cogno-sync 0.1.1__py3-none-any.whl

cogno_sync/__init__.py

@@ -0,0 +1,62 @@
+"""
+cogno-sync — distributed coordination primitives for the Cogno stack.
+
+Four small ports, each with a zero-infra in-process default and a Redis adapter that shares the
+state across workers: a lock, a token-bucket rate limiter, a claim-once idempotency store, and a
+rolling daily counter. Every adapter is async, every one **fails open**, and every one takes the
+same constructor — a ready client, or a URL to build one from.
+
+Redis is an optional extra (``cogno-sync[redis]``), imported only when a client is built; the
+ASGI middleware lives in :mod:`cogno_sync.asgi` behind ``cogno-sync[asgi]``. Importing this
+package pulls in neither.
+"""
+
+from importlib.metadata import PackageNotFoundError
+from importlib.metadata import version as _dist_version
+
+try:
+    __version__ = _dist_version("cogno-sync")
+except PackageNotFoundError:  # source tree without an installed dist (e.g. vendored checkout)
+    __version__ = "0.0.0"
+
+
+from cogno_sync._redis import client_from_url
+from cogno_sync.daily_cap import DailyCap, InMemoryDailyCap, RedisDailyCap, utc_day
+from cogno_sync.idempotency import (
+    DEFAULT_CLAIM_TTL,
+    IdempotencyStore,
+    InMemoryIdempotencyStore,
+    RedisIdempotencyStore,
+)
+from cogno_sync.locks import (
+    DEFAULT_LOCK_TTL,
+    InProcessLockProvider,
+    LockProvider,
+    RedisLockProvider,
+)
+from cogno_sync.ratelimit import (
+    InMemoryTokenBucketLimiter,
+    RateLimitDecision,
+    RateLimiter,
+    RedisTokenBucketLimiter,
+)
+
+__all__ = [
+    "client_from_url",
+    "LockProvider",
+    "InProcessLockProvider",
+    "RedisLockProvider",
+    "DEFAULT_LOCK_TTL",
+    "RateLimiter",
+    "RateLimitDecision",
+    "InMemoryTokenBucketLimiter",
+    "RedisTokenBucketLimiter",
+    "IdempotencyStore",
+    "InMemoryIdempotencyStore",
+    "RedisIdempotencyStore",
+    "DEFAULT_CLAIM_TTL",
+    "DailyCap",
+    "InMemoryDailyCap",
+    "RedisDailyCap",
+    "utc_day",
+]

cogno_sync/_redis.py

@@ -0,0 +1,66 @@
+"""One lazy Redis client factory, and the shared shape every ``Redis*`` adapter here follows.
+
+``redis`` is an **optional extra** (``cogno-sync[redis]``): it is imported the first time a
+client is actually built, so ``import cogno_sync`` never needs it.
+
+Every adapter in this package takes the same two arguments — a ready client, or a URL to build
+one from — and holds the same three fields. That uniformity is the point: the library this code
+came from had three different constructor conventions across five adapters, and each one grew
+its own copy of ``Redis.from_url``, so a process ended up holding one connection pool per
+adapter. **Build one client and inject it everywhere**; the URL form stays for a one-adapter
+script and for tests.
+"""
+
+from __future__ import annotations
+
+from typing import Any
+
+
+def client_from_url(url: str, *, decode_responses: bool = False) -> Any:
+    """Build a ``redis.asyncio.Redis`` from ``url``.
+
+    Nothing connects here — the client opens its pool on the first command — so this is safe
+    to call at start-up, before the server is up.
+
+    ``decode_responses=True`` makes the client return ``str`` instead of ``bytes``. Every
+    adapter in this package works either way: none of them compares a returned value against a
+    literal (the lock's owner check runs inside Redis, the limiter reads integers), so a caller
+    that shares one client with its own code is free to pick whichever that code needs.
+    """
+    if not url:
+        raise ValueError("client_from_url needs a non-empty redis url")
+    from redis.asyncio import Redis           # lazy: the [redis] extra
+    return Redis.from_url(url, decode_responses=decode_responses)
+
+
+class _RedisAdapter:
+    """Constructor + lazy-client behaviour shared by the ``Redis*`` adapters.
+
+    Subclasses declare the full public signature themselves — ``(client=None, *, url="",
+    key_prefix="cogno:…:")`` — so the default keyspace is readable at the class that owns it.
+    """
+
+    def __init__(self, client: Any = None, *, url: str = "", key_prefix: str = "",
+                 decode_responses: bool = False) -> None:
+        if isinstance(client, str):
+            # A URL handed in the positional (the CLIENT slot) would only surface on the first
+            # command — and every adapter here fails open, so it would be swallowed forever and
+            # the coordination would be silently off. Accept it as the url instead.
+            client, url = None, client
+        if client is None and not url:
+            raise ValueError(f"{type(self).__name__} needs a redis client or a url")
+        self._client: Any = client
+        self._url: str = url
+        self._decode = decode_responses
+        self._prefix = key_prefix
+
+    def _conn(self) -> Any:
+        if self._client is None:
+            self._client = client_from_url(self._url, decode_responses=self._decode)
+        return self._client
+
+    def _key(self, key: str) -> str:
+        return f"{self._prefix}{key}"
+
+
+__all__ = ["client_from_url"]

