nexo-brain 7.25.3 → 7.25.4

package/src/local_context/db.py

@@ -18,6 +18,7 @@ MAIN_CLEANUP_STATE_KEY = "local_context_main_tables_drained"
 LOCAL_CONTEXT_TABLES: tuple[str, ...] = (
     "local_index_roots",
     "local_index_exclusions",
+    "local_index_file_type_rules",
     "local_index_jobs",
     "local_index_checkpoints",
     "local_index_state",
@@ -103,10 +104,53 @@ def _ensure_schema(conn: sqlite3.Connection) -> None:
     _m63_local_context_layer(conn)
     _m64_local_context_live_dirs(conn)
     _ensure_entity_dossier_schema(conn)
-    conn.execute("PRAGMA user_version=64")
+    _ensure_local_context_v2_schema(conn)
+    conn.execute("PRAGMA user_version=65")
     conn.commit()
 
 
+def _table_columns(conn: sqlite3.Connection, table: str) -> set[str]:
+    rows = conn.execute(f"PRAGMA table_info({table})").fetchall()
+    return {str(row["name"] if isinstance(row, sqlite3.Row) else row[1]) for row in rows}
+
+
+def _add_column_if_missing(conn: sqlite3.Connection, table: str, column: str, definition: str) -> None:
+    if column in _table_columns(conn, table):
+        return
+    conn.execute(f"ALTER TABLE {table} ADD COLUMN {column} {definition}")
+
+
+def _ensure_local_context_v2_schema(conn: sqlite3.Connection) -> None:
+    """Idempotent local-index v2 schema for managed roots and file type rules."""
+    _add_column_if_missing(conn, "local_index_roots", "source", "TEXT NOT NULL DEFAULT 'legacy'")
+    _add_column_if_missing(conn, "local_index_roots", "remote", "INTEGER NOT NULL DEFAULT 0")
+    _add_column_if_missing(conn, "local_index_roots", "seed_version", "INTEGER NOT NULL DEFAULT 1")
+    _add_column_if_missing(conn, "local_index_exclusions", "source", "TEXT NOT NULL DEFAULT 'legacy'")
+    _add_column_if_missing(conn, "local_index_exclusions", "kind", "TEXT NOT NULL DEFAULT 'folder'")
+    conn.executescript(
+        """
+        CREATE TABLE IF NOT EXISTS local_index_file_type_rules (
+            id INTEGER PRIMARY KEY AUTOINCREMENT,
+            extension TEXT NOT NULL,
+            action TEXT NOT NULL DEFAULT 'ignore',
+            source TEXT NOT NULL DEFAULT 'user',
+            priority INTEGER NOT NULL DEFAULT 0,
+            reason TEXT NOT NULL DEFAULT '',
+            created_at REAL NOT NULL,
+            updated_at REAL NOT NULL,
+            UNIQUE(extension, source)
+        );
+
+        CREATE INDEX IF NOT EXISTS idx_local_index_roots_source
+            ON local_index_roots(source, status);
+        CREATE INDEX IF NOT EXISTS idx_local_index_exclusions_source
+            ON local_index_exclusions(source);
+        CREATE INDEX IF NOT EXISTS idx_local_index_file_type_rules_ext
+            ON local_index_file_type_rules(extension, source);
+        """
+    )
+
+
 def _ensure_entity_dossier_schema(conn: sqlite3.Connection) -> None:
     conn.executescript(
         """

package/src/local_context/extractors.py

@@ -68,6 +68,20 @@ def _read_text(path: Path) -> str:
     return data.decode("utf-8", errors="replace")[:MAX_CHARS]
 
 
+def _read_text_if_safe(path: Path) -> str:
+    data = path.read_bytes()[:MAX_TEXT_BYTES]
+    if not data or b"\x00" in data[:8192]:
+        return ""
+    text = _read_text(path)
+    if not text:
+        return ""
+    printable = sum(1 for char in text[:4096] if char.isprintable() or char.isspace())
+    sample_len = max(1, min(len(text), 4096))
+    if printable / sample_len < 0.85:
+        return ""
+    return text
+
+
 def _extract_csv(path: Path) -> str:
     text = _read_text(path)
     rows = []
@@ -291,7 +305,9 @@ def extract_text(path: Path) -> tuple[str, dict]:
     elif suffix == ".xlsx":
         text = _extract_xlsx(path)
     else:
-        text = ""
+        text = _read_text_if_safe(path)
+        if text:
+            metadata["extractor"] = "generic_text"
     if contains_secret(text):
         metadata["content_secret_detected"] = True
     return clean_text(text), metadata

package/src/server.py

@@ -1026,6 +1026,32 @@ def nexo_local_index_exclusions(action: str = "list", path: str = "", reason: st
     return json.dumps(result, ensure_ascii=False)
 
 
+@mcp.tool
+def nexo_local_index_filetypes(action: str = "list", extension: str = "", mode: str = "extract", reason: str = "user") -> str:
+    """List, include, exclude or reset local memory file extension rules."""
+    normalized = str(action or "list").strip().lower()
+    if normalized == "list":
+        result = local_context_api.list_file_type_rules(readonly=True)
+    elif normalized in {"include", "add"}:
+        result = local_context_api.set_file_type_rule(extension, action=mode or "extract", reason=reason)
+    elif normalized in {"exclude", "ignore"}:
+        result = local_context_api.set_file_type_rule(extension, action="ignore", reason=reason)
+    elif normalized in {"remove", "delete"}:
+        result = local_context_api.remove_file_type_rule(extension)
+    elif normalized == "reset":
+        result = local_context_api.reset_file_type_rules()
+    else:
+        result = {"ok": False, "error": "unknown_action", "allowed": ["list", "include", "exclude", "remove", "reset"]}
+    return json.dumps(result, ensure_ascii=False)
+
+
+@mcp.tool
+def nexo_local_index_migrate_roots_v2(apply: bool = False) -> str:
+    """Plan or apply Local Memory roots v2 cleanup."""
+    result = local_context_api.migrate_roots_seed_v2(dry_run=not bool(apply))
+    return json.dumps(result, ensure_ascii=False)
+
+
 @mcp.tool
 def nexo_local_context(
     query: str,

package/tool-enforcement-map.json

@@ -2502,6 +2502,37 @@
       },
       "triggers_after": []
     },
+    "nexo_local_index_filetypes": {
+      "description": "List, include, exclude, remove, or reset Local Context file extension rules",
+      "category": "local_context",
+      "source": "server",
+      "requires": [],
+      "provides": [
+        "local_index_file_type_rules"
+      ],
+      "internal_calls": [],
+      "enforcement": {
+        "level": "none",
+        "rules": []
+      },
+      "triggers_after": []
+    },
+    "nexo_local_index_migrate_roots_v2": {
+      "description": "Plan or apply Local Context roots v2 cleanup",
+      "category": "local_context",
+      "source": "server",
+      "requires": [],
+      "provides": [
+        "local_index_roots",
+        "local_index_cleanup"
+      ],
+      "internal_calls": [],
+      "enforcement": {
+        "level": "none",
+        "rules": []
+      },
+      "triggers_after": []
+    },
     "nexo_local_index_models": {
       "description": "Inspect or warm Local Context local model availability",
       "category": "local_context",