…and wire the trackers up (#1523)

* fix(metrics): aggregate Prometheus metrics across uvicorn workers

With MEMMACHINE_WORKERS > 1 uvicorn forks a worker per process and each
keeps its own registry, so /metrics returned whichever worker happened to
answer the scrape - about 1/N of the traffic, a different 1/N each time.
Consecutive scrapes then look like counter resets, and every rate,
histogram quantile and calls-per-request figure derived from them is
wrong in an unbounded direction.

Build a fresh registry per scrape and attach MultiProcessCollector when
PROMETHEUS_MULTIPROC_DIR is set, leaving the single-worker path on the
default registry so nothing changes for the default deployment.

Measured on an 8-tenant capacity run: counter resets over one ladder went
from 70 (4 workers) and 87 (8 workers) to 0, and five consecutive scrapes
of a loaded server returned identical values where they previously
diverged.

Requires PROMETHEUS_MULTIPROC_DIR to be set and a writable directory
mounted at it; the chart change that does so is separate.

* fix(metrics): hand every OperationTracker a metrics factory

OperationTracker accepts metrics_factory=None and then silently discards
every timing it takes - no error, no warning, no series. A component that
is fully instrumented but never given a factory is therefore
indistinguishable from one that was never instrumented at all, and the
only way to notice is to go looking for a metric that should exist.

The Neo4j store, the episode store and the session store each shipped
instrumented and unwired, which is why database latency appeared to be
unmeasurable: every call was timed and thrown away. Wire the factory
through the resource manager, the database manager and the event
backend's params so those trackers emit.

The tests assert at the call sites rather than on the components. An
earlier version tested that each store honours a factory it is given,
which passes whether or not anything passes one - it still passed with
the fix reverted. These fail when the wiring is removed.

* fix(metrics): choose a multiprocess directory when workers > 1

PROMETHEUS_MULTIPROC_DIR had to be set by whoever deployed the server. Set
MEMMACHINE_WORKERS=4 without it and nothing complains: each worker keeps its
own registry, a scrape is answered by whichever worker the load balancer
picked, and the numbers that come back look plausible. There is no error to
notice, so the only way to find out is to compare a counter against a request
count and see it come up short by a factor of the worker count.

The worker count is the thing that decides whether aggregation is needed, so
read it here and default the directory when it is above 1. An explicit setting
still wins, which is how two servers on one host keep their metrics apart. If
the chosen directory cannot be created the variable is removed again -
prometheus_client raises in every worker if it cannot open the directory it is
pointed at, and taking the server down over metrics would be the wrong trade.

Worker-count parsing moves into _worker_count() so start_http() and the
directory setup cannot disagree about it.

Verified: uvicorn spawns workers via multiprocessing.get_context("spawn"), so
children re-import in a fresh interpreter and inherit the variable set here.
13 new tests cover the default, the explicit override, nested creation, stale
file clearing, and both failure paths; 1228 server/common tests pass; ruff
check, ruff format --check and ty are clean (17 ty diagnostics, all
pre-existing and none in these files).

Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01Nr9kacmpFVTTfkZRw6esxP

---------

Co-authored-by: Claude Opus 5 <noreply@anthropic.com>

…odes (speedkick) (#1547)

* fix(event-backend): make expand_context return timeline-neighbor episodes

Fixes #1540.

On the event backend, expand_context was silently inert: EventMemory
fetched and materialized the expanded segment windows, but
LongTermMemory._search_scored_event read only the seed segment's
_episode_uid and score from each window, and the response schema has no
context field - so responses were byte-identical for expand_context 0
and 5 while every request paid the LATERAL fetch.

The declarative backend, by contrast, folds neighbor episodes into the
returned list (_unify_scored_anchored_episode_contexts). This brings
the event backend to parity:

- Each scored window now contributes the episodes its segments belong
  to (chronological within the window, the seed's episode as nucleus).
- Windows are unified best-score-first with the same fill algorithm as
  the declarative backend: taken whole while they fit within
  num_episodes_limit, then filled by weighted index-proximity to the
  nucleus (forward recall preferred) until the limit is met; an episode
  keeps the score of the first window that contributed it.
- The unified context is returned chronologically, matching the
  declarative backend's ordering contract for expanded results.
- expand_context is clamped to num_episodes_limit - 1 (declarative
  parity).

expand_context == 0 behavior is unchanged (score-ordered seeds, exactly
as before). Reranked configurations gain the same folding on top of
reranker-scored windows.

Tests: end-to-end via the in-memory event-backend wiring (neighbors
returned, chronological order, limit respected) and unit tests for the
window-to-episode-uid extraction and the unification algorithm
(whole-context fit, overflow proximity with forward preference,
first-window score retention).

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>

* style: ruff format

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>

* fix(event-backend): clamp expand_context above zero, and make the
expansion tests actually discriminate

Self-review of the two commits above turned up one defect and one hole.

Defect: the quota clamp `min(max(0, expand_context), num_episodes_limit - 1)`
goes negative when `num_episodes_limit == 0` -- reachable, since
`SearchMemoriesSpec.top_k` carries no lower bound. `EventMemory._query`
then derives `max_backward_segments = -1 // 3 = -1` and hands the segment
store a negative window, which the SegmentStorePartition contract does not
define: the SQLAlchemy store happens to short-circuit on `<= 0`, the
in-memory store computes an empty slice and drops the seed. Apply the floor
last so the clamp can only ever produce a non-negative window.

Hole: neither end-to-end test could tell the fix from its absence -- both
pass unmodified against the pre-fix `long_term_memory.py`. `FakeEmbedder`
maps text to `[len(text), -len(text)]`, so under cosine every document
scores exactly 1.0 against every query; all seven timeline episodes become
seeds of equal rank, ties keep insertion order (which is chronological),
and `num_episodes_limit=7` returns all seven with or without expansion. The
"expansion adds episodes" assertion compared a limit-2 search against a
limit-7 one, so the limit alone explained the difference.

Embed on a keyword instead: only `tl-3` matches the query, so `tl-4` and
`tl-5` -- which score zero -- can reach the result only through the
expansion. The tests now pin the exact window (`[tl-3, tl-4, tl-5]`,
chronological, each keeping the window's score), the clamp against an
oversized `expand_context`, and the non-negative window above. All three
fail against the code they cover.

Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01PVtg6Zea292Pb9L7GXnTJp

* test(event-backend): assert the expansion contract, not a ranking

The tests added in the previous commit discriminate, but they pin an
outcome: an exact episode list (`[tl-3, tl-4, tl-5]`) and exact score
values. Both are properties of the fixture's ranking and of the
backward/forward split `expand_context // 3`, neither of which the fix
claims -- change the split or the scoring and the tests fail while the
behaviour under test is still correct.

Restate them as the contract. Each episode now gets its own similarity from
an explicit search rank, with the match's four timeline neighbours ranked
last, so:

- no correct top-k can return those neighbours, and any nonzero window
  around the match reaches at least one of them whatever the split. The
  assertion is "expansion returned a neighbour the search itself would not",
  plus chronological order and the episode limit.
- the clamp is asserted on the call made to the segment store
  (0 <= backward + forward <= limit - 1, over several limit/expand_context
  pairs) rather than on which episodes come back.
- `expand_context == 0` is asserted as "matches only, best score first",
  without naming them.

Exact lists and score values stay in the unit tests, which own the fill
algorithm and the score-retention rule and are meant to track them. All
three expansion tests still fail against the code they cover.

Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01PVtg6Zea292Pb9L7GXnTJp

---------

Co-authored-by: Claude Fable 5 <noreply@anthropic.com>

The calculator turns a design peak in operations per second, plus a traffic
mix of adds, plain searches and agent-mode searches, into a hardware order:
API servers, vector-store machines and their RAM size, a PostgreSQL server,
embedding and agent-model GPU cards, hot vector RAM and disk for a chosen
retention period, the max_connections PostgreSQL will need, north-south and
east-west network peaks, and how many callers that capacity holds. It also
works backwards, turning a caller population into the capacity it demands.

Every number it prints is labelled measured, derived, estimate or assumption,
and every input is a command-line flag and a box on the web form, so a reader
changes any of them without editing code.

Two things that sound alike are kept apart throughout. Agent-mode search is a
property of one REQUEST - the agent_mode flag, which fans one search out into
about 22 - and it is a cost multiplier. An automated client is a property of a
CALLER - a program sending requests in a loop - and it is a rate multiplier.
Each caller population carries its own traffic mix and the two are blended,
because a person can send agent-mode searches and a program need not.

The unit everywhere is a session sending requests at the same moment, whoever
is behind it: one developer driving a ten-user load test is ten sessions. A
user count is not a session count, so converting one to the other takes two
figures - the share of users active at the busiest moment, and the sessions
one active user holds. Both ship as example defaults, labelled as such, with
a warning naming which were defaulted and what to replace them with. The
share is 10 per 100 people: a convention, not a measurement, plausibly 5 to
20, and every machine count moves with it.

The two per-caller rates carry published sources. A human chat session at
0.011 to 0.028 operations per second is the median to about the 90th
percentile of a session's busiest five minutes, measured across 55,295
sessions in the BurstGPT dataset (arXiv:2401.17644). An automated client at
0.4 is TraceLab's measured 5.0-second median step across roughly 4,300
production agent sessions (arXiv:2606.30560). Both reports also carry the
figures those sources revealed and no count uses: a heavy session at 0.06,
and an automated client at 0.07 sustained, six times below its burst because
an agent is idle most of the wall-clock time.