cogno_sync/asgi.py

@@ -0,0 +1,118 @@
+"""ASGI edge helpers — the rate-limit middleware and a bounded body read.
+
+This is the only module here that needs a web framework, so it lives apart and behind the
+``[asgi]`` extra (Starlette): importing :mod:`cogno_sync` itself never pulls one in.
+
+* :class:`RateLimitMiddleware` applies a :class:`~cogno_sync.ratelimit.RateLimiter` to the
+  guarded path prefixes and answers ``429`` with a ``Retry-After`` header.
+* :func:`read_capped_body` reads a request body incrementally and gives up past a limit,
+  instead of buffering an unbounded stream into the worker's heap.
+"""
+
+from __future__ import annotations
+
+import logging
+from typing import Any, Callable, Optional, Sequence
+
+from starlette.middleware.base import BaseHTTPMiddleware
+from starlette.requests import Request
+from starlette.responses import JSONResponse
+
+from cogno_sync.ratelimit import RateLimiter
+
+log = logging.getLogger(__name__)
+
+# Every path. Narrow it per deployment: which routes are worth guarding is a property of the
+# application, never of this library.
+ALL_PATHS: tuple[str, ...] = ("/",)
+
+# A generous ceiling for a machine-to-machine POST (batched webhook events and the like).
+DEFAULT_MAX_BODY_BYTES = 2 * 1024 * 1024
+
+
+def client_ip_key(request: Request, *, trusted_proxy_hops: int = 1) -> str:
+    """The caller's IP, resolved against the number of TRUSTED reverse proxies in front.
+
+    The real client is the entry the OUTERMOST trusted proxy recorded — the N-th from the RIGHT
+    of ``X-Forwarded-For``, since each proxy appends the address it received the connection
+    from. Taking the leftmost hop is spoofable: a caller sets ``X-Forwarded-For: <anything>``
+    and lands at position 0 with a fresh budget on every request.
+
+    ``trusted_proxy_hops=0`` ignores the header entirely (the app is directly exposed). A chain
+    shorter than the hop count — a direct hit, or a misconfigured deployment — falls back to the
+    peer address, never to a client-controlled value. ``anon`` when there is no peer either: a
+    limiter must not crash on a missing client.
+    """
+    peer = request.client.host if request.client else "anon"
+    if trusted_proxy_hops <= 0:
+        return peer
+    xff = request.headers.get("x-forwarded-for", "")
+    hops = [h.strip() for h in xff.split(",") if h.strip()] if xff else []
+    if len(hops) >= trusted_proxy_hops:
+        return hops[-trusted_proxy_hops]
+    return peer
+
+
+class RateLimitMiddleware(BaseHTTPMiddleware):
+    """429s a request whose key is over budget — but only on the guarded ``prefixes``;
+    everything else (a health probe, static assets) passes straight through.
+
+    ``key_fn`` decides what is being limited; the default is the caller's IP resolved with
+    ``trusted_proxy_hops``. That number is a **constructor argument on purpose**: read from the
+    environment at import time it would be fixed for the whole process, invisible to a test, and
+    impossible to differ between two apps in one runtime.
+    """
+
+    def __init__(self, app: Any, *, limiter: RateLimiter,
+                 prefixes: Sequence[str] = ALL_PATHS,
+                 key_fn: Optional[Callable[[Request], str]] = None,
+                 trusted_proxy_hops: int = 1) -> None:
+        super().__init__(app)
+        self._limiter = limiter
+        self._prefixes = tuple(prefixes)
+        self._hops = trusted_proxy_hops
+        self._key_fn = key_fn or (lambda r: client_ip_key(r, trusted_proxy_hops=self._hops))
+
+    async def dispatch(self, request: Request, call_next):   # noqa: ANN001,ANN201 — Starlette shape
+        if request.url.path.startswith(self._prefixes):
+            key = self._key_fn(request)
+            decision = await self._limiter.check(key)
+            if not decision.allowed:
+                log.warning("event=rate_limited key=%s path=%s", key, request.url.path)
+                return JSONResponse(
+                    {"detail": "rate limit exceeded"}, status_code=429,
+                    headers={"Retry-After": str(max(1, int(decision.retry_after) + 1))})
+        return await call_next(request)
+
+
+class BodyTooLarge(Exception):
+    """Raised when the request body exceeds the cap. Callers answer ``413``."""
+
+
+async def read_capped_body(request: Request, limit: int = DEFAULT_MAX_BODY_BYTES) -> bytes:
+    """Return the body, or raise :class:`BodyTooLarge` without ever buffering past ``limit``.
+
+    Unauthenticated routes are the reason this exists: a signature can only be verified AFTER
+    the payload is read, and such routes are usually exempt from an IP limiter too (a provider's
+    shared egress addresses would collapse onto one key). ``await request.body()`` buffers the
+    whole stream into the worker's heap, so a handful of concurrent multi-GB POSTs is enough to
+    OOM-kill the process and take every caller down with it.
+
+    ``Content-Length`` is advisory — simply absent on a ``Transfer-Encoding: chunked`` request —
+    so it is a fast pre-check here, never the guard: the running total is.
+    """
+    declared: Optional[str] = request.headers.get("content-length")
+    if declared and declared.isdigit() and int(declared) > limit:
+        raise BodyTooLarge()
+    parts: list[bytes] = []
+    total = 0
+    async for chunk in request.stream():
+        total += len(chunk)
+        if total > limit:
+            raise BodyTooLarge()
+        parts.append(chunk)
+    return b"".join(parts)
+
+
+__all__ = ["RateLimitMiddleware", "client_ip_key", "read_capped_body", "BodyTooLarge",
+           "DEFAULT_MAX_BODY_BYTES", "ALL_PATHS"]

cogno_sync/daily_cap.py

@@ -0,0 +1,81 @@
+"""Per-key DAILY counters — a rolling cost ceiling for anonymous surfaces.
+
+A per-second limiter bounds REQUESTS PER SECOND, not spend: at 10 rps one caller can still
+drive an unbounded amount of expensive work over a day. A per-session counter does not close it
+either whenever the session id comes from the client — omitting it starts a fresh session on
+every request. A rolling daily count per caller is the missing ceiling: cheap, anonymous-
+friendly (a real user never reaches it) and checkable BEFORE any money is spent.
+
+Both implementations **fail open** — a counter outage returns ``0``, which every caller must
+read as "unknown, let it through". A quota exists to bound cost, not to be a correctness
+barrier.
+"""
+
+from __future__ import annotations
+
+import logging
+from datetime import datetime, timezone
+from typing import Any, Protocol, runtime_checkable
+
+from cogno_sync._redis import _RedisAdapter
+
+log = logging.getLogger(__name__)
+
+
+def utc_day() -> str:
+    """The current UTC day stamp — the natural reset boundary for a rolling daily quota."""
+    return datetime.now(timezone.utc).strftime("%Y-%m-%d")
+
+
+@runtime_checkable
+class DailyCap(Protocol):
+    """Count one hit for ``key`` today; returns the count INCLUDING this hit (0 = unknown)."""
+
+    async def hit(self, key: str) -> int: ...
+
+
+class InMemoryDailyCap:
+    """Process-local counter — dev / single worker. A multi-worker deployment injects the Redis
+    one, or each worker hands out its own full allowance."""
+
+    def __init__(self) -> None:
+        self._day = utc_day()
+        self._counts: dict[str, int] = {}
+
+    async def hit(self, key: str) -> int:
+        today = utc_day()
+        if today != self._day:            # cheap rollover: drop yesterday wholesale
+            self._day, self._counts = today, {}
+        self._counts[key] = self._counts.get(key, 0) + 1
+        return self._counts[key]
+
+
+class RedisDailyCap(_RedisAdapter):
+    """Shared counter across workers and restarts: ``INCR`` + a TTL that expires the key itself.
+
+    The **caller** puts the day in the key (see :func:`utc_day`) — the window then rolls over on
+    its own and yesterday's keys expire, so nothing has to be swept. Only the first hit of a
+    window sets the expiry: re-setting it on every hit would let a caller who keeps hitting the
+    key hold it alive forever, and the count would never reset.
+    """
+
+    def __init__(self, client: Any = None, *, url: str = "", key_prefix: str = "cogno:cap:",
+                 ttl_seconds: int = 26 * 3600) -> None:
+        super().__init__(client, url=url, key_prefix=key_prefix)
+        self._ttl = ttl_seconds           # > 24h so a day boundary never truncates a window
+
+    async def hit(self, key: str) -> int:
+        try:
+            conn = self._conn()
+            full = self._key(key)
+            count = int(await conn.incr(full))
+            if count == 1:                # first hit of the window owns the expiry
+                await conn.expire(full, self._ttl)
+            return count
+        except Exception as exc:  # noqa: BLE001 — a cache blip must not wall a legitimate user
+            log.warning("event=daily_cap_error key=%s error=%s proceeding_fail_open",
+                        key, type(exc).__name__)
+            return 0
+
+
+__all__ = ["DailyCap", "InMemoryDailyCap", "RedisDailyCap", "utc_day"]

