fix: address PR #814 review + SonarCloud gate

SonarCloud: - Resolve 6 S5332 hotspots (http→https in test fixture URLs). - S6418: hoist the unauthenticated AsyncOpenAI placeholder to a named constant + NOSONAR (genuine non-secret; gateway ignores it when unauthenticated). - Fix two reliability bugs: None-index guard in the gateway token-cache test (S2259) and float `> 0.0` instead of `!= 0.0` in the sentinel test (S1244). - status.py idle path sleeps 0.1s instead of sleep(0) (S7491); NOSONAR on the protocol-required async no-await aclose() stubs (S7503). Claude review: - Remove three leftover debug print() calls in app.py (logger.info already covers them). - payload_backfill: drop parsed_at from the backfilled-keys docstring (it is per-document state, not a deployment scalar); add a clean 404 precondition for BasicAuth deployments without an OAuth token verifier. - status.py: task_status typed TaskStatus | None (drop type: ignore). - nats.py: TODO to thread etags for file/deck/news; note etag default → None. - factory: warn on unknown INGEST_BUS_URL scheme; raise ValueError instead of assert for the external-mode preconditions. - docs/configuration.md: document the decomposition hook-point env vars + that nats-py ships core (lazy-imported) and external+bus uses two NATS connections. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-05-29 18:36:42 +02:00
parent a92f6260fb
commit b5ed1e3b4d
13 changed files with 91 additions and 25 deletions
@@ -750,6 +750,40 @@ docker-compose up

 ---

+## Decomposition Hook Points (Optional, Advanced)
+
+The server can optionally offload document processing and embeddings to external
+services (the Astrolabe Cloud document-processor and embedding-gateway). These
+are **opt-in**; every default reproduces the in-process monolith behavior, so
+self-hosters can ignore this section.
+
+```bash
+# Embeddings via an OpenAI-compatible gateway (else: autodetect — see above)
+EMBEDDING_PROVIDER=gateway
+EMBEDDING_GATEWAY_URL=https://embedding-gateway.internal
+# Gateway M2M OIDC client (its own realm; leave unset to call it unauthenticated)
+EMBEDDING_GATEWAY_TOKEN_URL=...
+EMBEDDING_GATEWAY_CLIENT_ID=...
+EMBEDDING_GATEWAY_CLIENT_SECRET=...
+
+# External ingest: publish to NATS instead of the in-process processor pool
+INGEST_MODE=external          # local (default) | external
+STATUS_BACKEND=bus            # local (default) | bus — REQUIRED with external
+INGEST_BUS_URL=nats://nats:4222
+TENANT_ID=<uuid>              # NATS per-tenant subject token
+```
+
+Notes:
+- `STATUS_BACKEND=local` with `INGEST_MODE=external` is rejected at startup
+  (the in-process job state is empty for externally-dispatched work).
+- **`nats-py` ships as a core dependency** (small, pure-Python) and is imported
+  lazily — only when `INGEST_MODE=external`. Self-hosters who never enable
+  external ingest pay no runtime cost.
+- `INGEST_MODE=external` + `STATUS_BACKEND=bus` opens **two** NATS connections
+  per pod (the ingest producer and the status subscriber are separate roles).
+
+---
+
 ## Tag-Based File Exclusion (Optional)

 Some files (contracts, medical records, credentials, private notes) should
@@ -5,8 +5,10 @@ values for any *missing* decomposition payload keys (so existing corpora gain
 them without a re-index), then upserts the collection-metadata sentinel.

 Scope: this backfills the cheap, deployment-level scalar keys
-(``processor_version``, ``parsed_at``, ``pipeline_tier``, ``embedding_identity``)
-only. It deliberately does NOT synthesize ``acl_hash`` — a correct value needs
+(``processor_version``, ``pipeline_tier``, ``embedding_identity``) only.
+``parsed_at`` is per-document state (the local processor sets it at index time),
+not a deployment-level scalar, so it is intentionally not backfilled. It also
+deliberately does NOT synthesize ``acl_hash`` — a correct value needs
 per-document share enumeration (a separate job), and writing a placeholder
 ``acl_hash`` would be unsafe to pre-filter on. The query-side ACL pre-filter
 therefore stays disabled until a real ACL backfill runs (see
@@ -38,6 +40,14 @@ logger = logging.getLogger(__name__)