Sizing for the worst sustained five minutes follows ITU-T E.500, which
requires read-out periods greater than five minutes so that resources are not
dimensioned for infrequent small-interval peaks.

It deliberately does not price high-availability additions: a second copy of
every vector, a PostgreSQL standby or a second gateway are a separate
decision.

476 tests, standard-library unittest, covering the arithmetic at exact
machine boundaries, every subcommand, the web form driven end to end, and
rejection of bad input on every path. ruff check is clean.

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01DiqwYpXVGDet5NiHjdU7Ng

#1532)

QdrantVectorStore was the fourth component built without a metrics factory.
OperationTracker accepts metrics_factory=None and then discards every timing
without an error, so the store looked instrumented and emitted nothing - the
same defect as the Neo4j store, the episode store and the session store, which
is why no Qdrant latency was observable.

QdrantConf gains MetricsFactoryIdMixin so it can resolve one, and
database_manager passes it through.

The Dockerfile gains an EXTRAS build arg on all four uv sync lines. Without
--extra qdrant the qdrant-client package is never installed and the provider
raises ModuleNotFoundError on every request, so an image intended to exercise
this path could not run at all.

test_qdrant_creates_vector_store pinned the exact params and had to change.
It now asserts metrics_factory is not None rather than pinning it: passing the
keyword is not the property worth guarding, since None is accepted and silently
discards everything. Removing the wiring fails it; 1228 common/server tests
pass, ruff check and format are clean.


Claude-Session: https://claude.ai/code/session_01Nr9kacmpFVTTfkZRw6esxP

Co-authored-by: Claude Opus 5 <noreply@anthropic.com>

#1551)

Every resource a search needs is lazy-loaded behind a lock, so the first
search into a fresh worker imports its module tree and builds pydantic
schemas for it. That work is synchronous, so it blocks the event loop, and
it is paid once per uvicorn worker because each worker is a separate
process. Kubernetes has already routed traffic by then: /api/v2/health
touches none of it and answers in about a millisecond while a search on the
same worker takes seconds.

start() already resolves the session data manager. This extends that to the
rest of the search path - episode storage, the episodic memory manager, and
the embedder, reranker, vector store and segment store named by the
long-term-memory config - so the cost lands in front of the readiness probe
where it belongs.

Failures are logged and swallowed. Each of these is retried lazily on the
request path anyway, so a store that happens to be unreachable at boot
should cost a slow first request rather than a process that will not start.
The long-term-memory config is polymorphic over the backend, so the resource
names are read with getattr rather than assuming a concrete class.

Verified on a deployed platform (k3s, four uvicorn workers, 12k episodes,
Qdrant and PostgreSQL on separate hosts) by mounting this file over the
installed package - byte-identical to upstream/main beforehand, so the only
variable was this change. First search after a restart, three restarts each:

  stock       2.22 / 2.69 / 3.15 / 2.74 s     (mean 2.70)
  with warm   1.40 / 0.92 / 1.00 s            (mean 1.11)

No "Could not warm" warnings, so every resource resolved cleanly. Rollout to
Ready was unchanged - 41.8 / 43.8 s with the fix against 43.1 / 42.8 s
without - because the startup probe's 10 s period absorbs the shift.

The remaining ~1.1 s is not addressed here: partition creation is
session-scoped, so a project's first search still pays its own first touch.

ruff check and ruff format --check both pass.


Claude-Session: https://claude.ai/code/session_01Nr9kacmpFVTTfkZRw6esxP

Co-authored-by: Claude Opus 5 <noreply@anthropic.com>

…covers (#1552)

A pooled connection whose peer stops responding is not an error, it is a
wait. The kernel retransmits with exponential backoff - observed at
tcp_retries2 attempt 13, Send-Q stuck at 34 bytes and unchanged over
20 seconds - and asyncpg has no deadline of its own, so the request blocks
for the whole of it. pool_pre_ping does not help: its liveness SELECT is
written into the same dead socket and waits with everything else.

The failure this produces is worse than a slow request. The pool never
discards the dead connections, so the process stays wedged after the network
recovers and only a restart clears it. Seen twice on a benchmark deployment:
searches timing out indefinitely while /api/v2/health answered in about a
millisecond, four uvicorn workers idle in select(), no lock contention in
Postgres, and a fresh connection from the same pod completing in 18 ms.

create_async_engine was passed no connect_args, so nothing bounded either
the statement or the connect. This adds command_timeout and connect_timeout,
defaulted to 60 s and 10 s and settable to null for the old behaviour. Both
are gated on driver == "asyncpg": they reach asyncpg.connect() and mean
nothing to aiosqlite or aiomysql. The keyword assembly moves into a helper,
because adding a branch to async_get_sql_engine pushed it past ruff's
complexity limit and the repeated "if not None" lines were asking for it.

Verified by fault injection on a deployed platform - iptables DROP on the
pod's traffic to Postgres:5432, so pooled connections wedge exactly as they
did in the incident:

                        under the fault          after the fault is removed
  stock        no response at 150 s (cap)    still hung, no response at 90 s
  with fix     500 after 194 s               200 in 1.5 s

The recovery is the point: the stock build stays wedged once the network is
healthy again, which is what forced the restarts.

Under a total outage the request still takes ~194 s to fail rather than
~60 s. A search touches three separate stores, so it pays a timeout per
store; the bound is roughly command_timeout times the number of stores
touched, not command_timeout. Lowering the default would shorten it
proportionally - 60 s is chosen as generous against real query times here,
which are milliseconds, and is left tunable rather than tuned.

ruff check and ruff format both pass.


Claude-Session: https://claude.ai/code/session_01Nr9kacmpFVTTfkZRw6esxP

Co-authored-by: Claude Opus 5 <noreply@anthropic.com>

…1556)

qdrant-client is an optional extra (packages/server/pyproject.toml:67).
#1532 gave the Dockerfile an EXTRAS build arg so a build can include it,
but the publish workflow never passed one, so every image on Docker Hub
lacks the package: the core starts cleanly and then raises
ModuleNotFoundError on the first request that uses the event backend.
The platform chart can now select that backend, so there is no tag it
can point at.

Both the CPU and GPU builds pass EXTRAS=--extra qdrant unconditionally.
The GPU sync line becomes `--extra gpu --extra qdrant`.

Verified: uv.lock already carries qdrant-client under
`marker = "extra == 'qdrant'"`, so `uv sync --frozen --extra qdrant`
resolves without relocking. Parsed the workflow with PyYAML and confirmed
build-args is exactly ['GPU=...', 'SCM_VERSION=...', 'EXTRAS=--extra qdrant']
for both jobs -- the rationale sits in a YAML comment above the key, not
inside the block scalar, where it would have reached Docker as a literal
build arg.

Not verified: no image was published from this branch. The workflow is
tag-triggered and was not dispatched, so the built image has not been
run against a Qdrant deployment.


Claude-Session: https://claude.ai/code/session_01Nr9kacmpFVTTfkZRw6esxP

Co-authored-by: Claude Opus 5 <noreply@anthropic.com>

)

Dispatching this workflow with a pre-release tag publishes one image, not two.
metadata-action's semver type drops its pattern when the parsed version has a
pre-release component, so `v0.3.9-post1` produced the bare tag `0.3.9-post1`
for both jobs instead of `v0.3.9-post1-cpu` and `-gpu`. The CPU build pushed
first, the GPU build overwrote it twenty minutes later, and nothing reported an
error: the tag simply held a 2.9 GB GPU image under a name meant for CPU.

Observed on run 33439754567, where both jobs logged
`tags: memmachine/memmachine:0.3.9-post1` and pushed distinct manifests to it.

This has never been hit before because every release so far used a plain
vX.Y.Z, for which the pattern applies normally. It is not specific to this
branch - the rules are character-identical on main.

Two changes:

- The dispatch version tag is now type=raw with the variant appended, which
  cannot lose the suffix. For a release input it produces exactly what semver
  did (v0.3.9 -> v0.3.9-cpu), so nothing changes for the existing flow; for a
  pre-release it produces v0.3.9-post1-cpu. Simulating both rule sets over both
  input shapes gives no overlapping tag between the jobs.
- A `variant` input (both | cpu | gpu, default both) so a caller can skip the
  26-minute GPU build they do not want. build-gpu needs always() for this,
  because a deliberately skipped CPU job would otherwise skip it too; the
  result check still stops a *failed* CPU build from letting it through.

The `latest` rules are untouched and stay guarded by !contains(inputs.tag,'-'),
so a pre-release still cannot move latest/latest-cpu/latest-gpu.

Note for anyone pinning: after this, rebuilding the tag v0.3.9-post1 publishes
v0.3.9-post1-cpu rather than 0.3.9-post1. The existing 0.3.9-post1 image is
left as it is.

Not verified: no build was dispatched from this branch. The tag sets above were
derived by evaluating the rules, not by running the workflow.