cogno_sync/idempotency.py

@@ -0,0 +1,113 @@
+"""Claim-once idempotency — handle a delivery ONCE, however many times it arrives.
+
+A lock (:mod:`cogno_sync.locks`) **serializes** work; it does not **deduplicate** it. Once the
+first run finishes and releases, a re-delivery takes the lock and does the whole thing again.
+At-least-once transports (webhooks, queues, mobile clients on a flaky link) re-deliver by
+design, so the receiver is the only place that can make the effect once.
+
+Two decisions shape this port, and both are worth stating because the opposite choice looks
+equally reasonable until it ships:
+
+* **Claim BEFORE the work, release on failure.** Claiming afterwards leaves the concurrent
+  window open; claiming before with no way to release turns one transient crash into a message
+  lost forever. As written: a concurrent re-delivery is refused, a re-delivery after success is
+  refused, and a re-delivery after a crash — before anything committed — is let through, because
+  the caller called :meth:`forget`.
+* **Fail-OPEN.** A store error lets the delivery through with a warning, the same direction
+  :class:`~cogno_sync.locks.RedisLockProvider` takes. A store that goes down must not take the
+  service with it.
+
+The KEY is the caller's business: it must carry whatever makes one delivery distinct in that
+system (the sender, the source, the provider's own id), because a provider's ids are unique per
+provider, not globally. This package never composes it.
+"""
+
+from __future__ import annotations
+
+import logging
+import time
+from typing import Any, Callable, Protocol, runtime_checkable
+
+from cogno_sync._redis import _RedisAdapter
+
+log = logging.getLogger(__name__)
+
+# Senders retry within minutes; a day is generous and costs nothing (Redis expires it).
+DEFAULT_CLAIM_TTL = 24 * 60 * 60
+
+
+@runtime_checkable
+class IdempotencyStore(Protocol):
+    """Port: has this delivery been claimed, and claim it if not — in ONE atomic step."""
+
+    async def seen_before(self, key: str, *, ttl: int = DEFAULT_CLAIM_TTL) -> bool:
+        """True when ``key`` was already claimed. False claims it for this caller.
+
+        Must be atomic (test-and-set). A ``get`` followed by a ``set`` would let two concurrent
+        re-deliveries both read "unseen" and both proceed — which is the exact race this
+        exists to close.
+        """
+        ...
+
+    async def forget(self, key: str) -> None:
+        """Release a claim, so a re-delivery may retry work that never happened."""
+        ...
+
+
+class InMemoryIdempotencyStore:
+    """Process-local claims — the zero-infra default. Correct for ONE process; swap for
+    :class:`RedisIdempotencyStore` when more than one of them receives the same stream."""
+
+    def __init__(self, *, now: Callable[[], float] = time.monotonic) -> None:
+        self._seen: dict[str, float] = {}          # key → expiry (monotonic seconds)
+        self._now = now
+
+    def _sweep(self) -> None:
+        cutoff = self._now()
+        for key in [k for k, exp in self._seen.items() if exp <= cutoff]:
+            self._seen.pop(key, None)
+
+    async def seen_before(self, key: str, *, ttl: int = DEFAULT_CLAIM_TTL) -> bool:
+        self._sweep()
+        # Single-threaded asyncio: no await between the read and the write, so this IS atomic
+        # with respect to other tasks. Adding an await here would reopen the race.
+        if key in self._seen:
+            return True
+        self._seen[key] = self._now() + ttl
+        return False
+
+    async def forget(self, key: str) -> None:
+        self._seen.pop(key, None)
+
+
+class RedisIdempotencyStore(_RedisAdapter):
+    """Cross-process claims over Redis (``SET NX EX`` — atomic by construction).
+
+    Pass a ready ``redis.asyncio.Redis`` client, or a URL to build one lazily. Fail-open: any
+    Redis error logs a WARNING and reports "not seen", so the delivery is handled rather than
+    dropped.
+    """
+
+    def __init__(self, client: Any = None, *, url: str = "",
+                 key_prefix: str = "cogno:claim:") -> None:
+        super().__init__(client, url=url, key_prefix=key_prefix)
+
+    async def seen_before(self, key: str, *, ttl: int = DEFAULT_CLAIM_TTL) -> bool:
+        try:
+            # NX returns None when the key already exists → already claimed by someone.
+            claimed = await self._conn().set(self._key(key), "1", nx=True, ex=ttl)
+            return not claimed
+        except Exception as exc:                 # noqa: BLE001 — fail-open, see the module doc
+            log.warning("event=claim_store_unavailable key=%s error=%s — proceeding anyway",
+                        key, type(exc).__name__)
+            return False
+
+    async def forget(self, key: str) -> None:
+        try:
+            await self._conn().delete(self._key(key))
+        except Exception as exc:                 # noqa: BLE001 — best-effort release
+            log.warning("event=claim_release_failed key=%s error=%s", key, type(exc).__name__)
+
+
+__all__ = ["IdempotencyStore", "InMemoryIdempotencyStore", "RedisIdempotencyStore",
+           "DEFAULT_CLAIM_TTL"]

