Self-evolving Context Database for AI Agents. Unify Agent Memory, Knowledge RAG and Skills.
Python agent-contextagent-memoryagent-pluginsagentic-ragcontext-databasedsh-pluginknowledge-basememorymemory-managementself-evolvingvector-database
fix(server): read attrs tags by exact URI and level (#3707)
Jiahui Zhoua718da7e8525
diff --git a/openviking/server/routers/filesystem.py b/openviking/server/routers/filesystem.pyindex a0c415e85..557f98b2b 100644--- a/openviking/server/routers/filesystem.py+++ b/openviking/server/routers/filesystem.py@@ -17,6 +17,7 @@ from openviking.server.identity import RequestContextfrom openviking.server.models import Responsefrom openviking.server.routers.content import SetTagsRequestfrom openviking.server.routers.content import set_tags as content_set_tags+from openviking.storage.expr import And, Eq, Infrom openviking.storage.vikingdb_manager import VikingDBManagerProxyfrom openviking_cli.exceptions import InvalidArgumentError, NotFoundError@@ -35,14 +36,22 @@ def _clean_memory_attrs(raw: str) -> dict[str, Any]:return attrs-async def _tags_attr(service: Any, uri: str, ctx: RequestContext) -> list[str]:+async def _tags_attr(+ service: Any, uri: str, ctx: RequestContext, *, is_dir: bool+) -> list[str]:vikingdb_manager = getattr(service, "vikingdb_manager", None)if not vikingdb_manager:return []+ # Tags are written per level (see ContentWriteCoordinator.set_tags): a+ # directory carries them on its L0/L1 summary records, while a file carries+ # them on its L2 content record. Mirror that here so the exact node's tags+ # are read back instead of unrelated records. Eq("uri", ...) compiles to an+ # exact path match (-d=0), avoiding prefix/subtree matches over the path field.+ levels = [0, 1] if is_dir else [2]proxy = VikingDBManagerProxy(vikingdb_manager, ctx)records = await proxy.filter(- filter={"op": "must", "field": "uri", "conds": [uri]},+ filter=And([Eq("uri", uri), In("level", levels)]),limit=10,output_fields=_ATTR_INDEX_FIELDS,)@@ -181,7 +190,9 @@ async def attrs("uri": canonical_uri,"context_type": context_type_for_uri(canonical_uri),"attrs": {- "tags": await _tags_attr(service, canonical_uri, _ctx),+ "tags": await _tags_attr(+ service, canonical_uri, _ctx, is_dir=stat_result.get("isDir", False)+ ),},}if result["context_type"] == "memory" and not stat_result.get("isDir", False):