fix(export): tolerate id-less persisted hyperedges in attach_hyperedges (#2775)

[ousamabenyounes]

Summary

Fix #2775 — an incremental re-extract died at the merge step with KeyError: 'id'.

attach_hyperedges (graphify/export.py) seeded its dedup set from the
previously-persisted hyperedges with a hard subscript:

seen_ids = {h["id"] for h in existing}

The semantic/LLM extractor routinely emits hyperedges as
{"nodes": [...], "type": "...", "attributes": {...}} with no id, and
build.py persists them verbatim. So a prior graph.json can — and in practice
does — contain id-less hyperedges. The first graphify extract on a corpus has
no prior graph to read back and succeeds; every incremental re-extract
afterwards hits that comprehension and raises, writing nothing. Because the
first run works, it looks like flaky extraction; it is deterministic.

The loop directly below already tolerates id-less entries in the incoming set
via .get("id"). This change makes the read symmetric — it skips id-less
persisted entries when seeding the dedup set:

seen_ids = {h["id"] for h in existing if h.get("id")}

Scope is intentionally the crash only: the pre-existing behavior of the loop
(id-bearing incoming hyperedges are deduped and appended; id-less incoming are
skipped) is unchanged, and id-less persisted entries are retained in the graph.

Test verification (RED → GREEN)

New regression test test_attach_hyperedges_tolerates_id_less_persisted seeds an
id-less persisted hyperedge — the exact input the semantic extractor produces.

RED — on the unmodified v8 code (prod fix reverted):

tests/test_hypergraph.py:148: in test_attach_hyperedges_tolerates_id_less_persisted
    attach_hyperedges(G, [{"id": "flow_a", "label": "Flow A", "nodes": ["A", "B"]}])
graphify/export.py:166: in attach_hyperedges
    seen_ids = {h["id"] for h in existing}
E   KeyError: 'id'
1 failed in 1.72s

GREEN — with the fix:

tests/test_hypergraph.py::test_attach_hyperedges_tolerates_id_less_persisted PASSED
21 passed in 0.58s   # full tests/test_hypergraph.py

Full suite is unchanged aside from the new passing test (pytest tests/ →
4514 passed, 7 skipped; the 3 test_ollama::test_detect_backend_* failures are
pre-existing on v8 and unrelated — backend env-var detection). skillgen --check is clean; no generated skill files touched.

Files changed

File	Change
graphify/export.py	attach_hyperedges: skip id-less persisted entries when seeding the dedup set
tests/test_hypergraph.py	regression test for id-less persisted hyperedges
CHANGELOG.md	entry under 0.9.44

[graphify-labs]

Graphify reviewed this change.

Looks safe to merge — no coupling regressions and no blocking issues, checked against the code graph (not a self-assessment).

Formal verification. 1 change(s) tested, no difference found (not proven).

---

Graphify review — findings

This PR fixes a KeyError: 'id' that occurred in attach_hyperedges (in graphify/export.py) during incremental re-extraction when a prior graph.json contained hyperedges lacking an id field. The dedup-set construction now filters persisted entries to only those with a truthy id, mirroring the existing guard applied to incoming hyperedges. A new regression test (test_attach_hyperedges_tolerates_id_less_persisted) and a CHANGELOG entry are added to cover this scenario. The touched surface is limited to the one-line dedup change, the test file, and documentation.

No blocking issues surfaced.

Analysis details — impact, health, verification

Impact & health

Graphify review

Impact — 653 functions depend on the 260 functions this change touches.

Health — this change adds coupling hotspots:

• new: _rebuild_code() — 95 callers, 51 callees
• new: build_merge() — 41 callers, 14 callees
• new: to_obsidian() — 29 callers, 12 callees
• new: to_json() — 44 callers, 6 callees
• new: dispatch_command() — 2 callers, 117 callees
• new: _make_graph() — 30 callers, 6 callees
• new: run_pipeline() — 8 callers, 13 callees
• new: to_canvas() — 17 callers, 4 callees
• …and 4 more

Verification — 653 functions in the blast radius were not formally verified this run (proofs are advisory here).

Gate & verification

graphify gate

PASS — objectively clean (no health regressions, tests not run — proofs not run this pass (advisory)). Grounded, not self-assessed.

Advisory (not blocking):

• verification_scope: 510 function(s) in the blast radius were not formally verified this run

Formal verification

No difference found (not proven): No behavior difference found in attach\_hyperedges (not a proof).

The verifier ran both versions of attach\_hyperedges on many inputs and saw identical behavior every time. Strong evidence the change is safe, but evidence, not a proof.

Guarantee: Empirical: differential testing (both versions run on many generated inputs). A divergence on an untested input remains possible, so this is 'no counterexample found', not 'proven equivalent'.

Note: An input the sampler did not try could still differ.

· 12 more finding(s) on lines outside this diff (see the check run).

[safishamsi]

Landed in v0.9.45, just published to PyPI. Cherry-picked onto v8 with your authorship preserved in the commit, so both the change and the credit are in the history. Thanks @ousamabenyounes. Closing since it is now released.