cogno_sync/locks.py

@@ -0,0 +1,176 @@
+"""Mutual exclusion on a key — serialize concurrent work across workers.
+
+A unit of work that reads some state, changes it and writes it back races with itself whenever
+two copies run at once: a client retrying, a user acting twice in a second, two HTTP requests
+arriving at different processes. One process is covered by an :class:`asyncio.Lock`; **several
+processes are not** — they share no memory.
+
+* :class:`LockProvider` — the port: ``lock(key)`` → an async context manager that holds the
+  lock for the duration of the body.
+* :class:`InProcessLockProvider` — the zero-infra default: a per-key ``asyncio.Lock``. Correct
+  inside ONE process, and only there.
+* :class:`RedisLockProvider` — the cross-process adapter: ``SET NX EX`` with a per-acquisition
+  token plus an owner-only Lua compare-and-delete on release, so a worker can never delete a
+  lock someone else took after its own TTL expired.
+
+**The TTL is RENEWED while the body runs** (0.1.1). A heartbeat, every ``ttl / 3``, extends the
+key's TTL with an owner-only Lua compare-and-expire — so a body slower than the TTL keeps its lock
+and the next acquirer WAITS, instead of taking an expired key and running beside it (measured on
+the reference host: a turn's p99 of 126 s against the 120 s default). The heartbeat stops when the
+body ends or is cancelled, stops for good the moment the key holds someone else's token (it never
+renews a lock that is not its own), and treats a Redis error as one missed beat: it logs and tries
+again on the next one, and the body never waits on it.
+
+**Fail-open.** If Redis is unreachable, or the lock cannot be taken within ``max_wait``, the
+body runs anyway with a WARNING. These primitives guard against a rare double-run; refusing to
+do the work at all is the worse failure.
+"""
+
+from __future__ import annotations
+
+import asyncio
+import contextlib
+import logging
+import uuid
+from typing import (Any, AsyncIterator, Awaitable, Callable, Optional, Protocol,
+                    runtime_checkable)
+
+from cogno_sync._redis import _RedisAdapter
+
+log = logging.getLogger(__name__)
+
+DEFAULT_LOCK_TTL = 120          # seconds — must exceed the slowest critical section
+_RETRY_INTERVAL = 0.5           # seconds between acquisition attempts
+_MAX_WAIT = 60.0                # seconds — then fail-open (proceed without the lock)
+_RENEW_FRACTION = 3             # renew every ttl / 3: two beats can be missed before it expires
+_MIN_RENEW_INTERVAL = 0.05      # seconds — a floor, for a tiny ttl
+_HEARTBEAT_STOP_WAIT = 1.0      # seconds — how long the exit waits for the cancelled heartbeat
+
+# Owner-only release: delete the key only if it still holds OUR token (never delete a lock a
+# later acquirer took after our TTL expired).
+_LUA_RELEASE = """
+if redis.call("GET", KEYS[1]) == ARGV[1] then
+    return redis.call("DEL", KEYS[1])
+else
+    return 0
+end
+"""
+
+# Owner-only renewal: extend the TTL only if the key still holds OUR token — never a lock a later
+# acquirer took after ours expired.
+_LUA_RENEW = """
+if redis.call("GET", KEYS[1]) == ARGV[1] then
+    return redis.call("EXPIRE", KEYS[1], ARGV[2])
+else
+    return 0
+end
+"""
+
+
+@runtime_checkable
+class LockProvider(Protocol):
+    """Yields an async context manager that holds the lock for ``key`` while the body runs."""
+
+    def lock(self, key: str, *,
+             ttl: int = DEFAULT_LOCK_TTL) -> "contextlib.AbstractAsyncContextManager[None]": ...
+
+
+class InProcessLockProvider:
+    """A per-key ``asyncio.Lock`` — the zero-infra default. Serializes within ONE process only;
+    swap for :class:`RedisLockProvider` to serialize across processes."""
+
+    def __init__(self) -> None:
+        self._locks: dict[str, asyncio.Lock] = {}
+
+    @contextlib.asynccontextmanager
+    async def lock(self, key: str, *, ttl: int = DEFAULT_LOCK_TTL) -> AsyncIterator[None]:
+        existing = self._locks.get(key)
+        if existing is None:
+            existing = self._locks[key] = asyncio.Lock()
+        async with existing:
+            yield
+
+
+class RedisLockProvider(_RedisAdapter):
+    """Distributed lock over Redis (``SET NX EX`` + owner-only Lua release + an owner-only
+    renewal of the TTL while the body runs).
+
+    Pass a ready ``redis.asyncio.Redis`` client, or a URL to build one lazily. Fail-open by
+    design: any Redis error or an acquisition timeout logs a WARNING and runs the body.
+
+    ``renew=False`` turns the heartbeat off (the 0.1.0 behaviour). ``sleep`` is the clock the
+    acquisition retries and the heartbeat wait on — injectable so a test drives a fake one. When
+    it is not given, ``asyncio.sleep`` is looked up at CALL time, as 0.1.0 did, so a caller that
+    patches this module's ``asyncio`` keeps driving the waits.
+    """
+
+    def __init__(self, client: Any = None, *, url: str = "",
+                 key_prefix: str = "cogno:lock:", retry_interval: float = _RETRY_INTERVAL,
+                 max_wait: float = _MAX_WAIT, renew: bool = True,
+                 sleep: "Optional[Callable[[float], Awaitable[Any]]]" = None) -> None:
+        super().__init__(client, url=url, key_prefix=key_prefix)
+        self._retry = retry_interval
+        self._max_wait = max_wait
+        self._renew = renew
+        self._sleep = sleep
+        self._owner = uuid.uuid4().hex          # stable per process
+
+    @contextlib.asynccontextmanager
+    async def lock(self, key: str, *, ttl: int = DEFAULT_LOCK_TTL) -> AsyncIterator[None]:
+        lock_key = self._key(key)
+        token = f"{self._owner}:{uuid.uuid4().hex}"
+        acquired = False
+        try:
+            r = self._conn()
+            elapsed = 0.0
+            while elapsed < self._max_wait:
+                if await r.set(lock_key, token, nx=True, ex=ttl):
+                    acquired = True
+                    break
+                await self._wait(self._retry)
+                elapsed += self._retry
+            if not acquired:
+                log.warning("event=lock_timeout key=%s waited=%.1fs proceeding_fail_open",
+                            key, elapsed)
+        except Exception as exc:  # noqa: BLE001 — fail-open: never drop work on a lock error
+            log.warning("event=lock_error key=%s error=%s proceeding_fail_open", key, exc)
+            yield
+            return
+        beat = (asyncio.ensure_future(self._heartbeat(r, lock_key, token, ttl, key))
+                if acquired and self._renew else None)
+        try:
+            yield
+        finally:
+            if beat is not None:
+                # Stopped BEFORE the release, so no beat can renew a key we are about to free.
+                # `asyncio.wait`, never `wait_for`: on 3.10 `wait_for` hangs on a task that
+                # swallows its cancel; this returns after the bound whatever the task does.
+                beat.cancel()
+                await asyncio.wait({beat}, timeout=_HEARTBEAT_STOP_WAIT)
+            if acquired:
+                try:
+                    await r.eval(_LUA_RELEASE, 1, lock_key, token)
+                except Exception as exc:  # noqa: BLE001
+                    log.warning("event=lock_release_failed key=%s error=%s", key, exc)
+
+    async def _wait(self, seconds: float) -> None:
+        await (self._sleep or asyncio.sleep)(seconds)
+
+    async def _heartbeat(self, r: Any, lock_key: str, token: str, ttl: int, key: str) -> None:
+        """Renew ``lock_key``'s TTL every ``ttl / 3`` while it holds ``token``. Ends when the key
+        holds another token (``event=lock_lost``) or when cancelled; a Redis error is one missed
+        beat (``event=lock_renew_failed``), never an exception the body sees."""
+        interval = max(float(ttl) / _RENEW_FRACTION, _MIN_RENEW_INTERVAL)
+        while True:
+            await self._wait(interval)
+            try:
+                kept = await r.eval(_LUA_RENEW, 1, lock_key, token, int(ttl))
+            except Exception as exc:  # noqa: BLE001 — a missed beat, never the body's problem
+                log.warning("event=lock_renew_failed key=%s error=%s", key, type(exc).__name__)
+                continue
+            if not kept:
+                log.warning("event=lock_lost key=%s", key)
+                return
+
+
+__all__ = ["LockProvider", "InProcessLockProvider", "RedisLockProvider", "DEFAULT_LOCK_TTL"]