Claude-Session: https://claude.ai/code/session_01Nr9kacmpFVTTfkZRw6esxP

Co-authored-by: Claude Opus 5 <noreply@anthropic.com>

…keys (fixes #1544, #1546, #1549) (speedkick) (#1548)

* Fix: Detach segment store partitions before dropping them, and lock partition metadata on delete

Deleting a partition on PostgreSQL dropped its child tables with CASCADE.
The foreign key from segment_store_dv_ln to segment_store_sg is declared on
the partitioned parents, so the CASCADE dropped the parent-level constraint
rather than only the part belonging to the deleted partition. After the
first partition deletion the store stopped enforcing the link for every
remaining partition, and ON DELETE CASCADE stopped removing derivative
links with it, so delete_segments left orphaned rows that
get_derivative_uuids_by_segment_uuids still returned. Detaching each child
before dropping it keeps the constraint and the cascade intact.

delete_partition also took only a row lock on the partition row. ROW SHARE
does not conflict with the SHARE ROW EXCLUSIVE table lock the create paths
take, so a concurrent create and delete could reach segment_store_sg and
segment_store_dv_ln in opposite order and deadlock; that reproduced in 4 of
7 sampled interleavings against PostgreSQL 16, and in 0 of 7 once delete
takes the same table lock first. The row lock stays, because it is what
makes delete wait for in-flight writers holding FOR SHARE on that row.

Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01MpFwBnZ6SxdMe3pSuMTCHD

* fix: address partition child tables directly for segment DML

Fixes #1546. Parent-table queries carry the partition key as a bind
parameter; once asyncpg's prepared statement flips to a cached generic
plan (after five executions) PostgreSQL locks every child partition on
every execution before runtime pruning. With hundreds of partitions and
concurrent sessions this exhausts the lock table (searches fail 500
'out of shared memory') and saturates the database CPU with lock churn
and generic-plan startup.

The partition handle now maps the ORM entities onto its own child
tables (orm.aliased with adapt_on_names) and targets them for
insert/delete, so every plan references exactly one partition. SQLite
keeps the parent tables (it has no children).

Measured on a store with 316 partitions: max relation locks held by a
backend during a read loop drops 1276 -> 4; the 239 HTTP 500s in a
4-worker load test disappear; throughput at 128 concurrent requests
rises ~20-30% with PostgreSQL no longer pinned at its CPU cap.

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>

* Test: Pin the partition-delete lock ordering

The foreign-key half of this branch has a regression test; the lock
ordering did not. A deadlock test would be timing-dependent, so assert the
invariant the deadlock analysis rests on instead: delete_partition issues
LOCK TABLE segment_store_pt IN SHARE ROW EXCLUSIVE MODE, and issues it
before any DETACH or DROP of a child table. Without the lock the test fails
and reports the statement sequence it saw.

Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01MpFwBnZ6SxdMe3pSuMTCHD

* Style: Apply ruff format to the lock-ordering test

Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01MpFwBnZ6SxdMe3pSuMTCHD

* test: segment DML must address only the partition's child tables

Regression test for #1546. Captures the SQL the partition handle emits
across add / seed read / windowed read / filtered read / uuid maps /
delete and asserts no statement references the partitioned parents --
the deterministic observable of the generic-plan lock explosion (lock
counts would need timing-dependent pg_locks sampling). Fails against
the parent-table implementation, passes with per-partition DML.

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>

* style: drop annotation-only AliasedClass and Table imports

This SQLAlchemy version exports no public AliasedClass name (only the
aliased() factory), so the attribute annotations forced an import from
sqlalchemy.orm.util. The annotations were documentation only; the
branch comment already records that the attributes hold either the ORM
class or its child-table alias.

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>

* refactor: derive partition table names from one helper and the models

The parent names come from the models' __tablename__ and the child
naming pattern lives in _pg_child_table_name, used by child-table
creation, teardown, and the per-partition DML targets, so the three
sites cannot drift apart. Physical names are unchanged; the regression
tests keep literal names to pin the on-disk naming contract.

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>

* Fix: Drop detached children, and stop holding the store-wide lock while waiting for writers

Three follow-ups from review of this branch.

The child-table probe asked whether the table exists, but the statement it
guards is DETACH PARTITION, which requires that the table be attached. A
child left detached by manual maintenance or an interrupted DETACH
PARTITION CONCURRENTLY (that form is not transactional) passed the probe and
made DETACH raise "is not a partition of", rolling back the transaction, so
the partition became neither deletable nor recreatable -- a state the
CASCADE drop this branch replaced used to clean up. Probe pg_inherits for
attachment instead, in one round trip for both children, and drop an
unattached child directly.

delete_partition took the partitions-table lock before the row lock that
waits for in-flight writers, so a slow writer on one partition stalled
open_or_create_partition for every partition, which runs on the request
path. Take the row lock first; the table lock only has to be held across the
child DDL for the deadlock argument to hold. Re-measured: 4/7 sampled
interleavings deadlock with no table lock, 0/7 with either ordering.

Tests: the foreign key is now asserted to be enforced after a partition
delete, not only that the cascade fires -- the PR's measurements list those
as separate things the CASCADE drop broke. A second test leaves a child
detached and requires deletion to succeed and the key to be reusable.

Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01MpFwBnZ6SxdMe3pSuMTCHD

* fix: memoize partition entities per key; validate keys in the naming helper

Review follow-ups on the child-table change:

- The child Table objects and aliases are now built once per partition
  key (functools.cache) instead of per handle. SQLAlchemy's compiled-
  statement cache keys on the Table objects a statement references, so
  per-handle tables made every handle's statements recompile and
  polluted the cache for everything else (verified: cache keys differed
  across handles for the same partition; now identical).
- The tables carry columns only. The to_metadata copies dragged along
  foreign keys with unresolvable targets and duplicate index names --
  latent hazards for anything walking that MetaData.
- _pg_child_table_name validates the partition key itself, so every SQL
  string built from a child table name (including the DDL literals) is
  safe by construction rather than by call-site convention; the store's
  validator moved to module level beside it.

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>

* refactor: make partition key validation part of the segment store contract

Partition keys are embedded in native storage identifiers by any
implementation, so the alphabet/length rule is interface-level, not a
SQLAlchemy detail: validate_partition_key now lives in the package's
data_types (exported from the package), the SegmentStore.create_partition
docstring states the contract, and the SQLAlchemy store imports it.
Deliberately NOT unified with the vector store's identical identifier
rule: the repo-wide naming contract is not wired through yet, so the
convergence is treated as incidental for now.

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>

* refactor: move partition key validation to segment_store/utils.py

Mirrors the vector store's layout (validate_identifier in
vector_store/utils.py); the interface docstring states the key rule
plainly instead of referencing a code path.

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>

* style: state partition key naming constraints in the VectorStore format

Same ABC-level 'Naming constraints:' block the vector store uses, no
method-level restatement, and the length limit is enforced and
documented in bytes, matching validate_identifier.

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>

* style: state the key rule as the regex, not a prose fragment

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>

* style: restore the ABC's original naming-constraints docstring

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>

* refactor: make validate_partition_key boolean, call sites raise

Mirrors the vector store's validate_identifier: the predicate returns
bool so callers can compose it, and each entry point raises its own
error in the vector store's message style.

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>

* Revert "refactor: make validate_partition_key boolean, call sites raise"

This reverts commit 88be86603c36e4b759da0f876968e6dd6db8d913.

* fix: bound the partition-entity cache with lru_cache(4096)

functools.cache grew ~29 KiB per distinct partition key (measured) for
the life of the process, including deleted partitions. The LRU cap
bounds it at ~115 MiB per worker; eviction is harmless since a rebuilt
entry is identical and only costs recompiling that partition's
statements once.

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>

* style: drop point-in-time memory figures from the cache comment

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>

* fix: address second review round

- Type the partition entities: _pg_partition_entities returns a
  NamedTuple with real field types instead of a positional tuple of
  object, which was adding 55 ty diagnostics (the CI static check would
  have failed) and forcing blind unpacking; callers and the memoization
  test use named fields, and the private _generate_cache_key assertion
  (redundant given object identity) is dropped.
- DROP TABLE gains IF EXISTS back, so an out-of-band drop landing
  between the state probe and the drop cannot leave a partition
  half-deleted.
- open_or_create_partition opens existing partitions without the
  store-wide management lock (double-checked: unlocked read, then lock
  and re-check only when creating), so request-path opens no longer
  serialize behind a concurrent deletion's DDL window; pinned by
  test_open_existing_partition_takes_no_management_lock.
- The engine's compiled-statement cache is raised from the default 500
  (per-partition statements would thrash it once enough partitions are
  live concurrently).
- Comments and the DML test docstring scope the lock claim honestly:
  PostgreSQL's FK integrity triggers still address the parents
  internally, costing a one-shot per-backend lock spike on
  writes/deletes when a trigger plan first goes generic (verified:
  ~10 locks steady, one spike at execution six, then back) -- tracked
  on #1546.
- The detach test cleans up its detached child unconditionally so a
  failure cannot poison the session-scoped container for later tests;
  the statement recorder is a shared fixture instead of copy-paste;
  the byte-length check encodes once.

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>

* refactor: drop the typing casts from the partition entities

inspect(Model).columns yields real Column objects (the stubs type
__table__ as FromClause, which forced the cast), and SQLAlchemy's
typing convention represents an aliased entity as the mapped class
type, so the NamedTuple fields are type[SegmentRow] /
type[DerivativeLinkRow] and aliased() assigns without coercion.

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>

* design: shared segment store tables with incarnation-scoped keys

Design record for replacing the per-tenant partitioned layout with
shared tables on every dialect: the tenant registry carries an
incarnation, data rows are keyed by <logical_key>@<incarnation>,
deletion is an O(1) registry write plus a purge queue, and fencing
fails stale handles loudly. Records the measured comparison against
PARTITION OF and standalone-table layouts and the scaling requirements
(cheap tenant creation at 1e5-1e6 tenants, 1e4-1e7 rows per tenant)
that decided it.

