fix: address PR #813 review (owner_id index, cache bound, explicit param)

- vector/qdrant_client.py: add owner_id to _PAYLOAD_INDEX_FIELDS (BLOCKING).
  Every search applies MatchAny(key="owner_id", ...); without a keyword index
  Qdrant full-scans the collection and may 400 on Qdrant Cloud strict mode.
  _ensure_payload_indexes is idempotent so existing collections migrate at
  startup.
- search/access_filter.py: bound the process-global _owners_cache with an LRU
  cap (was one unbounded entry per active user, never evicted); document the
  owner-level over-fetch limitation (a prolific sharer floods the recall
  buffer with ghost candidates that verify-on-read drops, with no second
  Qdrant pass) as a TODO toward per-file filtering.
- search/algorithms.py + semantic.py + bm25_hybrid.py: promote
  accessible_owners from **kwargs to an explicit keyword-only parameter on the
  SearchAlgorithm ABC and both implementations, so a misspelled keyword is a
  type error rather than a silent fall back to self-only scope.
- search/verification.py: document that _verify_files now verifies by global
  file id (WebDAV SEARCH), not by path.
- tests/unit/search/test_access_filter.py: add cache-hit, TTL-expiry,
  failure-not-cached, and LRU-bound tests.

Bumps the astrolabe submodule with the matching #89 review fixes.

Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>

nextcloud_mcp_server/search/access_filter.py

@@ -18,6 +18,7 @@ from __future__ import annotations
 
 import logging
 import time
+from collections import OrderedDict
 from typing import Any, Protocol
 
 from qdrant_client.models import FieldCondition, Filter, MatchAny, MatchValue
@@ -30,7 +31,13 @@ logger = logging.getLogger(__name__)
 # an OCS round-trip per query. Safe: verify-on-read still gates each result
 # against Nextcloud, so a revoked share is caught there regardless of this cache.
 _OWNERS_CACHE_TTL_SECONDS = 30.0
-_owners_cache: dict[str, tuple[float, list[str]]] = {}
+# Cap the number of cached users so the process-global cache can't grow
+# unboundedly in a long-running multi-user deployment (one entry per active
+# user, never evicted otherwise). LRU eviction by insertion/access order via
+# OrderedDict; the cap is generous relative to any realistic concurrent-user
+# count, so steady state is effectively all-hit.
+_OWNERS_CACHE_MAXSIZE = 1024
+_owners_cache: OrderedDict[str, tuple[float, list[str]]] = OrderedDict()
 
 
 def clear_accessible_owners_cache() -> None:
@@ -64,12 +71,24 @@ async def list_accessible_owners(
     more incoming shares than the OCS page size could have some owners omitted;
     if that becomes real, add pagination to SharingClient.
 
+    Granularity / over-fetch limitation (TODO, finer-grained filtering): this
+    expansion is *owner-level*, not *file-level*. If a prolific content creator
+    shares a single item with the querying user, that owner's whole indexed
+    corpus becomes a Qdrant candidate set for the querier even though only the
+    shared item is accessible. Verify-on-read correctly drops the inaccessible
+    "ghost" candidates, but because there is no second Qdrant pass to replenish,
+    a ``limit=N`` search can return fewer than N results when the over-fetch
+    buffer (2× in nc_semantic_search / viz_routes) is dominated by ghosts. A
+    per-file ownership index would remove this tension and is the natural
+    starting point for future work (intentionally out of scope here).
+
     Sharing API failures are non-fatal — we degrade to ``[user_id]`` and log
     so a hiccup in OCS doesn't black-hole the user's own search.
     """
     now = time.monotonic()
     cached = _owners_cache.get(user_id)
     if cached is not None and now - cached[0] < _OWNERS_CACHE_TTL_SECONDS:
+        _owners_cache.move_to_end(user_id)  # mark as recently used (LRU)
         return list(cached[1])  # copy so callers can't mutate the cached value
 
     owners: set[str] = {user_id}
@@ -94,6 +113,9 @@ async def list_accessible_owners(
 
     result = list(owners)
     _owners_cache[user_id] = (now, result)
+    _owners_cache.move_to_end(user_id)  # newest = most-recently-used
+    while len(_owners_cache) > _OWNERS_CACHE_MAXSIZE:
+        _owners_cache.popitem(last=False)  # evict least-recently-used
     logger.debug("Accessible owners for user %s: %d entries", user_id, len(result))
     return list(result)
 