cogno_sync/ratelimit.py

@@ -0,0 +1,145 @@
+"""Rate limiting — a token-bucket port with an in-process default and a Redis adapter.
+
+A service that is reachable without an infrastructure edge in front of it carries its own
+limit. This module is the **pure** limiter (no web framework, so importing it never needs the
+``[asgi]`` extra); the ASGI middleware that applies it lives in :mod:`cogno_sync.asgi`.
+
+:class:`RateLimiter` is the port, :class:`InMemoryTokenBucketLimiter` the zero-infra default —
+**per process**, so N workers behind a balancer let one caller burst ~N× the configured rate —
+and :class:`RedisTokenBucketLimiter` the multi-worker adapter with shared counters.
+"""
+
+from __future__ import annotations
+
+import asyncio
+import logging
+import time
+from dataclasses import dataclass
+from typing import Any, Callable, Protocol, runtime_checkable
+
+from cogno_sync._redis import _RedisAdapter
+
+log = logging.getLogger(__name__)
+
+
+@dataclass(frozen=True)
+class RateLimitDecision:
+    """The outcome of a limiter check. ``retry_after`` (seconds) is meaningful only when denied."""
+
+    allowed: bool
+    retry_after: float = 0.0
+
+
+@runtime_checkable
+class RateLimiter(Protocol):
+    """Decides whether a request keyed by ``key`` may proceed. Async — a real adapter does I/O.
+    One ``check`` consumes one unit of budget."""
+
+    async def check(self, key: str) -> RateLimitDecision: ...
+
+
+@dataclass
+class _Bucket:
+    tokens: float
+    updated: float
+
+
+class InMemoryTokenBucketLimiter:
+    """Per-process token bucket: each ``key`` refills at ``rate_per_sec`` up to ``burst`` tokens;
+    a ``check`` spends one. ``burst`` is the max instantaneous spike; ``rate_per_sec`` the steady
+    state. Async-locked so concurrent requests for the same key don't race the refill."""
+
+    def __init__(self, *, rate_per_sec: float, burst: int,
+                 clock: Callable[[], float] = time.monotonic) -> None:
+        if rate_per_sec <= 0 or burst <= 0:
+            raise ValueError("rate_per_sec and burst must be positive")
+        self._rate = float(rate_per_sec)
+        self._burst = float(burst)
+        self._clock = clock
+        self._buckets: dict[str, _Bucket] = {}
+        self._lock = asyncio.Lock()
+
+    async def check(self, key: str) -> RateLimitDecision:
+        async with self._lock:
+            now = self._clock()
+            b = self._buckets.get(key)
+            if b is None:
+                b = _Bucket(tokens=self._burst, updated=now)
+                self._buckets[key] = b
+            else:                                    # refill for elapsed time, capped at burst
+                b.tokens = min(self._burst, b.tokens + (now - b.updated) * self._rate)
+                b.updated = now
+            if b.tokens >= 1.0:
+                b.tokens -= 1.0
+                return RateLimitDecision(True)
+            return RateLimitDecision(False, retry_after=(1.0 - b.tokens) / self._rate)
+
+
+# Atomic token bucket in one round-trip. Uses the Redis server clock (``TIME``) so every worker
+# shares one clock, works in integer milliseconds (Lua floats are lossy over the wire), and expires
+# idle buckets so the keyspace stays bounded. Returns ``{allowed(0|1), retry_after_ms}``.
+_LUA_TOKEN_BUCKET = """
+local key   = KEYS[1]
+local rate  = tonumber(ARGV[1])   -- tokens per second (steady state)
+local burst = tonumber(ARGV[2])   -- bucket capacity (max spike)
+local t = redis.call('TIME')
+local now_ms = tonumber(t[1]) * 1000 + math.floor(tonumber(t[2]) / 1000)
+local data = redis.call('HMGET', key, 'tokens', 'ts')
+local tokens = tonumber(data[1])
+local ts = tonumber(data[2])
+if tokens == nil then
+  tokens = burst
+  ts = now_ms
+end
+local elapsed = now_ms - ts
+if elapsed < 0 then elapsed = 0 end
+tokens = math.min(burst, tokens + elapsed * rate / 1000.0)
+local allowed = 0
+local retry_ms = 0
+if tokens >= 1.0 then
+  tokens = tokens - 1.0
+  allowed = 1
+else
+  retry_ms = math.ceil((1.0 - tokens) / rate * 1000.0)
+end
+redis.call('HSET', key, 'tokens', tokens, 'ts', now_ms)
+redis.call('EXPIRE', key, math.ceil(burst / rate) + 1)
+return {allowed, retry_ms}
+"""
+
+
+class RedisTokenBucketLimiter(_RedisAdapter):
+    """Token bucket over Redis — the **multi-worker** limiter (shared counters across processes).
+
+    The in-memory bucket is per-process, so N workers behind a balancer let a client burst ~N×
+    the configured rate; this adapter keeps one bucket per key in Redis (refilled atomically in a
+    Lua script under the Redis server clock), so the limit holds regardless of worker count. Same
+    :class:`RateLimiter` port as the in-memory default.
+
+    **Fail-open** (like the distributed lock): a Redis error → the request is *allowed* and a
+    warning is logged. A limiter is a cost/abuse guard, not an authentication check — refusing
+    every legitimate caller during a cache blip is worse than briefly not limiting.
+    """
+
+    def __init__(self, client: Any = None, *, url: str = "", key_prefix: str = "cogno:rl:",
+                 rate_per_sec: float, burst: int) -> None:
+        if rate_per_sec <= 0 or burst <= 0:
+            raise ValueError("rate_per_sec and burst must be positive")
+        super().__init__(client, url=url, key_prefix=key_prefix)
+        self._rate = float(rate_per_sec)
+        self._burst = int(burst)
+
+    async def check(self, key: str) -> RateLimitDecision:
+        try:
+            allowed, retry_ms = await self._conn().eval(
+                _LUA_TOKEN_BUCKET, 1, self._key(key), self._rate, self._burst)
+            if int(allowed) == 1:
+                return RateLimitDecision(True)
+            return RateLimitDecision(False, retry_after=float(retry_ms) / 1000.0)
+        except Exception as exc:   # noqa: BLE001 — fail-open: a Redis error never blocks a caller
+            log.warning("event=ratelimit_error key=%s error=%s proceeding_fail_open", key, exc)
+            return RateLimitDecision(True)
+
+
+__all__ = ["RateLimiter", "RateLimitDecision", "InMemoryTokenBucketLimiter",
+           "RedisTokenBucketLimiter"]

