correlation: rules decide, one cheap LLM call only when ambiguous; drop the consensus tiebreak (server-26#299)

The rules engine now commits every confident decision alone. _run_decision
marks a decision ambiguous (new_with_candidates, unfit_tg_candidate,
unverified_unit_link) and only then does upload._correlate make ONE
llm_correlator.decide call; a usable answer is committed, otherwise the rules
decision stands. Removed: _correlate_with_consensus, llm_correlator.tiebreak /
decisions_agree / tiebreak prompt, corr_smart_model, the correlation_smart
health tier, the LLM-orphan gate (_call_is_substanceless, effectively dead:
it could only fire on minor-severity untyped calls) and its sweep budget.
corr_consensus is now rules_only | llm | rules_fallback | rules_guard, with
corr_band naming why; admin debug and replay metrics read the new values.

Section fixes in the same pass:
- #115: no tiebreaker left to manufacture incidents; an LLM orphan of a
  rules "new" is honoured only for untyped, substanceless, routine/minor calls.
- #86: divergence veto bar moved to config (embedding_divergence_threshold);
  the band between it and the match bar is ambiguous, not a silent merge.
- #103: a unit link with no embedding is "unit_overlap_unverified" ->
  ambiguous, instead of skipping the veto and linking as verified.
- #189: unit-only and LLM links refused once an incident is open 90 min;
  unit links on incidents open >30 min are ambiguous; LLM links distance-
  guarded (incident_max_spread_km) and only to shown candidates.
- #190: severity steps down on an explicit stand-down phrase; new
  severity_peak keeps the monotonic max.
- #88: parse_llm_bool fails closed on "false"; scene closures honoured only
  on confident links or the incident the scene opened.
- #27: sweep budget is the scan window (wall clock), not 3 attempts.
- LLM prompt: "Reserve new" bullet no longer contradicts the #189 EXCEPT.
- updated_at never moves backward on a back-dated link.

Unmeasured: no live AI or replay (credits unfunded); band frequency, F1 and
thresholds are unverified. Must be re-scored on both answer keys in Replay
before merge.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

drb-c2-core/tests/test_ai_timeouts.py

@@ -153,8 +153,10 @@ def test_timeout_messages_classify_as_transient():
 
 
 def _corr_ctx():
+    from datetime import datetime, timezone
     return {"call_id": "c1", "is_thin_call": False, "recent": [{"incident_id": "i1"}],
-            "all_active": [], "call_units": ["12"], "call_cleared": [], "tags": ["mva"]}
+            "all_active": [], "call_units": ["12"], "call_cleared": [], "tags": ["mva"],
+            "now": datetime.now(timezone.utc)}
 
 
 @pytest.mark.asyncio
@@ -163,27 +165,14 @@ async def test_gemini_timeout_in_decide_falls_back_to_rules():
         raise httpx.ReadTimeout("The read operation timed out")
 
     with patch.object(settings, "gemini_api_key", "test-key"), \
-         patch.object(llm_correlator, "_build_decide_prompt", lambda ctx: "p"), \
+         patch.object(llm_correlator, "_build_decide_prompt", lambda ctx, cands: "p"), \
          patch.object(llm_correlator, "_sync_gemini", hang):
-        assert await llm_correlator.decide("c1", _corr_ctx()) is None  # None == rules_only
+        assert await llm_correlator.decide("c1", _corr_ctx()) is None  # None == keep the rules decision
     state = ai_health.snapshot()["correlation_cheap"]
     assert state["consecutive_failures"] == 1
     assert state["permanent"] is False
 
 
-@pytest.mark.asyncio
-async def test_gemini_timeout_in_tiebreak_keeps_the_rules_decision():
-    def hang(model, prompt):
-        raise httpx.ReadTimeout("The read operation timed out")
-
-    rules = {"action": "new", "matched_incident": None, "corr_debug": {}}
-    with patch.object(llm_correlator, "_build_tiebreak_prompt", lambda r, l, c: "p"), \
-         patch.object(llm_correlator, "_sync_gemini", hang):
-        out = await llm_correlator.tiebreak(rules, {"action": "orphan"}, _corr_ctx())
-    assert out is rules
-    assert ai_health.snapshot()["correlation_smart"]["permanent"] is False
-
-
 @pytest.mark.asyncio
 async def test_gemini_timeout_in_transcript_correction_keeps_raw_text():
     def hang(model, prompt):
@@ -233,8 +222,8 @@ def test_worst_case_bound_follows_the_settings():
          patch.object(settings, "gemini_timeout_seconds", 20.0), \
          patch.object(settings, "stt_retry_on_degenerate", True):
         bound = recorrelation_sweep.pipeline_worst_case_seconds(n_scenes=2)
-    # stt 2x2x100 + correction 20 + extraction 2x10 + embeds 2x2x10 + decide/tiebreak 2x20 + slack
-    assert bound == 400 + 20 + 20 + 40 + 40 + recorrelation_sweep._NON_AI_SLACK_SECONDS
+    # stt 2x2x100 + correction 20 + extraction 2x10 + embeds 2x2x10 + one correlation call 20 + slack
+    assert bound == 400 + 20 + 20 + 40 + 20 + recorrelation_sweep._NON_AI_SLACK_SECONDS
 
 
 def test_hold_back_covers_the_worst_case_and_never_drops_below_old_floor():