nextcloud_mcp_server/search/algorithms.py

@@ -274,6 +274,8 @@ class SearchAlgorithm(ABC):
         user_id: str,
         limit: int = 10,
         doc_type: str | None = None,
+        *,
+        accessible_owners: list[str] | None = None,
         **kwargs: Any,
     ) -> list[SearchResult]:
         """Execute search with the given parameters.
@@ -283,6 +285,12 @@ class SearchAlgorithm(ABC):
             user_id: User ID for multi-tenant filtering
             limit: Maximum number of results to return
             doc_type: Optional document type filter (note, file, calendar, etc.)
+            accessible_owners: Owner UIDs the user is allowed to read (self plus
+                the owners of content shared with them), pre-computed from the
+                OCS Sharing API by the caller. Declared explicitly — rather than
+                buried in ``**kwargs`` — so a misspelled keyword is a type error
+                instead of a silent fall back to self-only scope. ``None`` means
+                self-only (``[user_id]``).
             **kwargs: Algorithm-specific parameters
 
         Returns:

nextcloud_mcp_server/search/bm25_hybrid.py

@@ -71,6 +71,8 @@ class BM25HybridSearchAlgorithm(SearchAlgorithm):
         user_id: str,
         limit: int = 10,
         doc_type: str | None = None,
+        *,
+        accessible_owners: list[str] | None = None,
         **kwargs: Any,
     ) -> list[SearchResult]:
         """
@@ -89,6 +91,9 @@ class BM25HybridSearchAlgorithm(SearchAlgorithm):
             user_id: User ID for filtering
             limit: Maximum results to return
             doc_type: Optional document type filter
+            accessible_owners: Owner UIDs the user can read (self + share
+                senders), pre-computed by the caller from the OCS Sharing API.
+                Defaults to ``[user_id]`` (self-only) when ``None``.
             **kwargs: Additional parameters (score_threshold override)
 
         Returns:
@@ -99,7 +104,6 @@ class BM25HybridSearchAlgorithm(SearchAlgorithm):
         """
         settings = get_settings()
         score_threshold = kwargs.get("score_threshold", self.score_threshold)
-        accessible_owners: list[str] | None = kwargs.get("accessible_owners")
 
         logger.info(
             "BM25 hybrid search: query='%s', user=%s, limit=%s, score_threshold=%s, doc_type=%s, fusion=%s",

nextcloud_mcp_server/search/semantic.py

@@ -49,6 +49,8 @@ class SemanticSearchAlgorithm(SearchAlgorithm):
         user_id: str,
         limit: int = 10,
         doc_type: str | None = None,
+        *,
+        accessible_owners: list[str] | None = None,
         **kwargs: Any,
     ) -> list[SearchResult]:
         """Execute semantic search using vector similarity.
@@ -66,11 +68,11 @@ class SemanticSearchAlgorithm(SearchAlgorithm):
             user_id: User ID for filtering
             limit: Maximum results to return
             doc_type: Optional document type filter
+            accessible_owners: Owner UIDs the user can read (self + share
+                senders), pre-computed by the caller from the OCS Sharing API.
+                Defaults to ``[user_id]`` (self-only) when ``None``.
             **kwargs:
                 - score_threshold (float): override the instance default
-                - accessible_owners (list[str]): owner UIDs the user can read
-                  (self + share senders). Pre-computed by the caller from the
-                  OCS Sharing API. Defaults to ``[user_id]`` when omitted.
 
         Returns:
             List of unverified SearchResult objects ranked by similarity score
@@ -80,7 +82,6 @@ class SemanticSearchAlgorithm(SearchAlgorithm):
         """
         settings = get_settings()
         score_threshold = kwargs.get("score_threshold", self.score_threshold)
-        accessible_owners: list[str] | None = kwargs.get("accessible_owners")
 
         logger.info(
             "Semantic search: query='%s', user=%s, limit=%s, score_threshold=%s, doc_type=%s",

nextcloud_mcp_server/search/verification.py

@@ -132,6 +132,16 @@ async def _verify_files(
     results: list[SearchResult],
     semaphore: anyio.Semaphore,
 ) -> set[str]:
+    """Return the doc_ids of file results this user may actually access.
+
+    Verifies each file by its *global* Nextcloud file id via an ACL-aware
+    WebDAV SEARCH (``webdav.file_accessible_by_id``), NOT by path. This is the
+    ACL-aware-search fix: a file an owner shared with the querying user mounts
+    at a different path under each tree, so the previous path-based check
+    (``get_file_info``) produced false 404s and dropped legitimate shared-file
+    hits. Definitive 403/404 → inaccessible (dropped + scheduled for eviction
+    by the caller); transient/ambiguous errors → kept (fail-open).
+    """
     # safe: cooperative concurrency, no lock needed (see verify_search_results)
     accessible: set[str] = set()