cogno_sync-0.1.1.dist-info/METADATA

@@ -0,0 +1,191 @@
+Metadata-Version: 2.4
+Name: cogno-sync
+Version: 0.1.1
+Summary: Distributed coordination primitives for the Cogno stack — async lock, token-bucket rate limiter, claim-once idempotency store and rolling daily cap, each with an in-process default and a Redis adapter
+Author: Vinicius Vale
+Maintainer: Sudoers AI
+License-Expression: Apache-2.0
+Project-URL: Homepage, https://github.com/sudoers-ai/cogno-sync
+Project-URL: Repository, https://github.com/sudoers-ai/cogno-sync
+Project-URL: Issues, https://github.com/sudoers-ai/cogno-sync/issues
+Keywords: redis,distributed-lock,rate-limiting,token-bucket,idempotency,quota,asgi,agents
+Classifier: Development Status :: 3 - Alpha
+Classifier: Intended Audience :: Developers
+Classifier: Programming Language :: Python :: 3
+Classifier: Programming Language :: Python :: 3.10
+Classifier: Programming Language :: Python :: 3.11
+Classifier: Programming Language :: Python :: 3.12
+Classifier: Topic :: Software Development :: Libraries
+Classifier: Topic :: System :: Distributed Computing
+Classifier: Operating System :: OS Independent
+Classifier: Typing :: Typed
+Requires-Python: >=3.10
+Description-Content-Type: text/markdown
+License-File: LICENSE
+Provides-Extra: redis
+Requires-Dist: redis>=5.0; extra == "redis"
+Provides-Extra: asgi
+Requires-Dist: starlette>=0.36; extra == "asgi"
+Provides-Extra: dev
+Requires-Dist: pytest>=8.0; extra == "dev"
+Requires-Dist: pytest-asyncio>=0.23.0; extra == "dev"
+Requires-Dist: pytest-cov>=5.0; extra == "dev"
+Requires-Dist: redis>=5.0; extra == "dev"
+Requires-Dist: starlette>=0.36; extra == "dev"
+Requires-Dist: httpx>=0.24; extra == "dev"
+Requires-Dist: ruff>=0.5; extra == "dev"
+Requires-Dist: mypy>=1.8; extra == "dev"
+Dynamic: license-file
+
+# cogno-sync
+
+**Distributed coordination primitives for the [Cogno](https://github.com/sudoers-ai/cogno-anima) stack** — an async lock, a token-bucket rate limiter, a claim-once idempotency store and a rolling daily cap. Each is a `Protocol` port with a zero-infra in-process default and a Redis adapter that shares the state across workers.
+
+> Status: **alpha** — extracted from a production host, with its suite.
+
+## The problem this is for
+
+Everything here exists because **one process is not the deployment**. The in-process versions of
+these four things are five lines each and they are all correct — until a second worker starts.
+Then the lock guards nothing, the limiter hands out N× the budget, the same delivery is handled
+twice, and each worker gives every caller its own daily allowance.
+
+| Port | In-process default | Redis adapter | What the adapter buys |
+| --- | --- | --- | --- |
+| `LockProvider` | `InProcessLockProvider` | `RedisLockProvider` | mutual exclusion **across** processes (`SET NX EX` + owner-only Lua release) |
+| `RateLimiter` | `InMemoryTokenBucketLimiter` | `RedisTokenBucketLimiter` | one shared bucket, refilled atomically under the **server's** clock |
+| `IdempotencyStore` | `InMemoryIdempotencyStore` | `RedisIdempotencyStore` | a claim two workers cannot both win (`SET NX`) |
+| `DailyCap` | `InMemoryDailyCap` | `RedisDailyCap` | one allowance per caller, not one per worker |
+
+## Everything fails OPEN
+
+A Redis outage, a lock that cannot be acquired within `max_wait`, a counter that will not
+answer: the work runs anyway, with a `WARNING`. These are guards against a rare double-run and
+against cost — never correctness barriers. Refusing to serve a legitimate caller because a cache
+blipped is the worse failure, and it is the one that takes the product down with the cache.
+
+If you need the opposite (a strict guard that must refuse when it cannot verify), do not
+configure your way there — wrap the port and decide it in your own code, where the trade-off is
+visible.
+
+## One client, one pool
+
+Every `Redis*` adapter here takes the **same** constructor:
+
+```python
+Adapter(client=None, *, url="", key_prefix="cogno:…:")
+```
+
+Build one client in your process and inject it everywhere:
+
+```python
+from cogno_sync import (client_from_url, RedisLockProvider, RedisTokenBucketLimiter,
+                        RedisIdempotencyStore, RedisDailyCap)
+
+redis = client_from_url(os.environ["REDIS_URL"], decode_responses=True)
+
+locks   = RedisLockProvider(redis)
+limiter = RedisTokenBucketLimiter(redis, rate_per_sec=10, burst=20)
+claims  = RedisIdempotencyStore(redis)
+cap     = RedisDailyCap(redis)
+```
+
+That uniformity is the feature. The code this was extracted from had **three** constructor
+conventions across five adapters and a copy of `Redis.from_url` in each one, so a worker held
+one connection pool per adapter — and a URL passed in the client slot failed on the first
+command, where fail-open swallowed it and the coordination was silently off for months.
+
+Passing `url=` instead builds the client lazily, on first use: handy for a one-adapter script,
+and it keeps start-up independent of whether the server is up yet.
+
+## Using them
+
+```python
+# serialize a critical section by key (across workers)
+async with locks.lock(f"session:{session_id}"):
+    ...
+
+# spend one unit of budget
+decision = await limiter.check(client_ip)
+if not decision.allowed:
+    return too_many_requests(retry_after=decision.retry_after)
+
+# handle a delivery once (claim BEFORE the work, release if nothing happened)
+if await claims.seen_before(key):
+    return
+try:
+    ...
+except NothingHappenedYet:
+    await claims.forget(key)
+
+# a rolling daily ceiling (the caller owns the key, including the day stamp)
+from cogno_sync import utc_day
+used = await cap.hit(f"demo:{utc_day()}:{client_ip}")
+if used > DAILY_LIMIT:
+    return politely_closed()
+```
+
+The **key is always yours to compose**. This library never invents one: what makes two deliveries
+distinct, or two callers separate, is a property of your system — and getting it wrong is
+silent (one caller's budget spent by another, one message dropped as a duplicate of a different
+one).
+
+## ASGI extra
+
+```bash
+pip install "cogno-sync[asgi]"
+```
+
+```python
+from cogno_sync.asgi import RateLimitMiddleware, read_capped_body, BodyTooLarge
+
+app.add_middleware(RateLimitMiddleware, limiter=limiter,
+                   prefixes=("/api", "/chat"),   # default: every path
+                   trusted_proxy_hops=1)         # how many proxies you trust in front
+```
+
+`trusted_proxy_hops` decides which `X-Forwarded-For` entry is the real caller — the N-th from
+the **right**, because each proxy appends the address it saw. Reading the leftmost hop is
+spoofable: a caller sets the header and gets a fresh budget on every request. It is a
+constructor argument, never an environment read at import time, so a test can set it and two
+apps in one runtime can differ.
+
+`read_capped_body` reads a request body incrementally and raises `BodyTooLarge` past the limit —
+for the unauthenticated routes where the signature can only be verified *after* the payload is
+read, and where `await request.body()` would buffer an unbounded stream into the worker's heap.
+
+## Install
+
+```bash
+pip install cogno-sync                # zero hard dependencies
+pip install "cogno-sync[redis]"       # the cross-worker adapters
+pip install "cogno-sync[asgi]"        # the middleware + body cap
+pip install -e ".[dev]"               # tests + lint + type-check
+```
+
+## The Cogno ecosystem
+
+`cogno-sync` is one organ of **[Cogno](https://github.com/sudoers-ai)** — a family of
+small, composable, Apache-2.0 libraries that together form a complete
+conversational-agent platform. Each library owns a single concern and stays
+infra-agnostic; a **host** assembles them into a running agent:
+
+![The Cogno ecosystem](docs/assets/cogno-ecosystem.svg)
+
+The open-source libraries are the organs; the **host is the body** that joins
+them. Our reference host — `cogno-host`, with its `cogno-ui` dashboard — is the
+private product layer, but it holds no special powers: everything it does rides
+on the public seams documented in each library's `docs/HOST_INTEGRATION.md`, so
+you can assemble a body of your own.
+
+## Test
+
+```bash
+pytest tests/unit -q                                    # no server needed
+COGNO_TEST_REDIS_URL=redis://127.0.0.1:6379/0 pytest tests/integration -q
+```
+
+The unit suite drives the Redis adapters against fakes; the integration suite proves the parts
+only a server can show — two clients sharing one bucket, a claim that survives across
+connections, the TTL Redis itself holds, and the release script comparing the owner token
+inside Redis.