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>

* refactor!: shared segment store tables, incarnation fencing, O(1) delete

Implements design/segment_store_shared_tables.md. Fixes #1544, #1546,
and #1549 by construction:

- The ORM models are the physical schema on every dialect; PostgreSQL
  partitioning, per-tenant DDL, the detach machinery, the store-wide
  management lock, and the per-partition entity cache are all removed.
  No partitions means no generic-plan lock fan-out (client or
  RI-trigger) and no DDL for lifecycle deadlocks to live in -- the
  churn smoke that measured 41-83 deadlocks per 20s on every
  partitioned build measures zero, with 60x more write throughput.
- segment_store_pt becomes the tenant registry: partition_key +
  incarnation. Data rows are keyed by <logical_key>@<incarnation>, so
  a deleted-and-recreated tenant never sees its predecessor's rows.
- create_partition is a row insert (no DDL); delete_partition is O(1):
  FOR UPDATE on the registry row (drains writer pins), enqueue the
  physical key on segment_store_gc, delete the row.
  purge_deleted_partitions reclaims rows in chunked background batches.
- Writes pin the registry row FOR SHARE with an incarnation predicate;
  reads check it too: a stale handle raises
  SegmentStorePartitionStaleError on every dialect, SQLite included.
  Measured cost: one extra registry round trip per read operation.
- The segment table's FK to the registry is removed (registry and data
  rows are deliberately decoupled for O(1) deletion); the link-table FK
  and cascade remain.

Same-moment ABAB vs the partitioned build: ingest and windowed reads at
parity, lifecycle cycles 4-6x faster, tenant creation ~1000x cheaper
(row insert vs DDL).

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>

* refactor: identify tenant data rows by incarnation UUID alone

Data rows drop the composite <logical_key>@<incarnation> string for a
bare incarnation UUID column: a data query cannot be constructed
without resolving the registry, so referencing the wrong tenant is
structurally impossible; index entries narrow from a 41-byte varchar
to a native 16-byte uuid; random UUIDs are globally unique across
nodes without coordination, so tenant moves between databases carry
rows verbatim; and collisions among incarnations with live traces are
rejected by constraints (unique on the registry, primary key on the
purge queue) instead of left to probability. The purge queue keeps the
logical key for forensics.

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>

* docs: finish the physical-key -> incarnation wording in the design doc

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>

* refactor: fence by incarnation alone

The incarnation is unique-constrained, so it resolves the registry row
by itself; the logical-key predicate was a leftover from the composite
string design and contradicted the rule that the incarnation is the
handle's sole authority.

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>

* refactor: rename the stale error to name the handle, not the partition

The handle is what is stale -- the partition is deleted -- and
SegmentStorePartitionHandleStaleError follows the existing noun+state
convention (ConfigMismatch, AlreadyExists).

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>

* refactor: inline uuid4 for incarnation generation

new_incarnation() was a one-line wrapper adding indirection for no
behavior; the multi-node rationale lives in the design doc.

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>

* fix: normalize segment timestamps to UTC before persisting

Segment-store slice of #1462: SQLite's DateTime(timezone=True) discards
tzinfo and stores wall-clock fields verbatim, so a non-UTC timezone-aware
timestamp written without UTC normalization read back shifted by its
offset (13:30:45-08:00 came back as 05:30:45-08:00). The read path
already assumed UTC and reapplies the separately stored offset; only the
write was missing the conversion. PostgreSQL timestamptz stores a true
instant, so this is a no-op there.

Regression test parametrized over UTC/-08:00/+05:30 runs on both
backends; verified the non-UTC params fail without the fix and pass
with it (sqlite 54, pg 58).

The companion filter-bound normalization lives in shared
sql_filter_util.py (used by episode and cluster stores too) and stays
in #1462.

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>

* fix: normalize datetime filter bounds to UTC in SQL filter compilation

Second half of the #1462 segment-store slice: timestamp columns now hold
the UTC instant, so comparison bounds must be named in the same frame.
On SQLite an aware datetime bind is rendered as wall-clock digits with
tzinfo dropped and compared lexically, so `timestamp <=
2024-01-01T08:00+08:00` excluded a row stored at 00:00Z -- the same
instant. _normalize_column_value converts datetime values (Comparison
and In leaves) to UTC before binding; PostgreSQL compares timestamptz
by instant either way, so the two backends now agree.

The helper lives in the shared sql_filter_util because that is where
column leaves are compiled; other stores' write paths (episode, cluster)
are intentionally not touched here.

Regression test parametrized over the same instant named in +00:00,
+08:00, and -08:00, on both backends; verified the non-UTC bounds fail
without the fix and pass with it (sqlite 57, pg 61; full server suite
1880 passed, 3 skipped).

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>

* feat: promote purge_deleted_partitions to the SegmentStore ABC

Physical reclamation of deleted partitions is now an ABC capability:
callers schedule it however they want; the store never schedules it
itself. delete_partition's contract notes that reclamation may be
deferred. Implementations whose deletes reclaim physically implement
it as a no-op returning False.

Signature review against the prior purge iterations (#1199/#1205):
- The old three-step orphan-derivative API (get_orphaned / mark /
  purge) existed only because derivative purging interleaved with
  vector-collection deletes between steps; incarnation purge is fully
  internal to the store, so a single method suffices.
- The old scheduling knob (purge_interval loop in ExtraMemory) lived
  in the consumer -- preserved: no scheduling in the store.
- The bound is max_segments (domain unit; derivative links ride along
  uncounted) rather than max_batches, which presumed chunked-transaction
  implementations. batch_size stays as a keyword on the SQLAlchemy
  implementation only, as a transaction-size tuning knob.
- Returns bool ("reclaimable work may remain") instead of rows deleted:
  a row count cannot distinguish "drained" from "stopped at the bound"
  when a dead incarnation has zero data rows, and the scheduling caller
  needs exactly the more-work signal.

New test pins the bound and the completion signal on both dialects.

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>

* refactor: drop batch_size from purge_deleted_partitions

With max_segments as the caller's bound, a per-call batch_size is
redundant: bounded calls already cap every delete transaction at the
remaining budget, so the knob only governed the unbounded case --
where transaction sizing is engine policy, not caller policy. The
chunk is now an internal constant (_PURGE_CHUNK_SIZE); if a deployment
ever needs to tune it, it belongs in SQLAlchemySegmentStoreParams,
not per call. Tests exercise multi-chunk draining by patching the
constant.

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>

* refactor: purge is one atomic slice per call; fix open_or_create race

Purge contract resolved to atomic-slice-per-call: each
purge_deleted_partitions call is a single transaction that reclaims up
to max_segments rows and either commits that progress or nothing.
Draining a backlog is the caller's loop (call until False), so
reclamation never holds a long transaction, committed slices survive
interruption, and there is no internal chunking competing with the
caller's bound. max_segments=None means the store-chosen slice size
(_PURGE_SLICE_SEGMENTS), keeping engine-appropriate transaction sizing
out of callers' hands. Rationale over the alternatives: cross-call
atomicity is anti-useful for gc (a huge atomic purge is exactly the
long-transaction hazard, and an error would forfeit all progress),
while batch_size+max_batches exposes the store's transaction quantum
and bounds a call only as a product of two knobs.

Also fixes a TOCTOU in _open_or_create_partition caught by the new
lifecycle churn test: losing the insert race and then finding no row
(a concurrent delete removed the winner) raised RuntimeError; the
read-then-insert sequence now retries, since every retry implies
another actor changed the state. New deterministic fencing tests:
test_write_landing_during_delete_is_never_orphaned (the write pin means
rows can never land under an incarnation the purge queue no longer
tracks) and test_concurrent_remote_delete_yields_single_queue_entry
(the delete pin means racing deletions enqueue exactly once).

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>

* test: lock-necessity suite verified by per-lock ablation in both generations

New test_segment_store_locking.py (PostgreSQL integration lane) pins
each locking property through the public API, using only surface shared
with the pre-overhaul partitioned store so the module runs against both
generations. All interleavings are staged event-driven: blocked-ness is
decided by observing pg_stat_activity lock waits, not elapsed time;
there are no grace sleeps, and paused writers are released in finally
blocks so a failing assertion cannot wedge fixture teardown. The lane
adds under a second of CI time.

Ablation matrix (each lock removed one at a time via source-patched
variant trees, PYTHONPATH-shadowed; old = pre-overhaul partitioned
store at 14b8f0a2~1):

- write pin ablated (either generation): write-pin test fails, plus
  the no-orphaned-writes fencing test on the new store.
- delete pin ablated (new store): churn, concurrent-delete, and
  single-queue-entry tests fail (double-enqueue IntegrityError).
- delete row pin ablated (old store): write-pin test fails (the delete
  no longer waits out the in-flight writer).
- ordered delete_segments row locks ablated (either generation): no
  test fails -- identical DELETE shapes lock rows in identical orders
  on PostgreSQL (sorted scalar-array probes, TID-ordered bitmap scans),
  so the AB/BA cycle needs plan divergence the store never produces.
  The overlap test is kept as a regression canary and documented as
  such; whether to keep the pre-lock itself is a separate decision.
- old store with ALL locks intact: churn and concurrent-delete tests
  fail with DeadlockDetectedError in the two cycle shapes documented
  on #1546 (delete-vs-delete lock upgrade over the table mutex;
  create-vs-delete DDL cycles through the shared parents). Those
  deadlocks are inherent to the partitioned layout -- the property the
  shared-table overhaul removes, and these tests now pin.

