feat(temporal): canonical half-open bounds for claim windows (D107 WP-T.0a)

.github/ci/integration-paths.txt

@@ -5,6 +5,7 @@ src/tests/eval/test_s58_consumption.py
 src/tests/profiles/test_selfhost_profile.py
 src/tests/spikes/test_operational_scale.py
 src/tests/spine/test_backfill.py
+src/tests/spine/test_canonical_bounds_sql.py
 src/tests/spine/test_clustering.py
 src/tests/spine/test_component_versions.py
 src/tests/spine/test_deployment_bootstrap.py

.github/ci/unit-paths.txt

@@ -39,6 +39,7 @@ src/tests/core/test_knowledge_planner.py
 src/tests/core/test_knowledge_writer.py
 src/tests/core/test_section_snap.py
 src/tests/core/test_structure_skeleton.py
+src/tests/core/test_temporal.py
 src/tests/core/test_text_metering.py
 src/tests/eval/test_s55_hard_forget.py
 src/tests/eval/test_s55_selfhost_restore.py

.github/workflows/ci.yml

@@ -371,7 +371,7 @@ jobs:
           trap - EXIT
           test "$(docker compose --env-file .env.example exec -T postgres \
             psql -U rememberstack -d rememberstack -Atc \
-            'SELECT version_num FROM alembic_version')" = 'p9_25_0046'
+            'SELECT version_num FROM alembic_version')" = 'p9_26_0047'
           test "$(docker compose --env-file .env.example exec -T postgres \
             psql -U rememberstack -d rememberstack -Atc \
             'SELECT count(*) FROM deployments')" = '1'