 async def handle_payload_backfill(request: Request) -> JSONResponse:
+    # require_admin_scope authenticates via the OAuth token verifier; in
+    # BasicAuth-only deployments there is no oauth_context, so surface a clean
+    # 404 rather than a confusing 401 from the broad except below.
+    if getattr(request.app.state, "oauth_context", None) is None:
+        return JSONResponse(
+            {"error": "admin API requires an OAuth-capable deployment"},
+            status_code=404,
+        )
    try:
        await require_admin_scope(request)
    except AdminScopeRequired:
@@ -1120,9 +1120,6 @@ def get_app(transport: str = "streamable-http", enabled_apps: list[str] | None =
        and settings.vector_sync_enabled
        and settings.enable_background_operations
    ):
-        print(
-            f"DEBUG: Multi-user BasicAuth mode detected, vector_sync={settings.vector_sync_enabled}, background_operations={settings.enable_background_operations}"
-        )
        logger.info(
            "Multi-user BasicAuth with vector sync - checking for OAuth/app password credentials"
        )
@@ -1132,12 +1129,10 @@ def get_app(transport: str = "streamable-http", enabled_apps: list[str] | None =
        static_client_secret = os.getenv("NEXTCLOUD_OIDC_CLIENT_SECRET")

        if static_client_id and static_client_secret:
-            print("DEBUG: Using static OAuth credentials")
            logger.info("Using static OAuth credentials for background operations")
            multi_user_basic_oauth_creds = (static_client_id, static_client_secret)
        else:
            # Perform DCR before uvicorn starts (same lifecycle as OAuth modes)
-            print("DEBUG: No static credentials, attempting DCR...")
            logger.info(
                "OAuth credentials not configured - attempting Dynamic Client Registration..."
            )
@@ -35,6 +35,11 @@ logger = logging.getLogger(__name__)
 # token never expires mid-flight (matches AstrolabeClient / CP CLI behavior).
 _EARLY_REFRESH_SECONDS = 60

+# Non-secret placeholder for AsyncOpenAI, which rejects an empty key. In
+# unauthenticated mode the gateway ignores the bearer; when a token provider is
+# configured, the real M2M token replaces this before each request.
+_UNAUTHENTICATED_PLACEHOLDER = "unauthenticated"
+

 class GatewayTokenProvider:
    """Caches a gateway M2M access token via the ``client_credentials`` grant.
@@ -106,7 +111,7 @@ class GatewayProvider(OpenAIProvider):
        # the gateway is unauthenticated. When a token provider is configured,
        # the real Bearer is set on the client before each request.
        super().__init__(
-            api_key="gateway-unauthenticated",
+            api_key=_UNAUTHENTICATED_PLACEHOLDER,  # NOSONAR: placeholder, not a secret
            base_url=base_url,
            embedding_model=embedding_model,
            generation_model=None,  # gateway never generates
@@ -9,16 +9,25 @@ later) so moving the external processor to Postgres needs no new INGEST_MODE.

 from __future__ import annotations

+import logging
 from urllib.parse import urlsplit

 from ...config import Settings
 from .ports import TaskProducer

+logger = logging.getLogger(__name__)
+

 def _transport_for(url: str) -> str:
    scheme = urlsplit(url).scheme.lower()
    if scheme.startswith("postgres"):
        return "postgres"
+    if not scheme.startswith("nats"):
+        logger.warning(
+            "INGEST_BUS_URL scheme %r is neither nats:// nor postgres://; "
+            "defaulting to the NATS transport",
+            scheme,
+        )
    return "nats"


@@ -28,8 +37,13 @@ async def build_external_producer(settings: Settings) -> TaskProducer:
    Precondition: ``settings.ingest_mode == "external"`` (so __post_init__ has
    guaranteed ``ingest_bus_url`` and ``tenant_id`` are set).
    """
-    assert settings.ingest_bus_url is not None
-    assert settings.tenant_id is not None
+    # Defence-in-depth (robust under ``python -O``, which strips asserts):
+    # __post_init__ already guarantees these when ingest_mode == external.
+    if settings.ingest_bus_url is None or settings.tenant_id is None:
+        raise ValueError(
+            "build_external_producer requires INGEST_BUS_URL and TENANT_ID "
+            "(guaranteed by Settings validation when INGEST_MODE=external)"
+        )

    transport = _transport_for(settings.ingest_bus_url)
    if transport == "postgres":
@@ -36,7 +36,13 @@ def _modified_at_rfc3339(modified_at: int) -> str:

 def _content_hash(task: DocumentTask) -> str:
    """etag is the change-detection token; fall back to modified_at when it is
-    absent (e.g. deletes, or sources whose etag we don't thread through)."""
+    absent (e.g. deletes, or sources whose etag we don't thread through).
+
+    TODO(follow-up): thread etags for file / deck_card / news_item scans too
+    (only note scans pass etag today). Until then their JetStream Nats-Msg-Id
+    dedup keys off modified_at, which misses content changes that leave
+    modified_at unchanged (e.g. a file move/rename).
+    """
    return task.etag or str(task.modified_at)


@@ -133,7 +139,7 @@ class NatsTaskProducer:
    ) -> None:
        return None

-    async def aclose(self) -> None:
+    async def aclose(self) -> None:  # NOSONAR: async required by TaskProducer protocol
        # Per-handle close (e.g. a per-user scanner clone exiting). The bus
        # connection is shared and owned by the lifespan, so this is a no-op;
        # the connection is torn down once via ``drain()`` on shutdown.
@@ -45,5 +45,5 @@ class PostgresTaskProducer:
    ) -> None:  # pragma: no cover
        return None