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>

* fix: purge claims queue entries with SKIP LOCKED; correct lock-order rationale

Concurrent purgers were a real deadlock surface: two processes draining
the same dead incarnation delete overlapping row sets through unordered
scans. Claiming queue entries with FOR UPDATE SKIP LOCKED removes the
contention instead of ordering it -- racing purgers partition the queue,
and only the claiming call touches a dead incarnation's rows (writers
cannot; the fence pins live incarnations only), so reclamation is
deadlock-free by construction. This is the claiming half of the purger
scale-out design in the design doc; the ABC now states the contract
(concurrent calls, including cross-process, must neither error nor
deadlock).

Tests: test_purge_skips_entries_claimed_by_concurrent_purger stages a
purger from another process holding its claim uncommitted -- a
concurrent purge must skip the entry and complete without blocking;
verified to fail (blocks on the held queue row) with the claim ablated
and pass with it. test_concurrent_purges_reclaim_everything pins the
correctness property on both dialects: racing drain loops terminate
cleanly with full reclamation.

Also rewords the ordered-row-lock rationale in the locking suite: the
consistent acquisition order that makes the ablation unobservable is
current PostgreSQL executor behavior, not a guarantee any engine
documents -- the pre-lock imposes the order deliberately, and the canary
catches divergence if an engine or plan change ever produces it.

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>

* test: accept any task result type in _wait_until_blocked_or_done

The helper only observes done-ness; Task[None] rejected the purge
task (Task[bool]).

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>

* fix: reject minting an incarnation whose garbage is still awaiting purge

Data rows are keyed by incarnation alone, so a fresh mint colliding with
a dead-but-unpurged incarnation would adopt its garbage and then be
erased by the purger. The registry's unique constraint only guarded
collisions with live incarnations; the purge-queue case was guarded by
uuid randomness alone.

The mint (shared by create_partition and open_or_create_partition) now
re-checks the purge queue inside the insert transaction and re-mints on
collision. The check is race-free with the existing tables -- no ledger
table needed: it runs after the registry insert, so a concurrent
deletion moving a colliding row to the queue (the insert waited on its
uncommitted registry delete) is already visible, and no new queue entry
for the minted value can appear before commit because the only registry
row carrying it is uncommitted. The locking read sees latest-committed
state on dialects whose plain reads serve transaction-start snapshots;
SQLite serializes whole transactions. An incarnation value can therefore
never be reused while any trace of it remains within a database; across
databases, uniqueness still rests on random-uuid collision resistance.

test_incarnation_with_garbage_left_is_never_reused forces the collision
by stubbing the mint (both creation paths, both dialects); verified to
fail with the re-check ablated and pass with it.

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>

* refactor: one collision error for live and garbage incarnation mints

Both collision causes look the same to the mint's callers and share the
same remedy -- mint a fresh incarnation and retry -- so they now share
one error, with the cause classification (key taken vs incarnation
collision) resolved inside _insert_partition_row: an IntegrityError with
a committed row under the key means the key is taken
(SegmentStorePartitionAlreadyExistsError: open or delete it instead);
without one, the incarnation collided with a live row. Errors are typed
by the decision the caller makes, not by the failing constraint, and
both call sites shrink to one remedy branch per error.

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>

* test: fold locking tests into the segment store test file; drop "slice"

All SQLAlchemy segment store tests live in one file. The separate
locking module existed so the same tests could import against the
pre-overhaul partitioned store for the lock-ablation matrix; that
verification is done and recorded, so the split's constraint is spent.
The per-lock coverage map moves to a section comment.

Also replaces the "one slice per call" purge wording, which was
circular (a slice being defined as whatever one call does), with the
actual contract: each call reclaims at most max_segments segments --
in this store, one transaction that commits that progress or nothing --
and None means the store's default bound (_DEFAULT_PURGE_MAX_SEGMENTS,
renamed from _PURGE_SLICE_SEGMENTS).

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>

* docs: correct design-doc drift; drop dead _is_postgresql flag

Accuracy review of the design doc against the code:
- The create bullet said "one row insert"; the mint transaction also
  re-checks the purge queue.
- The locking model omitted the purger's SKIP LOCKED queue claims and
  the mint's collision-case queue pin; it now lists every row lock and
  why reclamation cannot contend with anything.
- The consequences section claimed the only remaining dialect split is
  the LATERAL-vs-loop read strategy; the PostgreSQL-only ordered row
  locks in delete_segments and SQLite's foreign-key pragma are splits
  too.

_is_postgresql was assigned and never read -- dead since the overhaul
removed the DDL branches.

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>

* feat: make the default purge bound a store construction parameter

SQLAlchemySegmentStoreParams.default_purge_max_segments (default 10000)
replaces the module constant: each purge call is one transaction, so
the right default bound is dialect- and deployment-dependent, and the
construction parameter lets an application set it once instead of
every purge caller reading configuration to pass max_segments.

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>

* test: cover live-incarnation collision and purge-bound params wiring

Coverage audit of the recent additions found two unpinned paths:

- The live half of the mint's collision handling (registry unique
  violation classified by the key re-read, then re-mint) had no test --
  only the garbage half did. test_incarnation_colliding_with_live_
  partition_is_never_reused forces the collision on both creation paths
  and both dialects; verified to fail with the classification ablated
  (create_partition misreports AlreadyExists) and pass with it.
- The purge tests patched the store's default-bound attribute directly,
  leaving the SQLAlchemySegmentStoreParams.default_purge_max_segments
  wiring itself untested. test_default_purge_bound_comes_from_params
  constructs a store with a small configured bound and observes it
  govern an unbounded purge call.

Also converts the two override-method docstrings (delete_partition,
purge_deleted_partitions) to body comments: the contract lives on the
ABC; overrides keep only implementation mechanics.

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>

* refactor: rename _logical_partition_key to _partition_key

The "logical" qualifier contrasted with the physical partition key of
the composite-key era; data rows now carry no key at all, so there is
nothing physical to distinguish from.

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>

* refactor: rename PurgeRow to PurgeQueueRow

The model classes are named for what a row represents (PartitionRow,
SegmentRow, DerivativeLinkRow); a segment_store_gc row is not a purge
but an entry of the purge queue.

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>

* refactor: restore @staticmethod on _resolve_segment_field

It became an instance method when field resolution went through the
handle's per-partition aliased entities; the shared-table overhaul
resolves against the module-level SegmentRow again, leaving self
unused. Call sites return to the original class-qualified form.

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>

* test: pin the mint's insert-then-check statement ordering

The collision guard relies on checking the purge queue AFTER the
registry insert: under READ COMMITTED, the insert's unique-index wait
on a concurrent deletion's uncommitted registry delete is what forces
that deletion's queue entry to be committed -- and therefore visible to
the later check. Checked before the insert, the queue is read too
early and the mint commits a live partition whose incarnation is on
the purge queue, handing its rows to the purger.

Only a concurrent interleave distinguishes the orderings, so the
sequential collision tests cannot pin it: verified by swapping the two
statements -- the sequential tests all still pass (the opposite order
is correct for non-concurrent use), while the new
test_mint_detects_collision_with_concurrent_deletion fails (it is
incorrect for concurrent use). The test stages the interleave
deterministically: a raw-session deletion held uncommitted, the
colliding mint observed blocking on it via pg_stat_activity, then the
deletion committed.

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>

* docs: plain "maximum number of segment rows purged per call" wording

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>

* docs: match params docstring to the pydantic field description

Convention in the class: the Attributes entry carries the field
description plus the default; the field expresses the default via its
default attribute only.

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>

* fix: purge claims queue entries one at a time

The claim SELECT had no limit: it materialized and row-locked every
unclaimed queue entry even when max_segments exhausted on the first
incarnation -- a mass-deletion backlog was fetched wholesale per call,
and the first purger claimed the entire queue, so concurrent purgers
skipped everything and exited instead of sharing the backlog.

Claims are now LIMIT 1 FOR UPDATE SKIP LOCKED, issued as the call
processes entries: a bounded call locks exactly what it works on.
Within the transaction each claimed entry is retired before the next
claim, so the call's own claims (which SKIP LOCKED does not skip)
cannot recur and the loop terminates.

test_purge_claims_queue_entries_incrementally pins the property via
recorded SQL: every queue claim carries LIMIT, and a call whose bound
exhausts on its first incarnation issues exactly one claim; verified
to fail against the previous claim-all form.

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>

* refactor: batch, not chunk, for the purge deletion unit

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>

* refactor: purge runs on an engine connection for typed rowcount

AsyncSession.execute has no DML overload -- it is typed Result[Any] for
every non-typed statement, so reading rowcount needed an isinstance
narrowing to CursorResult (whose unreachable else-branch would have
fabricated a zero count). AsyncConnection.execute is typed CursorResult
in every overload, and the purge transaction is pure Core DML with no
session features, so it now runs on self._engine.begin(): the library's
own annotations carry the type and the narrowing disappears.

