core/chat
Shay 401ae53328
chore(generate): make stop-tokens caller-overridable via RuntimeConfig (#87)
Closes audit Finding 6 (2026-05-20).

Pre-fix ``_STOP_TOKENS = frozenset({"it", "to", "word"})`` was
hardcoded inside ``generate.stream.generate()`` and inhibited those
three tokens unconditionally across every pack, every language, and
every domain.  If a pack legitimately needed one of them as a content
word — e.g. a philosophy pack where ``"word"`` maps to λόγος, or a
syntax pack where ``"to"`` is a content node — there was no override
path.  The ``_try_index`` guard handled the case where the token was
absent from the pack, but offered nothing for packs that contained
the token and meant it.

Changes:

  * ``generate.stream.generate`` accepts ``stop_tokens: frozenset[str]
    | None = None``.  ``None`` resolves to the historical
    ``_STOP_TOKENS`` constant, preserving byte-identity for every
    pre-Finding-6 caller.
  * ``RuntimeConfig.stop_tokens: tuple[str, ...] | None = None`` —
    operator-level override threaded through ``ChatRuntime`` into
    ``generate()``.
  * Default ``None`` preserves byte-identical behavior for every
    existing pack and every existing test.

Scope notes:

  * This PR delivers the *runtime override* surface.  Manifest-driven
    per-pack overrides (``generation_stop_tokens`` field in the pack
    manifest) are the natural next step but require a pack-schema
    ADR and re-ratification of every affected pack, so the wiring
    lands first and the manifest field follows on a separate ADR.
  * ``agenerate`` was identified as unreachable and is being deleted
    in a sibling PR (Finding 7); its hardcoded ``_STOP_TOKENS``
    reference disappears with it, so it is intentionally not touched
    here.

Verification:

  * 4 new tests in ``tests/test_stop_tokens_override.py``:
      - ``RuntimeConfig.stop_tokens`` defaults to ``None``
      - ``generate()`` signature exposes ``stop_tokens`` with default
        ``None``
      - the historical constant is unchanged
      - an explicit override flows through the runtime end-to-end
  * ``core eval cognition`` — public 100/100/91.7/100, byte-identical
    to the MEMORY baseline.
  * ``core test --suite cognition`` — 120/0/1.
  * ``core test --suite smoke`` — 67/0.
  * ``core test --suite runtime`` — 19/0.
2026-05-20 19:59:33 -07:00
..
__init__.py
__main__.py
anaphora.py feat(adr-0066): session-thread context + opt-in anaphora prefix (Phase 3.1 + 3.2) 2026-05-18 17:01:34 -07:00
atom_equivalence.py feat(telemetry): ADR-0078 Phase 1 — composer/graph atom equivalence (observational) 2026-05-20 06:14:25 -07:00
cross_pack_grounding.py feat(register): R1–R4 register pack subsystem — deterministic surface variation 2026-05-19 16:52:36 -07:00
example_surface.py feat(register): R1–R4 register pack subsystem — deterministic surface variation 2026-05-19 16:52:36 -07:00
narrative_surface.py feat(register): R1–R4 register pack subsystem — deterministic surface variation 2026-05-19 16:52:36 -07:00
oov_surface.py feat(adr-0066): NARRATIVE + EXAMPLE intents with multi-clause composers (Phase 3.3 + 3.4) 2026-05-18 17:01:55 -07:00
pack_grounding.py feat(adr-0085): gloss-aware CAUSE composer — explanation frame from glosses (#70) 2026-05-20 15:55:08 -07:00
pack_resolver.py feat(packs): en_collapse_anchors_v1 — activate chesed/shalom/tzedek lenses on EN input 2026-05-20 10:58:07 -07:00
pack_surface_candidate.py feat(pack-grounding): selector-ready gloss wiring via PackSurfaceCandidate 2026-05-19 07:26:46 -07:00
partial_surface.py feat(adr-0065): OOV cliff → five-tier honesty gradient (Phase 2.1 + 2.2) 2026-05-18 16:41:45 -07:00
refusal.py
register_substantive.py feat(register): ADR-0077 — substantive register knobs + layering boundary (R6) 2026-05-19 23:39:11 -07:00
register_variation.py feat(register): R5 — operator-visible register telemetry + tour demo 2026-05-19 19:03:07 -07:00
runtime.py chore(generate): make stop-tokens caller-overridable via RuntimeConfig (#87) 2026-05-20 19:59:33 -07:00
teaching_grounding.py feat(adr-0083): transitive (multi-hop) teaching-grounded surface (#63) 2026-05-20 14:11:40 -07:00
telemetry.py feat(telemetry): ADR-0078 Phase 1 — composer/graph atom equivalence (observational) 2026-05-20 06:14:25 -07:00
thread_context.py feat(adr-0066): session-thread context + opt-in anaphora prefix (Phase 3.1 + 3.2) 2026-05-18 17:01:34 -07:00
verdicts.py feat(adr-0039): audit completeness — TurnVerdicts bundle, stub TurnEvent, hedge_injected 2026-05-17 21:32:46 -07:00