fix: disable chromadb telemetry in nexus clients (#1427)

nexus/mempalace/searcher.py

@@ -53,13 +53,28 @@ def _get_client(palace_path: Path):
             "Run: pip install chromadb  (or: pip install mempalace)"
         ) from exc
 
+    try:
+        from chromadb.config import Settings  # type: ignore
+    except Exception:  # pragma: no cover - supports MagicMock-based tests
+        Settings = getattr(getattr(chromadb, "config", None), "Settings", None)
+        if Settings is None:
+            Settings = getattr(chromadb, "Settings", None)
+        if Settings is None:
+            raise MemPalaceUnavailable(
+                "ChromaDB Settings API unavailable. "
+                "Upgrade chromadb or verify the install."
+            )
+
     if not palace_path.exists():
         raise MemPalaceUnavailable(
             f"Palace directory not found: {palace_path}\n"
             "Run 'mempalace mine' to initialise the palace."
         )
 
-    return chromadb.PersistentClient(path=str(palace_path))
+    return chromadb.PersistentClient(
+        path=str(palace_path),
+        settings=Settings(anonymized_telemetry=False),
+    )
 
 
 def search_memories(

scaffold/deep-dive/relevance/relevance_engine.py

@@ -5,6 +5,7 @@ Filters and ranks content by Hermes/Timmy relevance
 """
 
 import chromadb
+from chromadb.config import Settings
 from chromadb.utils import embedding_functions
 from typing import List, Dict, Any
 import json
@@ -26,7 +27,10 @@ HERMES_CONTEXT = [
 
 class RelevanceEngine:
     def __init__(self, collection_name: str = "deep_dive"):
-        self.client = chromadb.PersistentClient(path="./chroma_db")
+        self.client = chromadb.PersistentClient(
+            path="./chroma_db",
+            settings=Settings(anonymized_telemetry=False),
+        )
         self.embedding_fn = embedding_functions.SentenceTransformerEmbeddingFunction(
             model_name="all-MiniLM-L6-v2"
         )

scripts/audit_mempalace_privacy.py

@@ -16,6 +16,7 @@ from pathlib import Path
 
 try:
     import chromadb
+    from chromadb.config import Settings
 except ImportError:
     print("ERROR: chromadb not installed")
     sys.exit(1)
@@ -34,7 +35,10 @@ VIOLATION_KEYWORDS = [
 
 def audit(palace_path: Path):
     violations = []
-    client = chromadb.PersistentClient(path=str(palace_path))
+    client = chromadb.PersistentClient(
+        path=str(palace_path),
+        settings=Settings(anonymized_telemetry=False),
+    )
     try:
         col = client.get_collection("mempalace_drawers")
     except Exception as e:

scripts/mempalace_export.py

@@ -10,6 +10,7 @@ import json
 import sys
 from pathlib import Path
 import chromadb
+from chromadb.config import Settings
 
 PALACE_PATH = "/root/wizards/bezalel/.mempalace/palace"
 FLEET_INCOMING = "/var/lib/mempalace/fleet/incoming"
@@ -18,7 +19,10 @@ DOCS_PER_ROOM = 5
 
 
 def main():
-    client = chromadb.PersistentClient(path=PALACE_PATH)
+    client = chromadb.PersistentClient(
+        path=PALACE_PATH,
+        settings=Settings(anonymized_telemetry=False),
+    )
     col = client.get_collection("mempalace_drawers")
 
     # Discover rooms in this wing

tests/test_chroma_telemetry_optout.py

@@ -0,0 +1,55 @@
+"""Regression tests for ChromaDB telemetry hardening.
+
+Issue #1427: ChromaDB defaults to anonymous telemetry unless explicitly disabled.
+These tests ensure every direct PersistentClient() call opts out.
+"""
+
+from __future__ import annotations
+
+from pathlib import Path
+from unittest.mock import MagicMock, patch
+
+from nexus.mempalace.searcher import _get_client
+
+PROJECT_ROOT = Path(__file__).parent.parent
+DIRECT_CLIENT_FILES = [
+    PROJECT_ROOT / "nexus/mempalace/searcher.py",
+    PROJECT_ROOT / "scripts/mempalace_export.py",
+    PROJECT_ROOT / "scripts/audit_mempalace_privacy.py",
+    PROJECT_ROOT / "scaffold/deep-dive/relevance/relevance_engine.py",
+]
+
+
+def test_get_client_disables_chroma_telemetry(tmp_path):
+    mock_chroma = MagicMock()
+    mock_settings = MagicMock(name="Settings")
+    mock_chroma.PersistentClient.return_value = "CLIENT"
+    mock_settings.return_value = "SETTINGS"
+
+    with patch.dict(
+        "sys.modules",
+        {
+            "chromadb": mock_chroma,
+            "chromadb.config": MagicMock(Settings=mock_settings),
+        },
+    ):
+        (tmp_path / "chroma.sqlite3").touch()
+        client = _get_client(tmp_path)
+
+    assert client == "CLIENT"
+    mock_settings.assert_called_once_with(anonymized_telemetry=False)
+    mock_chroma.PersistentClient.assert_called_once_with(path=str(tmp_path), settings="SETTINGS")
+
+
+def test_all_direct_persistent_clients_explicitly_disable_telemetry():
+    missing = []
+    needle = "anonymized_telemetry=False"
+    for path in DIRECT_CLIENT_FILES:
+        text = path.read_text(encoding="utf-8")
+        if needle not in text:
+            missing.append(path.relative_to(PROJECT_ROOT).as_posix())
+
+    assert not missing, (
+        "Direct Chroma PersistentClient call(s) still missing explicit telemetry opt-out:\n"
+        + "\n".join(f"  - {item}" for item in missing)
+    )