benchmarks/locomo/README.md

@@ -1,4 +1,4 @@
-# RS-LoCoMo-Full-v21 setup
+# RS-LoCoMo-Full-v22 setup
 
 This directory contains the unshipped full-system LoCoMo adapter. It does not vendor or
 auto-download LoCoMo. Supply the exact pinned `locomo10.json` only after confirming its
@@ -16,12 +16,12 @@ The safe first command is local and makes no API or model call:
 uv run --extra benchmark python -m benchmarks.locomo prepare \
   --dataset /absolute/path/locomo10.json \
   --tier smoke \
-  --protocol full-v21 \
+  --protocol full-v22 \
   --output .benchmark-runs/locomo-smoke
 ```
 
 The harness validates the pinned bytes, renders session documents, and fingerprints the
-eight-question smoke plan. `--protocol` is prepare-only; `full-v21` is the one
+eight-question smoke plan. `--protocol` is prepare-only; `full-v22` is the one
 current-system protocol, and every later stage reads that immutable choice from
 `run.json`. Do not run remote stages until reviewing
 [`locomo_benchmark_design.md`](../../plan/designs/locomo_benchmark_design.md).
@@ -64,6 +64,11 @@ call sees the complete bounded candidate snapshot and returns a supplied
 candidate id or `new`. There is no insufficient-evidence result or
 confidence-routed frontier call.
 
+V22 fingerprints D107 WP-T.0a: the observation adjudicator and `claims_as_of`
+compare canonical half-open bounds (a day is the whole calendar day, an
+instant a non-empty point, adjacent units do not overlap); everything else is
+v21's.
+
 V21 fingerprints the D106 observation adjudicator: dated events with disjoint
 resolved windows never collapse onto or supersede each other (they may only
 contradict or stay distinct), a dated event is never `evidence` for an undated

benchmarks/locomo/__init__.py

@@ -1 +1 @@
-"""The pinned RS-LoCoMo-Full-v21 benchmark adapter."""
+"""The pinned RS-LoCoMo-Full-v22 benchmark adapter."""

benchmarks/locomo/cli.py

@@ -118,7 +118,7 @@ def _parser() -> argparse.ArgumentParser:
     parser = argparse.ArgumentParser(
         prog="python -m benchmarks.locomo",
         description=(
-            "RS-LoCoMo-Full-v21: prepare is local; ingest/answer/judge require "
+            "RS-LoCoMo-Full-v22: prepare is local; ingest/answer/judge require "
             "explicit execution acknowledgements"
         ),
     )

benchmarks/locomo/dataset.py

@@ -140,10 +140,10 @@ def load_manifest(tier: str) -> QuestionManifest:
         )
     if manifest.dataset_commit != DATASET_COMMIT:
         raise DatasetValidationError(
-            "manifest dataset commit is not RS-LoCoMo-Full-v21"
+            "manifest dataset commit is not RS-LoCoMo-Full-v22"
         )
     if manifest.dataset_sha256 != DATASET_SHA256:
-        raise DatasetValidationError("manifest dataset hash is not RS-LoCoMo-Full-v21")
+        raise DatasetValidationError("manifest dataset hash is not RS-LoCoMo-Full-v22")
     actual = item_ids_hash(item_ids=manifest.item_ids)
     if actual != manifest.item_ids_sha256:
         raise DatasetValidationError(

benchmarks/locomo/model.py

@@ -1,4 +1,4 @@
-"""Typed values for the full-system RS-LoCoMo-Full-v21 protocol."""
+"""Typed values for the full-system RS-LoCoMo-Full-v22 protocol."""
 
 from __future__ import annotations
 
@@ -28,8 +28,8 @@
 Category = Literal[1, 2, 3, 4, 5]
 RetainedCategory = Literal[1, 2, 3, 4]
 Tier = Literal["smoke", "development", "publication"]
-ProtocolKey = Literal["full-v21"]
-ProtocolName = Literal["RS-LoCoMo-Full-v21"]
+ProtocolKey = Literal["full-v22"]
+ProtocolName = Literal["RS-LoCoMo-Full-v22"]
 SourceTimezoneBasis = Literal["assumed_utc"]
 AnswerAgentModel = Literal["openai/gpt-5.6-luna"]
 JudgeModel = Literal["openai/gpt-5.6-luna"]
@@ -120,7 +120,7 @@ class QuestionManifest(FrozenModel):
 class RunConfiguration(FrozenModel):
     """Immutable identity of one prepared benchmark run."""
 
-    protocol_name: ProtocolName = "RS-LoCoMo-Full-v21"
+    protocol_name: ProtocolName = "RS-LoCoMo-Full-v22"
     adapter_version: NonEmpty
     prepared_at: datetime
     repository_revision: NonEmpty
@@ -393,7 +393,7 @@ class SessionDiagnosticSummary(FrozenModel):
 class RunSummary(FrozenModel):
     """Publication-ready local aggregate with no hidden denominator."""
 
-    protocol_name: ProtocolName = "RS-LoCoMo-Full-v21"
+    protocol_name: ProtocolName = "RS-LoCoMo-Full-v22"
     protocol_fingerprint: NonEmpty
     tier: Tier
     questions: int = Field(ge=1)

benchmarks/locomo/protocol.py

@@ -29,9 +29,9 @@
 from rememberstack.model import Envelope
 from rememberstack.model import ToolDescriptor
 
-PROTOCOL_NAME: Final = "RS-LoCoMo-Full-v21"
-DEFAULT_PROTOCOL_KEY: Final = "full-v21"
-ADAPTER_VERSION: Final = "locomo-full-adapter-2026.09-temporal-gate-v21"
+PROTOCOL_NAME: Final = "RS-LoCoMo-Full-v22"
+DEFAULT_PROTOCOL_KEY: Final = "full-v22"
+ADAPTER_VERSION: Final = "locomo-full-adapter-2026.09-canonical-bounds-v22"
 MAX_TOOL_CALLS: Final = 8
 MAX_AGENT_CALLS: Final = 9
 ANSWER_READER_RETRY_BUDGET: Final = 2
@@ -74,7 +74,7 @@
             "no-types-1:binary-t4-1:document-t0-1"
         ),
         "adjudicate_observations": (
-            "e3-obs-flush-2026.09a:temporal-gate-1:claim-fanout-1:entity-fanout-1"
+            "e3-obs-flush-2026.09b:canonical-bounds-1:temporal-gate-1:claim-fanout-1:entity-fanout-1"
         ),
         "adjudicate_supersession": "adjudicator-2026.07b:temp0-1",
         "embed_claim": "p1-embed-claims-2026.07",
@@ -211,8 +211,8 @@ class LoCoMoProtocol:
     answer_word_cap: int | None = None
 
 
-_FULL_V21 = LoCoMoProtocol(
-    key="full-v21",
+_FULL_V22 = LoCoMoProtocol(
+    key="full-v22",
     name=PROTOCOL_NAME,
     answer_agent_model=ANSWER_AGENT_MODEL,
     judge_model=JUDGE_MODEL,
@@ -234,7 +234,7 @@ class LoCoMoProtocol:
 )
 
 PROTOCOL_REGISTRY: Final[Mapping[ProtocolKey, LoCoMoProtocol]] = MappingProxyType(
-    {_FULL_V21.key: _FULL_V21}