cogno_sync-0.1.1.dist-info/RECORD

@@ -0,0 +1,13 @@
+cogno_sync/__init__.py,sha256=MBlNR5wt5lmACgK9pPMqaiHyaIDEoDd-WizzJDXO4a0,1896
+cogno_sync/_redis.py,sha256=f9VIqH3IdPgXiREaxoIZFBQ1B5Tgm5Yzw-5QTH_86Ak,3031
+cogno_sync/asgi.py,sha256=BQitlm_ZLaASV_wSnHcMKgtFJEJgWW9xMuH2cx1pl9Y,5524
+cogno_sync/daily_cap.py,sha256=CNUliqw6QL9PZ4jyK6dU7M3ZhT6IcSDt1uU1eVSxuzo,3361
+cogno_sync/idempotency.py,sha256=mniLF5Q6rmqBTtPgWgpP7JeA3IaxmhywL0TnQ_2az7k,5037
+cogno_sync/locks.py,sha256=i461BtPwbMoHFqSMZq9mHHFhPQTXEhyObesdbHAdrSc,8238
+cogno_sync/py.typed,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
+cogno_sync/ratelimit.py,sha256=wmyc8QzTZu7N_M8fmDpHjTlHJIjBXPN0hqi0iI_Am1o,5970
+cogno_sync-0.1.1.dist-info/licenses/LICENSE,sha256=4cK96oGhr1hZsZT3Bo5cnKQ3mx6DJ0u8XuTtJ-TNuso,11340
+cogno_sync-0.1.1.dist-info/METADATA,sha256=oVMF9Cc4KJPJlR2dtobSTULpHgPG_1aMl_jj40WGxm4,8449
+cogno_sync-0.1.1.dist-info/WHEEL,sha256=mOdXR9C3wMVov7cSY8PY-XCzo3SmmWCHvGQr5csJoao,91
+cogno_sync-0.1.1.dist-info/top_level.txt,sha256=OQxpIzp-QL2YqCL0EyzdaEm9CurWrBv7LcTGKMobZbc,11
+cogno_sync-0.1.1.dist-info/RECORD,,

cogno_sync-0.1.1.dist-info/WHEEL

@@ -0,0 +1,5 @@
+Wheel-Version: 1.0
+Generator: setuptools (84.2.1)
+Root-Is-Purelib: true
+Tag: py3-none-any
+