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>

* refactor: manual formatting

Signed-off-by: Edwin Yu <edwinyyyu@gmail.com>

* fix: address code-review findings (round 3)

Six confirmed-or-verified defects from a second review session, each
fixed with a test verified to fail on the pre-fix code:

- SQLite write fence was a no-op: the driver defers BEGIN to the first
  data-modifying statement, so the fence SELECT ran outside the write
  transaction and a write racing a delete-plus-purge committed rows no
  queue entry tracked. The write fence (and deletion's row check) now
  issue a no-op registry UPDATE first, opening the write transaction so
  the check is transactional and racing deletions serialize.
- The shared column-leaf UTC normalization silently changed OTHER
  stores' datetime filters: their write paths still store wall clock,
  so on SQLite their filters stopped matching rows they had just
  written. Normalization is now an explicit compile_sql_filter opt-in
  (column_datetimes_are_utc) that only the segment store sets; other
  stores regain their previous behavior, and #1462 flips the opt-in
  for the stores whose write paths it fixes.
- Mint collision retries were unbounded: any persistent IntegrityError
  with the key absent became an infinite hot loop. Both creation paths
  cap consecutive collision retries (_MAX_MINT_ATTEMPTS) and re-raise
  the underlying error -- consecutive failures at that depth mean a
  permanent cause, not a race.
- purge_deleted_partitions accepted non-positive bounds and returned
  True unconditionally, spinning the documented drain loop; it now
  raises ValueError. Empty incarnations charge one segment of budget,
  so a backlog of empty tenants is bounded per call instead of drained
  in one unbounded transaction.
- open_or_create committed the registry row before materializing the
  payload codec, leaving an unopenable partition behind on codec
  failure; the codec is loaded before the insert again.
- validate_partition_key used re.match with $, accepting keys with a
  trailing newline; now re.fullmatch.

Also from the review: the ABC documents the stale-handle contract on
SegmentStorePartition and corrects purge's False semantics (work owned
by a concurrent purger is not counted); the blocked-or-done test helper
scopes pg_stat_activity to the current database.

Rejected findings, with grounds recorded in the PR discussion: the
read-fence round trip is the deliberate loud-fencing contract (#1549);
fence/live-check unification, the forensic enqueued_at column, and
FIFO claiming are declined as taste; the partition-key rule's overlap
with service_locator stays per the incidental-convergence ruling.

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>

* fix: LongTermMemory erasure drains the purge queue inline

delete_partition's physical reclamation is deferred by design, but the
review found its one production caller now leaked: session deletion
previously removed data physically (DROP on PostgreSQL, cascade on
SQLite) and nothing anywhere called purge_deleted_partitions. The
erasure path drains the queue inline before returning, restoring
physical removal semantics; background scheduling remains available to
other callers.

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>

* fix: UTC-normalize every SQL store; self-checking SQLite fence; FIFO purge

Four follow-ups to the review round, per direction:

- The datetime-normalization opt-in is gone: instead of scoping the
  shared compiler's UTC bound normalization to the segment store, every
  SQL store's write path is fixed honestly in this PR. #1462's episode
  and cluster fixes (created_at + start/end bounds; last_ts + pending
  created_at) are ported with their regression tests, and the compiler
  normalizes column datetime bounds unconditionally -- correct for all
  consumers, since the semantic-storage columns it also serves are
  server-generated UTC (func.now()). Fixes #1558 and #1559 here.

- The SQLite fence is one self-checking statement instead of a no-op
  UPDATE plus a SELECT: the proper primitive, BEGIN IMMEDIATE, is only
  expressible engine-wide in SQLAlchemy (it would put every read
  transaction behind the write lock), so the registry-row UPDATE
  acquires the same write lock scoped to the transaction, and its match
  count is the staleness check. Deletion opens its transaction the same
  way, with zero matches as the idempotent no-op case.

- The purge queue is FIFO: claims order by enqueued_at (indexed), so
  the oldest garbage is reclaimed first and the name is honest. Queue
  entries carry their own per-call bound
  (SQLAlchemySegmentStoreParams.purge_max_partitions, default 1000)
  instead of charging a fake segment of budget: their cost is round
  trips rather than row deletions, and empty partitions are cheap to
  mass-create-and-delete, normally or adversarially. Empty entries no
  longer consume max_segments.

- PostgreSQL-only concurrency coverage gains SQLite counterparts
  wherever the property exists on both dialects: lifecycle churn,
  racing deletions (plus a single-enqueue assertion), overlapping
  segment deletes now run on both; new SQLite tests pin the
  mint-vs-deletion collision race and O(1) deletion via recorded SQL.

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>

* docs: precise rationale for the SQLite fence primitive

BEGIN IMMEDIATE is expressible per-transaction in principle, but only
atop engine-wide rewiring (isolation_level=None plus a begin-event
hook) that the store cannot apply to a caller-owned, possibly shared
engine; say that instead of "only expressible engine-wide".

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>

* refactor: two-character index name tokens, matching the prior convention

pk_ev / pk_ts_ev_bk_ix / pk_su used two characters per indexed column;
in (incarnation) and ea (enqueued_at) follow, replacing inc and enq.

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>

* refactor!: purge takes no arguments; cap derivatives at ingestion

purge_deleted_partitions() -> bool. Callers cannot know
engine-appropriate transaction sizing -- the same argument that made
the default a construction parameter removes the per-call override:
the caller's whole protocol is "call until False", and every bound
(purge_max_segments, renamed from default_purge_max_segments;
purge_max_partitions) is implementation policy set once at
construction. Non-positive bounds are now impossible by pydantic
validation, superseding the runtime ValueError.

The derivative side is bounded where it is created, not where it is
reclaimed: purge keeps relying on the link-table ON DELETE CASCADE --
benchmarked against manual link deletion on the real schema and 50-68%
faster (1 link/segment: ~312k vs ~209k segs/s; 4 links: ~266k vs
~158k; the manual pattern's extra round trips and array shipping cost
more than the per-row indexed trigger probes) -- and ingestion rejects
more than max_derivatives_per_segment links per segment (default 100),
so one purge call's work is at most purge_max_segments segment rows
plus that many times the cap in link rows.

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>

* revert: drop the store-level derivatives-per-segment cap

The cap rejected at add_segments time, when the caller has already
segmented and derived and can do nothing to obey it -- the bound on
link fan-out is ingestion-pipeline policy (deriver/segmenter design),
not a store contract. Performance also gives the cap no case: measured
across densities, cascade deletion saturates around 3M link rows/s
(380k segments/s at one link per segment, 302k at 4, 175k at 16, 46k
at 64 -- per-row cost FALLS with density, 1.3us/row at 1 link to
0.34us at 64), so a purge_max_segments=10000 call finishes in ~0.45s
even at 64 links per segment. The design doc records where the bound
lives and the measured sensitivity.

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>

* docs: state the purge contract's promise to callers

The bounds are implementation policy; what the caller is promised is
that a purge call does not noticeably degrade concurrent request
serving. The design doc also records why a store-level link cap would
be unactionable (only the deployment's segmenter/deriver choice can
change the ingested shape, so a dedicated error type would have no
useful handler).

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>

* fix: rebalance the purge entry bound to measured cost

Retiring an empty queue entry measures ~0.95 ms through the store
(four round trips), roughly 200x a segment row at the measured purge
rate -- not the ~10x the old default implied. purge_max_partitions
drops from 1000 (a ~0.95 s transaction when saturated, 20x the row
bound's ~46 ms) to 50, putting a full-entry call and a full-row call
at comparable transaction duration. Backlog drain throughput is
unchanged (~1k entries/s regardless of slicing); only per-call
transaction length shrinks.

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>

* fix: public SegmentStorePermanentError; power-of-ten entry bound

Mint-retry exhaustion raised "last_collision.__cause__ or
last_collision" -- expedient plumbing that leaked either the wrapped
SQLAlchemy IntegrityError or the private collision type to callers.
Per the error-design principle (type by the caller's decision), the
decision here is "retrying will not fix this; diagnose", so both
creation paths now raise the ABC-declared SegmentStorePermanentError
with the underlying error chained as the cause. The ABC documents it
on create_partition and open_or_create_partition.

purge_max_partitions defaults to 100 instead of 50: sibling fields of
one config keep to the same numeric family (powers of ten, alongside
purge_max_segments=10000); a saturated entry call (~95 ms measured)
and a saturated row call (~46 ms) stay within the same order of
transaction duration.

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>

* refactor: rename SegmentStorePermanentError to SegmentStoreRetriesExhaustedError

"Permanent" asserted a diagnosis the store cannot make -- sustained
adversarial churn could in principle clear on a later attempt. The
name now states only what happened (internal retries exhausted), with
the guidance phrased as likelihood: an immediate retry is unlikely to
succeed; diagnose the chained cause.

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>

* docs: drop illustrative examples from contract docstrings

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>

* fix: bound open_or_create's lost-race arm with the same retry cap

The collision arm was capped but the AlreadyExists arm looped
unboundedly -- the reviewer's livelock finding. Both non-terminating
outcomes now count toward one retry budget, and exhausting it raises
SegmentStoreRetriesExhaustedError with the last error chained. With
this, every retry construct in the store is bounded: purge makes
guaranteed progress per call, deletion is a single idempotent
transaction, fences raise stale, and reads are single-pass -- the
creation paths were the only sites with retries to exhaust.

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>

* refactor: attempts, not retries

SegmentStoreAttemptsExhaustedError, with the counter and docstrings
using the same word: "retry" is ambiguous between a re-attempt and the
whole attempt sequence, and _MAX_MINT_ATTEMPTS already counted
attempts.

