross-encoder hook, query expansion
• configurable ISA predicates and prompt templates (per domain/language)

CI / Releasing

• ci.yml — runs pytest (3.10–3.12) + python -m build + twine check on every push/PR.
• publish.yml — on a GitHub Release, builds and uploads to PyPI via Trusted Publishing (no token in the repo). One-time PyPI setup: project → Publishing → add pending publisher PlateerLab / xgen-omnifuse / publish.yml / pypi. (Token mode: add secrets.PYPI_API_TOKEN.)

Build locally:

pip install build && python -m build      # dist/*.tar.gz + *.whl

License

TBD.