+    {_FULL_V22.key: _FULL_V22}
 )
 
 

benchmarks/locomo/runner.py

@@ -411,7 +411,7 @@ def ingest_sample(
         _require_current_ingest_bindings(model_bindings=build.model_bindings)
         if build.document_binding_generation != EXPECTED_DOCUMENT_BINDING_GENERATION:
             raise ExecutionGuardError(
-                "deployment document binding generation differs from RS-LoCoMo-Full-v21"
+                "deployment document binding generation differs from RS-LoCoMo-Full-v22"
             )
         _require_current_query_surface(context=context, client=client)
         _require_exact_live_ingests(
@@ -570,7 +570,7 @@ def answer_sample(
     ):
         raise ExecutionGuardError(
             "the deployment did not report the exact completed"
-            " RS-LoCoMo-Full-v21 pipeline, live graph, and fresh P3 projection"
+            " RS-LoCoMo-Full-v22 pipeline, live graph, and fresh P3 projection"
         )
     _require_serving_revision(context=context, readiness=readiness)
     prior_readiness = context.state.readiness.get(sample_id)
@@ -1177,7 +1177,7 @@ def _validate_run(
     """Recompute immutable run identity before any local or remote stage."""
     selected_protocol = protocol_for_name(configuration.protocol_name)
     if configuration.dataset_sha256 != DATASET_SHA256:
-        raise BenchmarkRunError("run dataset hash is not RS-LoCoMo-Full-v21")
+        raise BenchmarkRunError("run dataset hash is not RS-LoCoMo-Full-v22")
     if item_ids_hash(item_ids=manifest.item_ids) != manifest.item_ids_sha256:
         raise BenchmarkRunError("run manifest item hash changed")
     if manifest_bytes_hash(manifest=manifest) != configuration.manifest_sha256:
@@ -1187,7 +1187,7 @@ def _validate_run(
     if manifest.tier != configuration.tier:
         raise BenchmarkRunError("run manifest tier changed")
     if configuration.dataset_commit != DATASET_COMMIT:
-        raise BenchmarkRunError("run dataset commit is not RS-LoCoMo-Full-v21")
+        raise BenchmarkRunError("run dataset commit is not RS-LoCoMo-Full-v22")
     if configuration.adapter_version != ADAPTER_VERSION:
         raise BenchmarkRunError("run adapter version differs from current code")
     if _models_hash(values=documents) != configuration.documents_sha256:
@@ -1445,7 +1445,7 @@ def _require_current_ingest_bindings(*, model_bindings: dict[str, str]) -> None:
             if model_bindings.get(name) != expected.get(name)
         )
         raise ExecutionGuardError(
-            "deployment ingest model bindings differ from RS-LoCoMo-Full-v21: "
+            "deployment ingest model bindings differ from RS-LoCoMo-Full-v22: "
             + ", ".join(mismatches)
         )
 

benchmarks/locomo/sharding/README.md

@@ -178,7 +178,7 @@ as source-derived data.
 | --- | ---: | --- |
 | `LOCOMO_PYTHON` | `.venv/bin/python` | repository virtual-environment Python |
 | `LOCOMO_TIER` | `publication` | prepared manifest tier |
-| `LOCOMO_PROTOCOL` | `full-v21` | prepare-time protocol key |
+| `LOCOMO_PROTOCOL` | `full-v22` | prepare-time protocol key |
 | `LOCOMO_MOUNT_ROOT` | `$RUN_DIR/.mounts` | host/container-identical P3 mount root |
 | `LOCOMO_MAX_DOCUMENTS` | `100` | per-sample ingest authorization |
 | `LOCOMO_MAX_QUESTIONS` | `1540` | run-absolute answer item authorization |

benchmarks/locomo/sharding/run_shard.sh

@@ -25,7 +25,7 @@ run_dir=$2
 dataset_path=$3
 python_bin=${LOCOMO_PYTHON:-.venv/bin/python}
 tier=${LOCOMO_TIER:-publication}
-protocol=${LOCOMO_PROTOCOL:-full-v21}
+protocol=${LOCOMO_PROTOCOL:-full-v22}
 mount_root=${LOCOMO_MOUNT_ROOT:-$run_dir/.mounts}
 max_documents=${LOCOMO_MAX_DOCUMENTS:-100}
 max_questions=${LOCOMO_MAX_QUESTIONS:-1540}
@@ -50,7 +50,7 @@ export GOOGLE_APPLICATION_CREDENTIALS=${LOCOMO_GCP_CREDENTIALS_FILE:-/etc/rememb
 export GOOGLE_API_CERTIFICATE_CONFIG=${LOCOMO_GCP_CERTIFICATE_CONFIG_FILE:-/etc/rememberstack/locomo-gcs/certificate-config.json}
 export GOOGLE_API_USE_CLIENT_CERTIFICATE=true
 
-# RS-LoCoMo-Full-v21's non-secret ingest identity. Override ambient self-host
+# RS-LoCoMo-Full-v22's non-secret ingest identity. Override ambient self-host
 # defaults so every shard runs the exact Luna/Qwen pipeline the protocol checks.
 export REMEMBERSTACK_STRUCTURER_MODEL=openai/gpt-5.6-luna
 export REMEMBERSTACK_SKELETON_CHECK_MODEL=openai/gpt-5.6-luna

plan/designs/locomo_benchmark_design.md

@@ -7,8 +7,19 @@
 > assured surface rolls this protocol; the sequencing is
 > `plan/plans/temporal_clocks.md`. Contract: `temporal_clocks_design.md` §6–§8.
 
-> **Binding D106 amendment (2026-09-03).** The current protocol is
-> `RS-LoCoMo-Full-v21`. It retains v20's dataset, rendered documents, models,
+> **Binding D107 amendment, WP-T.0a (2026-09-03).** The current protocol is
+> `RS-LoCoMo-Full-v22`. It retains v21's dataset, rendered documents, models,
+> tools, budgets, prompts, and scoring. Its pinned `adjudicate_observations`
+> component version now carries canonical half-open bounds (D107 §5): the
+> observation adjudicator compares a day as the whole calendar day and an
+> instant as a non-empty point, and `claims_as_of` finds a day-precision
+> claim from an intraday window. Ingestion provenance, the library's
+> `claims_as_of` result set, protocol identity, and fingerprint roll; the
+> benchmark's exposed read-tool set, retry, model-effort, and call-budget
+> behavior are unchanged.
+
+> **Historical D106 amendment (2026-09-03; superseded by D107 WP-T.0a).** The
+> D106 protocol was `RS-LoCoMo-Full-v21`. It retained v20's dataset, rendered documents, models,
 > tools, budgets, answer and judge prompts, counterfactual and complete-answer
 > instructions, content-before-`Unknown` harness guard, and no-review scoring
 > rule. Its pinned `adjudicate_observations` component version now carries the
@@ -98,7 +109,7 @@ and spend ceiling.
 ## 2. Fixed protocol
 