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>

* docs: attempts vocabulary in the mint-exhaustion message

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>

* nit: increase max mint attempts from 8 to 10

Signed-off-by: Edwin Yu <edwinyyyu@gmail.com>

* nit: manual formatting

Signed-off-by: Edwin Yu <edwinyyyu@gmail.com>

* review: fold the read fence into the data statement; harden creation, drain, purge

Third review round (15 findings; 12 acted on, 3 declined with grounds
in the PR body).

- Reads no longer issue a separate registry round trip: the liveness
  predicate rides in each data statement as an EXISTS conjunct (one
  statement, one snapshot -- a stale handle reads nothing), and the
  registry check is issued on its own only when a read returns no rows,
  to tell an empty partition from a stale handle. The write fence and
  the read check share one query builder (`_registry_row_query`) and
  one checker (`_ensure_partition_live(pin=...)`).
- `create_partition` materializes the payload codec before inserting,
  like `open_or_create_partition`; its mint loop uses the same
  attempt-counter idiom and message as the other path, which also
  removes the possibly-unbound `last_collision`.
- `drop_session_partition` nulls its handles before the inline drain,
  so a drain failure cannot leave them pointing at deleted resources;
  the drain's comment states exactly what it guarantees (the queue is
  global, the drain uncapped, and an entry a concurrent drain claimed
  is finished by that drain).
- The purge queue's enqueue stamp is the database clock (`now()`), so
  every server's entries order on one clock; the unreachable
  `remaining <= 0` guard is gone; the purge comment and design doc
  state SQLite's actual claiming behavior (plain read, serialized on
  the database write lock at the DELETE; duplicated round trips only).
- `startup()` refuses the old partitioned layout (registry without the
  incarnation column) with a directive to recreate the schema, instead
  of letting create_all leave the old tables in place for an opaque
  missing-column error later.
- Cluster store reads use the shared `ensure_tz_aware`; the private
  clone is deleted. Contract wording: "every data operation" raises
  the stale-handle error (the config property never did).

Tests: unloadable codec guard parametrized over both creation paths,
FIFO pinned with explicit stamps set against insertion order, the
database-clock stamp and the folded liveness check pinned via recorded
SQL, the startup probe on both dialects, and the LTM nulling order
under a failing drain. The codec and nulling tests were each verified
to fail with their fix ablated.

Read-path ABAB against the previous HEAD (interleaved rounds, medians):
seed context reads 1.17 vs 1.42 ms, event lookups 1.04 vs 1.61 ms,
derivative lookups 1.05 vs 1.33 ms (5 rounds), windowed context
expansion 8.74 vs 9.67 ms (8 rounds x 600 reps, paired median
-0.91 ms); reads that find nothing unchanged (two statements either
way). Server-side EXPLAIN ANALYZE: the EXISTS conjunct plans as a
one-time InitPlan (~3 us per statement); a windowed read's 3
statements execute in 0.069 ms vs the previous 4 statements' 0.067 ms.

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>

* docs: purge bound scales with link fan-out; queue stamp is transaction time

Follow-up notes from the review session: the purge_max_segments
description says its derivative links cascade uncounted, so a call's
transaction also scales with the deployment's links per segment (the
promise in the ABC is kept by sizing this bound with that fan-out in
mind, which is the deployment's knob, not the caller's); the enqueue
stamp comment records that PostgreSQL's now() is transaction-start
time and that one deletion per transaction makes it one stamp per
entry.

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>

* docs: a deleted partition's handle is permanently invalid

"Obtain a fresh handle to continue" read as if deletion-and-recreation
were a routine flow; the contract is simply that deletion permanently
invalidates the handle, including against a later same-key creation.

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>

* revert: drop the old-layout startup probe

Handling pre-existing partitioned-layout deployments is out of scope
for the opt-in, pre-GA event backend; existing databases recreate
their schema, and startup stays a plain create_all.

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>

* nit: slim the purge claiming comment

The code comment keeps only the invariants the loop relies on; the
full rationale stays in design/segment_store_shared_tables.md, which
the comment now points at.

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>

* nit: params docstring matches field descriptions, defaults at the end

The purge bounds' field descriptions carry the full text and the
docstring repeats them verbatim, with (default: N) moved to the end
of each description per the params convention.

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>

* review: second-round fixes across locator, stores, and race tests

Second review round from the local review session (9 findings; 7
acted on here, the background-purger suggestion lands separately, and
the unbounded link-retire guard is kept with its tradeoff stated in a
comment -- bounding it would add a budget for a case that indicates a
broken schema).

- partition_key_for_session validated with a drifted private copy of
  the store's key contract: its re.match passed a trailing-newline
  session id through unhashed, and the store's re.fullmatch then
  hard-rejected it, failing session creation where hashing would have
  succeeded. The copy is deleted; the locator (and its tests) now call
  the store's own validate_partition_key, and the hash slice length
  comes from the now-public PARTITION_KEY_MAX_BYTES, so the two can
  never disagree again. Regression test verified to fail pre-fix.
- The SegmentStorePartition contract states that a call with empty
  input does no work and returns without checking the handle -- the
  empty-set guards return before any fence, which the docstring's
  "from then on" overstated.
- delete_partition on SQLite resolves the incarnation in the pin
  UPDATE itself via RETURNING; the locking select is PostgreSQL's path
  only, removing SQLite's extra round trip and its unreachable
  row-is-None branch.
- _open_or_create_partition loads the payload codec only on the create
  path (still before any registry write); opening an existing
  partition no longer materializes a codec it discards.
- Episode-store reads use ensure_tz_aware instead of an inline clone
  in the same file that imports it for writes.
- The purge's link-retire guard comment states it is normally a
  zero-row delete and unbounded only if referential integrity was
  actually broken.
- The two SQLite race tests gained started-events proving the racing
  task ran before the sample, so a loaded box cannot pass them
  vacuously by never scheduling it; the remaining grace periods are
  annotated (SQLite exposes no lock-wait state to observe).

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>

* feat: background purge tick in the resource manager

Nothing but the inline drain in drop_session_partition ever called
purge_deleted_partitions, so a drain interrupted by a crash or a
dropped connection left its queue entry (and the partition's rows)
waiting for the next session deletion anywhere in the deployment.

The resource manager -- the component that owns each segment store --
now runs one background task per store: one bounded purge call per
fixed tick, exceptions logged and retried next tick, cancelled in
close() before the stores shut down. One call per tick keeps the
background work bounded by construction (a backlog drains over
successive ticks), and no purger coordination is needed at any
instance count because the store's claiming already makes racing
purgers safe. The store itself still never schedules reclamation;
this loop is the caller-side scheduler the ABC contract calls for.

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>

* fix: purge loop reads the backlog signal

purge_deleted_partitions() returning True is the API's statement that
more work remains; discarding it drained a backlog at one bounded
call per tick (~167 rows/s at the defaults). The loop now runs
bounded calls back-to-back while the store reports more and sleeps
one tick only when it reports done or a call fails -- full-rate
recovery, still bounded per call, still one idle call per tick.
Pinned by a test that drains a three-call backlog under a
deliberately huge tick interval.

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>

* docs: empty-input calls MAY skip the handle check

The contract permits the shortcut rather than mandating it; an
implementation that checks anyway still conforms.

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>

* test: pin the config-mismatch guard directly

The guard, its error type, and the ABC declaration predate this
branch, but no test staged a mismatch -- only the lifecycle-churn
test tolerated it as a domain outcome. Plaintext is the only concrete
codec config, so the test stands in a subclass for a future variant
(pydantic instances survive validation unrevalidated and compare
unequal by class). Verified to fail with the guard ablated.

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>

* purge: bound the integrity-escape link delete; warn on it and on collisions

The retire-path guard delete was the one unbounded statement in the
purge, unbounded precisely when it was not a no-op. It is now batched
under the same per-call budget as the segment rows: a full batch
leaves the queue entry for the next call (the existing call-until-
False contract absorbs it, callers unchanged), the normal case still
costs one zero-row statement, and reclaiming rows there logs a
warning naming the incarnation, since it means referential integrity
failed somewhere. Pinned by a test that stages orphan link rows
through a second SQLite engine without the foreign-key pragma and
drains them in warned batches; verified to fail against the unbounded
form.

The module's logger also gains the only other events worth an
operator's attention: a minted incarnation colliding (with garbage or
in the registry) is warning-logged at the detection site -- a genuine
collision is astronomically unlikely, so the log marks either broken
randomness or a misclassified persistent database error, visible even
when retries eventually succeed. Everything else either raises to the
caller or is normal operation, and stays unlogged.

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>

* docs: shared purge budget is measured, not a guessed ratio

Measured with the purge's batched-delete shape on 100k rows each
(3 interleaved rounds): a segment row deletes at ~3.3 us and a
derivative-link row at ~1.0 us, so link rows are about 3x cheaper --
they are narrower, carry fewer indexes, and fire no cascade. That is
why integrity-escaped links draw count-for-count on the segment
budget instead of getting their own limit: one budget calibrated on
the most expensive row type upper-bounds the call, whereas a separate
link limit would be safe only under an assumed cost ratio.

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>

* docs: the row-cost direction is the shared budget's precondition

The shared purge budget stays conservative only while a link row
deletes cheaper than a segment row; widening the link table or adding
indexes to it revisits the choice.

