feat: allow Python and Typescript users to create Sessions (#2530)

## Summary
- Exposes `Session` in Python and Typescript so users can set the
`index_cache_size_bytes` and `metadata_cache_size_bytes`
* The `Session` is attached to the `Connection`, and thus shared across
all tables in that connection.
- Adds deprecation warnings for table-level cache configuration


🤖 Generated with [Claude Code](https://claude.ai/code)

---------

Co-authored-by: Claude <noreply@anthropic.com>

nodejs/__test__/session.test.ts

@@ -0,0 +1,46 @@
+// SPDX-License-Identifier: Apache-2.0
+// SPDX-FileCopyrightText: Copyright The LanceDB Authors
+
+import * as tmp from "tmp";
+import { Session, connect } from "../lancedb";
+
+describe("Session", () => {
+  let tmpDir: tmp.DirResult;
+  beforeEach(() => {
+    tmpDir = tmp.dirSync({ unsafeCleanup: true });
+  });
+  afterEach(() => tmpDir.removeCallback());
+
+  it("should configure cache sizes and work with database operations", async () => {
+    // Create session with small cache limits for testing
+    const indexCacheSize = BigInt(1024 * 1024); // 1MB
+    const metadataCacheSize = BigInt(512 * 1024); // 512KB
+
+    const session = new Session(indexCacheSize, metadataCacheSize);
+
+    // Record initial cache state
+    const initialCacheSize = session.sizeBytes();
+    const initialCacheItems = session.approxNumItems();
+
+    // Test session works with database connection
+    const db = await connect({ uri: tmpDir.name, session: session });
+
+    // Create and use a table to exercise the session
+    const data = Array.from({ length: 100 }, (_, i) => ({
+      id: i,
+      text: `item ${i}`,
+    }));
+    const table = await db.createTable("test", data);
+    const results = await table.query().limit(5).toArray();
+
+    expect(results).toHaveLength(5);
+
+    // Verify cache usage increased after operations
+    const finalCacheSize = session.sizeBytes();
+    const finalCacheItems = session.approxNumItems();
+
+    expect(finalCacheSize).toBeGreaterThan(initialCacheSize); // Cache should have grown
+    expect(finalCacheItems).toBeGreaterThanOrEqual(initialCacheItems); // Items should not decrease
+    expect(initialCacheSize).toBeLessThan(indexCacheSize + metadataCacheSize); // Within limits
+  });
+});

nodejs/lancedb/connection.ts

@@ -85,6 +85,9 @@ export interface OpenTableOptions {
   /**
    * Set the size of the index cache, specified as a number of entries
    *
+   * @deprecated Use session-level cache configuration instead.
+   * Create a Session with custom cache sizes and pass it to the connect() function.
+   *
    * The exact meaning of an "entry" will depend on the type of index:
    * - IVF: there is one entry for each IVF partition
    * - BTREE: there is one entry for the entire index

nodejs/lancedb/index.ts

@@ -10,6 +10,7 @@ import {
 import {
   ConnectionOptions,
   Connection as LanceDbConnection,
+  Session,
 } from "./native.js";
 
 export {
@@ -51,6 +52,8 @@ export {
   OpenTableOptions,
 } from "./connection";
 