 ```text
-protocol                RS-LoCoMo-Full-v21
+protocol                RS-LoCoMo-Full-v22
 dataset commit           3eb6f2c585f5e1699204e3c3bdf7adc5c28cb376
 dataset SHA-256          79fa87e90f04081343b8c8debecb80a9a6842b76a7aa537dc9fdf651ea698ff4
 categories               1, 2, 3, 4
@@ -670,7 +681,7 @@ compatibility form. The response contains:
   same-snapshot proven-absent-anchor execution checks when live graph is required;
 - an overall `ready` that is the conjunction of the requested capabilities;
 - every non-secret ingestion/query model binding; and
-- the non-secret `document_binding_generation`, which Full-v21 requires to be
+- the non-secret `document_binding_generation`, which Full-v22 requires to be
   exactly `document-t0-v1` and stores in `run.json` plus the protocol
   fingerprint.
 
@@ -816,11 +827,11 @@ Local preparation:
 uv run --extra benchmark python -m benchmarks.locomo prepare \
   --dataset /absolute/path/locomo10.json \
   --tier smoke \
-  --protocol full-v21 \
+  --protocol full-v22 \
   --output .benchmark-runs/locomo-smoke
 ```
 
-`--protocol` exists only on `prepare`. The sole choice is `full-v21`; ingest,
+`--protocol` exists only on `prepare`. The sole choice is `full-v22`; ingest,
 answer, judge, and summarize read it from the prepared run and expose no
 protocol override.
 

plan/designs/temporal_clocks_design.md

@@ -348,12 +348,15 @@ ends as +∞. A caller's inclusive request `claims_as_of(from, to)` converts
 to `[from, to + 1 µs)`, so `from == to` is a point-in-time query, not an
 empty one.
 
-**Where it runs.** The function ships as one immutable SQL function in the
-`memory_v1` query space (`memory_v1.canonical_bounds`) with a companion
+**Where it runs.** The engine's own SQL calls two IMMUTABLE public-schema
+functions, `claim_canonical_start(from, precision)` and
+`claim_canonical_end(from, until, precision)`, backed by an expression index
+so the as-of scan stays indexed; the query space exposes the same
+canonicalisation as `memory_v1.canonical_bounds` with a companion
 `claims_canonical` view exposing `canon_start`/`canon_end` beside the raw
 columns, so saved examples, open SQL, the catalog metadata and the
-open-query prose use the same canonicalisation as the engine; the Python
-side calls the same definition. **Fact windows are stored canonical** —
+open-query prose use it too; the Python side (`core/temporal.py`) mirrors
+the same table and a test pins the twins equal. **Fact windows are stored canonical** —
 `valid_from`/`valid_until` and `occurs_from`/`occurs_until` are written from
 `canonical_bounds` at seeding, so every fact predicate the engine already
 has (`valid_until > :as_of`, `tstzrange(valid_from, valid_until)` in the