cogno_sync-0.1.1.dist-info/licenses/LICENSE

@@ -0,0 +1,201 @@
+                                 Apache License
+                           Version 2.0, January 2004
+                        http://www.apache.org/licenses/
+
+   TERMS AND CONDITIONS FOR USE, REPRODUCTION, AND DISTRIBUTION
+
+   1. Definitions.
+
+      "License" shall mean the terms and conditions for use, reproduction,
+      and distribution as defined by Sections 1 through 9 of this document.
+
+      "Licensor" shall mean the copyright owner or entity authorized by
+      the copyright owner that is granting the License.
+
+      "Legal Entity" shall mean the union of the acting entity and all
+      other entities that control, are controlled by, or are under common
+      control with that entity. For the purposes of this definition,
+      "control" means (i) the power, direct or indirect, to cause the
+      direction or management of such entity, whether by contract or
+      otherwise, or (ii) ownership of fifty percent (50%) or more of the
+      outstanding shares, or (iii) beneficial ownership of such entity.
+
+      "You" (or "Your") shall mean an individual or Legal Entity
+      exercising permissions granted by this License.
+
+      "Source" form shall mean the preferred form for making modifications,
+      including but not limited to software source code, documentation
+      source, and configuration files.
+
+      "Object" form shall mean any form resulting from mechanical
+      transformation or translation of a Source form, including but
+      not limited to compiled object code, generated documentation,
+      and conversions to other media types.
+
+      "Work" shall mean the work of authorship, whether in Source or
+      Object form, made available under the License, as indicated by a
+      copyright notice that is included in or attached to the work
+      (an example is provided in the Appendix below).
+
+      "Derivative Works" shall mean any work, whether in Source or Object
+      form, that is based on (or derived from) the Work and for which the
+      editorial revisions, annotations, elaborations, or other modifications
+      represent, as a whole, an original work of authorship. For the purposes
+      of this License, Derivative Works shall not include works that remain
+      separable from, or merely link (or bind by name) to the interfaces of,
+      the Work and Derivative Works thereof.
+
+      "Contribution" shall mean any work of authorship, including
+      the original version of the Work and any modifications or additions
+      to that Work or Derivative Works thereof, that is intentionally
+      submitted to Licensor for inclusion in the Work by the copyright owner
+      or by an individual or Legal Entity authorized to submit on behalf of
+      the copyright owner. For the purposes of this definition, "submitted"
+      means any form of electronic, verbal, or written communication sent
+      to the Licensor or its representatives, including but not limited to
+      communication on electronic mailing lists, source code control systems,
+      and issue tracking systems that are managed by, or on behalf of, the
+      Licensor for the purpose of discussing and improving the Work, but
+      excluding communication that is conspicuously marked or otherwise
+      designated in writing by the copyright owner as "Not a Contribution."
+
+      "Contributor" shall mean Licensor and any individual or Legal Entity
+      on behalf of whom a Contribution has been received by Licensor and
+      subsequently incorporated within the Work.
+
+   2. Grant of Copyright License. Subject to the terms and conditions of
+      this License, each Contributor hereby grants to You a perpetual,
+      worldwide, non-exclusive, no-charge, royalty-free, irrevocable
+      copyright license to reproduce, prepare Derivative Works of,
+      publicly display, publicly perform, sublicense, and distribute the
+      Work and such Derivative Works in Source or Object form.
+
+   3. Grant of Patent License. Subject to the terms and conditions of
+      this License, each Contributor hereby grants to You a perpetual,
+      worldwide, non-exclusive, no-charge, royalty-free, irrevocable
+      (except as stated in this section) patent license to make, have made,
+      use, offer to sell, sell, import, and otherwise transfer the Work,
+      where such license applies only to those patent claims licensable
+      by such Contributor that are necessarily infringed by their
+      Contribution(s) alone or by combination of their Contribution(s)
+      with the Work to which such Contribution(s) was submitted. If You
+      institute patent litigation against any entity (including a
+      cross-claim or counterclaim in a lawsuit) alleging that the Work
+      or a Contribution incorporated within the Work constitutes direct
+      or contributory patent infringement, then any patent licenses
+      granted to You under this License for that Work shall terminate
+      as of the date such litigation is filed.
+
+   4. Redistribution. You may reproduce and distribute copies of the
+      Work or Derivative Works thereof in any medium, with or without
+      modifications, and in Source or Object form, provided that You
+      meet the following conditions:
+
+      (a) You must give any other recipients of the Work or
+          Derivative Works a copy of this License; and
+
+      (b) You must cause any modified files to carry prominent notices
+          stating that You changed the files; and
+
+      (c) You must retain, in the Source form of any Derivative Works
+          that You distribute, all copyright, patent, trademark, and
+          attribution notices from the Source form of the Work,
+          excluding those notices that do not pertain to any part of
+          the Derivative Works; and
+
+      (d) If the Work includes a "NOTICE" text file as part of its
+          distribution, then any Derivative Works that You distribute must
+          include a readable copy of the attribution notices contained
+          within such NOTICE file, excluding those notices that do not
+          pertain to any part of the Derivative Works, in at least one
+          of the following places: within a NOTICE text file distributed
+          as part of the Derivative Works; within the Source form or
+          documentation, if provided along with the Derivative Works; or,
+          within a display generated by the Derivative Works, if and
+          wherever such third-party notices normally appear. The contents
+          of the NOTICE file are for informational purposes only and
+          do not modify the License. You may add Your own attribution
+          notices within Derivative Works that You distribute, alongside
+          or as an addendum to the NOTICE text from the Work, provided
+          that such additional attribution notices cannot be construed
+          as modifying the License.
+
+      You may add Your own copyright statement to Your modifications and
+      may provide additional or different license terms and conditions
+      for use, reproduction, or distribution of Your modifications, or
+      for any such Derivative Works as a whole, provided Your use,
+      reproduction, and distribution of the Work otherwise complies with
+      the conditions stated in this License.
+
+   5. Submission of Contributions. Unless You explicitly state otherwise,
+      any Contribution intentionally submitted for inclusion in the Work
+      by You to the Licensor shall be under the terms and conditions of
+      this License, without any additional terms or conditions.
+      Notwithstanding the above, nothing herein shall supersede or modify
+      the terms of any separate license agreement you may have executed
+      with Licensor regarding such Contributions.
+
+   6. Trademarks. This License does not grant permission to use the trade
+      names, trademarks, service marks, or product names of the Licensor,
+      except as required for reasonable and customary use in describing the
+      origin of the Work and reproducing the content of the NOTICE file.
+
+   7. Disclaimer of Warranty. Unless required by applicable law or
+      agreed to in writing, Licensor provides the Work (and each
+      Contributor provides its Contributions) on an "AS IS" BASIS,
+      WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or
+      implied, including, without limitation, any warranties or conditions
+      of TITLE, NON-INFRINGEMENT, MERCHANTABILITY, or FITNESS FOR A
+      PARTICULAR PURPOSE. You are solely responsible for determining the
+      appropriateness of using or redistributing the Work and assume any
+      risks associated with Your exercise of permissions under this License.
+
+   8. Limitation of Liability. In no event and under no legal theory,
+      whether in tort (including negligence), contract, or otherwise,
+      unless required by applicable law (such as deliberate and grossly
+      negligent acts) or agreed to in writing, shall any Contributor be
+      liable to You for damages, including any direct, indirect, special,
+      incidental, or consequential damages of any character arising as a
+      result of this License or out of the use or inability to use the
+      Work (including but not limited to damages for loss of goodwill,
+      work stoppage, computer failure or malfunction, or any and all
+      other commercial damages or losses), even if such Contributor
+      has been advised of the possibility of such damages.
+
+   9. Accepting Warranty or Additional Liability. While redistributing
+      the Work or Derivative Works thereof, You may choose to offer,
+      and charge a fee for, acceptance of support, warranty, indemnity,
+      or other liability obligations and/or rights consistent with this
+      License. However, in accepting such obligations, You may act only
+      on Your own behalf and on Your sole responsibility, not on behalf
+      of any other Contributor, and only if You agree to indemnify,
+      defend, and hold each Contributor harmless for any liability
+      incurred by, or claims asserted against, such Contributor by reason
+      of your accepting any such warranty or additional liability.
+
+   END OF TERMS AND CONDITIONS
+
+   APPENDIX: How to apply the Apache License to your work.
+
+      To apply the Apache License to your work, attach the following
+      boilerplate notice, with the fields enclosed by brackets "[]"
+      replaced with your own identifying information. (Don't include
+      the brackets!)  The text should be enclosed in the appropriate
+      comment syntax for the file format. We also recommend that a
+      file or class name and description of purpose be included on the
+      same "printed page" as the copyright notice for easier
+      identification within third-party archives.
+
+   Copyright 2026 Sudoers AI
+
+   Licensed under the Apache License, Version 2.0 (the "License");
+   you may not use this file except in compliance with the License.
+   You may obtain a copy of the License at
+
+       http://www.apache.org/licenses/LICENSE-2.0
+
+   Unless required by applicable law or agreed to in writing, software
+   distributed under the License is distributed on an "AS IS" BASIS,
+   WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+   See the License for the specific language governing permissions and
+   limitations under the License.

cogno_sync-0.1.1.dist-info/top_level.txt

@@ -0,0 +1 @@
+cogno_sync