+export { Session } from "./native.js";
+
 export {
   ExecutableQuery,
   Query,
@@ -131,6 +134,7 @@ export { IntoSql, packBits } from "./util";
 export async function connect(
   uri: string,
   options?: Partial<ConnectionOptions>,
+  session?: Session,
 ): Promise<Connection>;
 /**
  * Connect to a LanceDB instance at the given URI.
@@ -149,31 +153,43 @@ export async function connect(
  *   storageOptions: {timeout: "60s"}
  * });
  * ```
+ *
+ * @example
+ * ```ts
+ * const session = Session.default();
+ * const conn = await connect({
+ *   uri: "/path/to/database",
+ *   session: session
+ * });
+ * ```
  */
 export async function connect(
   options: Partial<ConnectionOptions> & { uri: string },
 ): Promise<Connection>;
 export async function connect(
   uriOrOptions: string | (Partial<ConnectionOptions> & { uri: string }),
-  options: Partial<ConnectionOptions> = {},
+  options?: Partial<ConnectionOptions>,
 ): Promise<Connection> {
   let uri: string | undefined;
+  let finalOptions: Partial<ConnectionOptions> = {};
+
   if (typeof uriOrOptions !== "string") {
     const { uri: uri_, ...opts } = uriOrOptions;
     uri = uri_;
-    options = opts;
+    finalOptions = opts;
   } else {
     uri = uriOrOptions;
+    finalOptions = options || {};
   }
 
   if (!uri) {
     throw new Error("uri is required");
   }
 
-  options = (options as ConnectionOptions) ?? {};
-  (<ConnectionOptions>options).storageOptions = cleanseStorageOptions(
-    (<ConnectionOptions>options).storageOptions,
+  finalOptions = (finalOptions as ConnectionOptions) ?? {};
+  (<ConnectionOptions>finalOptions).storageOptions = cleanseStorageOptions(
+    (<ConnectionOptions>finalOptions).storageOptions,
   );
-  const nativeConn = await LanceDbConnection.new(uri, options);
+  const nativeConn = await LanceDbConnection.new(uri, finalOptions);
   return new LocalConnection(nativeConn);
 }

nodejs/src/connection.rs

@@ -74,6 +74,10 @@ impl Connection {
             builder = builder.host_override(&host_override);
         }
 
+        if let Some(session) = options.session {
+            builder = builder.session(session.inner.clone());
+        }
+
         Ok(Self::inner_new(builder.execute().await.default_error()?))
     }
 

nodejs/src/lib.rs

@@ -14,6 +14,7 @@ pub mod merge;
 mod query;
 pub mod remote;
 mod rerankers;
+mod session;
 mod table;
 mod util;
 
@@ -34,6 +35,9 @@ pub struct ConnectionOptions {
     ///
     /// The available options are described at https://lancedb.github.io/lancedb/guides/storage/
     pub storage_options: Option<HashMap<String, String>>,
+    /// (For LanceDB OSS only): the session to use for this connection. Holds
+    /// shared caches and other session-specific state.
+    pub session: Option<session::Session>,
 
     /// (For LanceDB cloud only): configuration for the remote HTTP client.
     pub client_config: Option<remote::ClientConfig>,

nodejs/src/session.rs

@@ -0,0 +1,102 @@
+// SPDX-License-Identifier: Apache-2.0
+// SPDX-FileCopyrightText: Copyright The LanceDB Authors
+
+use std::sync::Arc;
+
+use lancedb::{ObjectStoreRegistry, Session as LanceSession};
+use napi::bindgen_prelude::*;
+use napi_derive::*;
+
+/// A session for managing caches and object stores across LanceDB operations.
+///
+/// Sessions allow you to configure cache sizes for index and metadata caches,
+/// which can significantly impact memory use and performance. They can
+/// also be re-used across multiple connections to share the same cache state.
+#[napi]
+#[derive(Clone)]
+pub struct Session {
+    pub(crate) inner: Arc<LanceSession>,
+}
+
+impl std::fmt::Debug for Session {
+    fn fmt(&self, f: &mut std::fmt::Formatter<'_>) -> std::fmt::Result {
+        f.debug_struct("Session")
+            .field("size_bytes", &self.inner.size_bytes())
+            .field("approx_num_items", &self.inner.approx_num_items())
+            .finish()
+    }
+}
+
+#[napi]
+impl Session {
+    /// Create a new session with custom cache sizes.
+    ///
+    /// # Parameters
+    ///
+    /// - `index_cache_size_bytes`: The size of the index cache in bytes.
+    ///   Index data is stored in memory in this cache to speed up queries.
+    ///   Defaults to 6GB if not specified.
+    /// - `metadata_cache_size_bytes`: The size of the metadata cache in bytes.
+    ///   The metadata cache stores file metadata and schema information in memory.
+    ///   This cache improves scan and write performance.
+    ///   Defaults to 1GB if not specified.
+    #[napi(constructor)]
+    pub fn new(
+        index_cache_size_bytes: Option<BigInt>,
+        metadata_cache_size_bytes: Option<BigInt>,
+    ) -> napi::Result<Self> {
+        let index_cache_size = index_cache_size_bytes
+            .map(|size| size.get_u64().1 as usize)
+            .unwrap_or(6 * 1024 * 1024 * 1024); // 6GB default
+
+        let metadata_cache_size = metadata_cache_size_bytes
+            .map(|size| size.get_u64().1 as usize)
+            .unwrap_or(1024 * 1024 * 1024); // 1GB default
+
+        let session = LanceSession::new(
+            index_cache_size,
+            metadata_cache_size,
+            Arc::new(ObjectStoreRegistry::default()),
+        );
+
+        Ok(Self {
+            inner: Arc::new(session),
+        })
+    }
+
+    /// Create a session with default cache sizes.
+    ///
+    /// This is equivalent to creating a session with 6GB index cache
+    /// and 1GB metadata cache.
+    #[napi(factory)]
+    pub fn default() -> Self {
+        Self {
+            inner: Arc::new(LanceSession::default()),
+        }
+    }
+
+    /// Get the current size of the session caches in bytes.
+    #[napi]
+    pub fn size_bytes(&self) -> BigInt {
+        BigInt::from(self.inner.size_bytes())
+    }
+
+    /// Get the approximate number of items cached in the session.
+    #[napi]
+    pub fn approx_num_items(&self) -> u32 {
+        self.inner.approx_num_items() as u32
+    }
+}
+
+// Implement FromNapiValue for Session to work with napi(object)
+impl napi::bindgen_prelude::FromNapiValue for Session {
+    unsafe fn from_napi_value(
+        env: napi::sys::napi_env,
+        napi_val: napi::sys::napi_value,
+    ) -> napi::Result<Self> {
+        let object: napi::bindgen_prelude::ClassInstance<Session> =
+            napi::bindgen_prelude::ClassInstance::from_napi_value(env, napi_val)?;
+        let copy = object.clone();
+        Ok(copy)
+    }
+}