-    async def aclose(self) -> None:  # pragma: no cover
+    async def aclose(self) -> None:  # pragma: no cover  # NOSONAR: protocol stub
        return None
@@ -134,7 +134,7 @@ class NatsStatusSubscriber:
        self,
        shutdown_event: anyio.Event,
        *,
-        task_status: TaskStatus = None,  # type: ignore[assignment]
+        task_status: TaskStatus | None = None,
    ) -> None:
        """Durable pull-consumer loop. Requires a live broker (integration)."""
        import anyio  # noqa: PLC0415
@@ -149,8 +149,9 @@ class NatsStatusSubscriber:
            try:
                msgs = await sub.fetch(batch=16, timeout=5)
            except Exception:
-                # fetch timeout when idle — loop and re-check shutdown.
-                await anyio.sleep(0)
+                # fetch timeout when idle — brief pause before re-checking
+                # shutdown, avoiding a tight check-and-sleep spin on quiet tenants.
+                await anyio.sleep(0.1)
                continue
            for msg in msgs:
                self.handle_message(msg.subject, msg.data)
@@ -350,7 +350,7 @@ async def scan_user_documents(
                        doc_type="note",
                        operation="index",
                        modified_at=modified_at,
-                        etag=note.get("etag", ""),
+                        etag=note.get("etag"),
                    )
                )
                queued += 1
@@ -417,7 +417,7 @@ async def scan_user_documents(
                            doc_type="note",
                            operation="index",
                            modified_at=modified_at,
-                            etag=note.get("etag", ""),
+                            etag=note.get("etag"),
                        )
                    )
                    queued += 1
@@ -29,7 +29,7 @@ def _patch_settings(monkeypatch, settings):
 def test_gateway_selected_unauthenticated(monkeypatch):
    settings = Settings(
        embedding_provider="gateway",
-        embedding_gateway_url="http://gateway:8083",
+        embedding_gateway_url="https://gateway:8083",
        embedding_gateway_model="mistral-embed",
    )
    _patch_settings(monkeypatch, settings)
@@ -44,7 +44,7 @@ def test_gateway_selected_unauthenticated(monkeypatch):
 def test_gateway_selected_with_m2m_oidc(monkeypatch):
    settings = Settings(
        embedding_provider="gateway",
-        embedding_gateway_url="http://gateway:8083",
+        embedding_gateway_url="https://gateway:8083",
        embedding_gateway_token_url="https://idp.example/oauth2/token",
        embedding_gateway_client_id="mcp-server",
        embedding_gateway_client_secret="shh",
@@ -60,7 +60,7 @@ def test_partial_m2m_creds_rejected():
    with pytest.raises(ValueError, match="must be set together"):
        Settings(
            embedding_provider="gateway",
-            embedding_gateway_url="http://gateway:8083",
+            embedding_gateway_url="https://gateway:8083",
            embedding_gateway_client_id="mcp-server",  # missing token_url/secret
        )

@@ -111,6 +111,7 @@ async def test_token_provider_caches_and_refreshes(monkeypatch):
    assert calls["n"] == 1

    # Expire the cache → next call refreshes.
+    assert tp._cache is not None
    tp._cache = (tp._cache[0], time.time() - 1)
    t3 = await tp.get_token()
    assert t3 == "tok2"
@@ -95,7 +95,7 @@ class TestConditionalRequired:
    def test_gateway_happy_path(self):
        s = Settings(
            embedding_provider="gateway",
-            embedding_gateway_url="http://gateway:8083",
+            embedding_gateway_url="https://gateway:8083",
        )
        assert s.embedding_provider == "gateway"

@@ -45,7 +45,7 @@ async def test_qdrant_error_falls_back_to_env(mocker):
 async def test_api_source(mocker):
    settings = Settings(
        collection_metadata_source="api",
-        collection_metadata_api_url="http://cp",
+        collection_metadata_api_url="https://cp",
    )

    def handler(request: httpx.Request) -> httpx.Response:
@@ -85,7 +85,7 @@ async def test_upsert_sentinel_builds_point(mocker):
    point = kwargs["points"][0]
    assert str(point.id) == cm.SENTINEL_POINT_ID
    # Non-zero dense (cosine-safe), empty sparse.
-    assert point.vector["dense"][0] != 0.0
+    assert point.vector["dense"][0] > 0.0
    assert len(point.vector["dense"]) == 4
    assert point.payload[EMBEDDING_IDENTITY] == "mistral-embed"
    assert point.payload[cm.IS_SENTINEL] is True
@@ -124,7 +124,7 @@ def test_publisher_matches_shared_fixture(mocker):
        ("nats://nats:4222", "nats"),
        ("postgres://h/db", "postgres"),
        ("postgresql://h/db", "postgres"),
-        ("http://elsewhere", "nats"),
+        ("https://elsewhere", "nats"),
    ],
 )
 def test_transport_for(url, expected):