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>

* review: pace the purger, state the drain's real guarantee, close cleanly

Round-3 findings 1, 2, 5 and 6 -- the first three introduced by the
background purger itself.

- The purge loop now pauses briefly after every productive call
  instead of running delete transactions back-to-back: the pause
  yields the database (and SQLite's single write lock) to request
  serving, while a backlog still drains at one bounded call per pause
  and an idle store costs one call per tick. This also removes the
  in-process busy-timeout window between the background task and the
  inline drain on SQLite.
- The inline drain's comment claimed "the server schedules no other
  purger", which the purger commit falsified, and "reclaimed before
  returning", which SKIP LOCKED claiming never strictly guaranteed
  under any concurrent purger. Comment, design doc, and PR body now
  state the actual promise: rows are reclaimed promptly -- normally
  before the drain returns, and otherwise within the bounded call of
  whichever purger claimed the entry, moments later.
- close() clears the purge-task list and the store registry, so a
  second close is a no-op and a post-close get_segment_store can no
  longer hand back a shut-down store that silently never purges.
- The design doc no longer implies deployments can already tune the
  purge bounds through server configuration: the server constructs
  its stores with the defaults, and config plumbing is future work.

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>

* review: one datetime-normalization rule per filter path

Round-3 findings 3 and 11.

- The properties_json In branch bound raw values while its Comparison
  sibling normalized through _cast_properties_json_value; a datetime
  In list would bind datetime objects against the stored ISO-string
  form (an InterfaceError on Python 3.14's sqlite3, a never-matching
  comparison on PostgreSQL). Both leaf shapes now cast and normalize
  through the one function, which also aligns the float and bool
  casts the old branch fell through to as_string/as_integer. Same
  defensive-reachability status as the column-leaf In normalization
  kept deliberately: unreachable by In's declared value types,
  reachable at runtime.
- The episode store's start_time/end_time bounds re-implemented the
  UTC normalization inline; sql_filter_util's normalize_column_value
  is now public and both bounds use it, so the storage convention has
  one definition across compiled filters and dedicated bounds.

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>

* review: StaticPool guard raises; empty add_segments short-circuits

Round-3 findings 9 and 10.

- The params validator's StaticPool guard was a bare assert, stripped
  under python -O -- and the design depends on multiple connections
  (the registry fence, deletion waiting out writers, SKIP LOCKED
  claiming all degrade on one shared connection). It now raises
  ValueError like the ephemeral-SQLite check beside it; pinned by a
  test, and the check stays a ValueError because pydantic converts
  only ValueError/AssertionError into a ValidationError.
- add_segments returns early on empty input, matching delete_segments
  and the ABC's empty-input permission; previously it opened a
  transaction and, on SQLite, took the write lock to insert nothing.
  The stale-handle test pins the shortcut.

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>

* filter: datetime values normalize to UTC at node construction

The filter language now owns datetime semantics: a value denotes an
instant, and a naive value means UTC. Comparison.__post_init__
normalizes datetime values to UTC-aware instants (In gets the same
defensively -- its declared types exclude datetimes, but runtime lists
are unchecked), so every consumer -- parsed trees and programmatically
built ones, SQL compilers and vector stores alike -- receives
normalized instants by construction, and compilers only choose a
representation.

This is where the rule the recent fixes kept restating per leaf
actually belongs: the same aware-to-UTC-or-naive-means-UTC conversion
appeared in the SQL column leaf, the properties_json leaf, the episode
bounds, and twice in the Milvus store, and two of the drifted copies
were bugs fixed this round. With the invariant at the node, the SQL
column leaf's re-normalization became redundant and is reverted (it
binds tree values as-is); the properties_json leaf keeps its routing
because datetime-to-ISO-string is representation, not normalization;
the episode start/end bounds keep the shared helper because they are
raw API values outside any tree; other backends' now-idempotent
defenses are left for separate cleanup.

Pinned by tests that a programmatically built Comparison and a parsed
date() literal with a non-UTC offset both carry the UTC instant.

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>

* filter: drop normalize_column_value; contract stated on the protocol

With datetime normalization at node construction, the compiler-side
helper had no filter role left, and its one remaining consumer -- the
episode store's start/end bounds, which arrive outside any filter
tree -- now spells the convention inline as the two explicit steps,
ensure_tz_aware(...).astimezone(UTC). A composed to_utc() helper was
considered and rejected: the name does not pin the naive-means-UTC
tagging decision (an alternative design under the same name could
reject naive datetimes entirely), so the explicit steps are clearer
at each site.

The FilterExpr protocol docstring now states the construction-time
contract where the next value-carrying node's author will read it:
such a node normalizes datetime values to UTC-aware instants, and
compilers bind instants without re-normalizing.

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>

* docs: the two-step datetime spelling is deliberate

Record the datetime convention in the design doc so a future cleanup
does not consolidate the repeated ensure_tz_aware(...).astimezone(UTC)
sequences back into the composed helper b636d62a deliberately removed:
a name that pins only the conversion, not the naive-means-UTC tagging,
hides a real design decision, so the repetition is load-bearing.

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>

* docs: second ground for rejecting the composed datetime helper

A shared helper earns its place only when the name honestly pins the
unit AND the composition structurally prevents half-applied
normalization. The second condition fails here regardless of naming:
read paths legitimately need the tagging step alone (segment reads
reapply the stored original offset; cluster and episode reads only tag
naive database values), so ensure_tz_aware stays independently
available and the helper could not have removed the partial-use error
class.

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>

* fix: SQLite foreign keys enforced from engine creation

Round-4 findings 1, 4 and 5. The store registered its foreign_keys
pragma as a per-store connect listener, which has two structural
faults: connections the caller's shared engine pooled before the store
existed never receive the pragma, so cascade deletes silently leave
orphaned link rows for LIV…

…ists (#1578)

_create_native_collection wrapped create_collection and every
create_payload_index in one try and swallowed "already exists" for the whole
block. A collection that already existed therefore raised on the first call,
took the already-exists path, and was left with no payload indexes at all -
despite the docstring promising both were created idempotently.

Two creators are easy to arrive at. The guarding lock is keyed on the
AsyncQdrantClient object, so it serialises callers inside one process and
nothing across them; with MEMMACHINE_WORKERS above 1 each worker has its own
client and its own lock. A crash between the two calls leaves the same state.

The collection and the indexes now sit under separate guards, and each index
is created individually and tolerant of already-exists.

Verified against Qdrant 1.19 in testcontainers. Before the change,
test_indexes_are_created_when_the_collection_already_exists fails with an
empty payload_schema - not a missing index, none of the twelve. After it, both
new tests pass, along with the rest of the vector store suite: 293 unit and
142 integration.

Scope, checked rather than assumed: a filtered query on an unindexed
collection returns only the matching tenant's points, so what a missing
sys-partition_key index costs is the multitenant storage layout and query
speed, not isolation. Note also that only the already-exists path reproduces;
two clients creating simultaneously both succeed, which the second test pins.

The tests need a real server and are marked integration - local-mode Qdrant
ignores payload indexes, so the defect is invisible there and CI, which runs
with `-m "not integration"`, will not exercise them.


Claude-Session: https://claude.ai/code/session_01Nr9kacmpFVTTfkZRw6esxP

Co-authored-by: Claude Opus 5 <noreply@anthropic.com>

…the backend (#1580)

A deployment that runs the server on its own, with a hand-written config and no
chart, gets whatever the code defaults are. Those defaults assumed a full
installation: short-term memory on, semantic memory on, and long-term memory on
Neo4j. A lab running MemMachine for retrieval alone had to turn two things off
and switch a third before it matched what it wanted.

Short-term memory is now off unless asked for. Two changes were needed, not one:
the field default, and the merge, which read True whenever a short_term_memory
section existed at all - so the field default never applied to any real config.
Configuring the block and enabling it are now separate statements.

Semantic memory is off by default. Note this only decides the fully-configured
case: _auto_disable_when_incomplete already forced it off whenever llm_model,
embedding_model or the storage fields were missing, so the old default of True
only took effect for a deployment that had wired all of them up.

The long-term backend is now inferred from which fields a config fills in
rather than defaulting. The two backends share no field names, so naming one is
an unambiguous statement: a config pointing at a `vector_store` cannot mean the
declarative backend, which has no such field.

The discriminator itself is untouched, deliberately. Flipping `None ->
declarative` would repoint every pre-discriminator config at Qdrant, against
data sitting in Neo4j - and those configs name a `vector_graph_store`, so
inference resolves them exactly as before. test_old_param_data_without_backend_
loads_as_declarative still passes unchanged. Defaulting to event when nothing is
named would not work either: such a config has no store ids, so the event
backend cannot start.

Two shipped configs relied on the old defaults and now state what they want, so
that this change does not quietly alter them: locomo_config.yaml ran with
short-term memory on, and deployments/helm configures both memories and would
otherwise have shipped an llm_model and message_capacity that were never used.

Verified: 1964 tests pass. Seven new tests cover the changed behaviour, and each
was checked against the old code to confirm it fails there - which caught one
that did not. Asserting the semantic default on a bare instance passes whatever
the default is, because the auto-disable fires first; the test now supplies
every required field so that nothing but the default decides.


Claude-Session: https://claude.ai/code/session_01Nr9kacmpFVTTfkZRw6esxP

Co-authored-by: Claude Opus 5 <noreply@anthropic.com>