refactor(node): async generator for RecordBatchIterator (#2744)

JS native Async Generator, more efficient asynchronous iteration, fewer
synthetic promises, and the ability to handle `catch` or `break` of
parent loop in `finally` block

docs/src/js/classes/Query.md

@@ -80,7 +80,7 @@ AnalyzeExec verbose=true, metrics=[]
 ### execute()
 
 ```ts
-protected execute(options?): RecordBatchIterator
+protected execute(options?): AsyncGenerator<RecordBatch<any>, void, unknown>
 ```
 
 Execute the query and return the results as an
@@ -91,7 +91,7 @@ Execute the query and return the results as an
 
 #### Returns
 
-[`RecordBatchIterator`](RecordBatchIterator.md)
+`AsyncGenerator`&lt;`RecordBatch`&lt;`any`&gt;, `void`, `unknown`&gt;
 
 #### See
 

docs/src/js/classes/QueryBase.md

@@ -81,7 +81,7 @@ AnalyzeExec verbose=true, metrics=[]
 ### execute()
 
 ```ts
-protected execute(options?): RecordBatchIterator
+protected execute(options?): AsyncGenerator<RecordBatch<any>, void, unknown>
 ```
 
 Execute the query and return the results as an
@@ -92,7 +92,7 @@ Execute the query and return the results as an
 
 #### Returns
 
-[`RecordBatchIterator`](RecordBatchIterator.md)
+`AsyncGenerator`&lt;`RecordBatch`&lt;`any`&gt;, `void`, `unknown`&gt;
 
 #### See
 

docs/src/js/classes/RecordBatchIterator.md

@@ -1,43 +0,0 @@
-[**@lancedb/lancedb**](../README.md) • **Docs**
-
-***
-
-[@lancedb/lancedb](../globals.md) / RecordBatchIterator
-
-# Class: RecordBatchIterator
-
-## Implements
-
-- `AsyncIterator`<`RecordBatch`>
-
-## Constructors
-
-### new RecordBatchIterator()
-
-```ts
-new RecordBatchIterator(promise?): RecordBatchIterator
-```
-
-#### Parameters
-
-* **promise?**: `Promise`<`RecordBatchIterator`>
-
-#### Returns
-
-[`RecordBatchIterator`](RecordBatchIterator.md)
-
-## Methods
-
-### next()
-
-```ts
-next(): Promise<IteratorResult<RecordBatch<any>, any>>
-```
-
-#### Returns
-
-`Promise`&lt;`IteratorResult`&lt;`RecordBatch`&lt;`any`&gt;, `any`&gt;&gt;
-
-#### Implementation of
-
-`AsyncIterator.next`

docs/src/js/classes/TakeQuery.md

@@ -76,7 +76,7 @@ AnalyzeExec verbose=true, metrics=[]
 ### execute()
 
 ```ts
-protected execute(options?): RecordBatchIterator
+protected execute(options?): AsyncGenerator<RecordBatch<any>, void, unknown>
 ```
 
 Execute the query and return the results as an
@@ -87,7 +87,7 @@ Execute the query and return the results as an
 
 #### Returns
 
-[`RecordBatchIterator`](RecordBatchIterator.md)
+`AsyncGenerator`&lt;`RecordBatch`&lt;`any`&gt;, `void`, `unknown`&gt;
 
 #### See
 

docs/src/js/classes/VectorQuery.md

@@ -221,7 +221,7 @@ also increase the latency of your query. The default value is 1.5*limit.
 ### execute()
 
 ```ts
-protected execute(options?): RecordBatchIterator
+protected execute(options?): AsyncGenerator<RecordBatch<any>, void, unknown>
 ```
 
 Execute the query and return the results as an
@@ -232,7 +232,7 @@ Execute the query and return the results as an
 
 #### Returns
 
-[`RecordBatchIterator`](RecordBatchIterator.md)
+`AsyncGenerator`&lt;`RecordBatch`&lt;`any`&gt;, `void`, `unknown`&gt;
 
 #### See
 

docs/src/js/functions/RecordBatchIterator.md

@@ -0,0 +1,19 @@
+[**@lancedb/lancedb**](../README.md) • **Docs**
+
+***
+
+[@lancedb/lancedb](../globals.md) / RecordBatchIterator
+
+# Function: RecordBatchIterator()
+
+```ts
+function RecordBatchIterator(promisedInner): AsyncGenerator<RecordBatch<any>, void, unknown>
+```
+
+## Parameters
+
+* **promisedInner**: `Promise`&lt;`RecordBatchIterator`&gt;
+
+## Returns
+
+`AsyncGenerator`&lt;`RecordBatch`&lt;`any`&gt;, `void`, `unknown`&gt;

docs/src/js/globals.md

@@ -32,7 +32,6 @@
 - [PhraseQuery](classes/PhraseQuery.md)
 - [Query](classes/Query.md)
 - [QueryBase](classes/QueryBase.md)
-- [RecordBatchIterator](classes/RecordBatchIterator.md)
 - [Session](classes/Session.md)
 - [StaticHeaderProvider](classes/StaticHeaderProvider.md)
 - [Table](classes/Table.md)
@@ -105,6 +104,7 @@
 
 ## Functions
 
+- [RecordBatchIterator](functions/RecordBatchIterator.md)
 - [connect](functions/connect.md)
 - [makeArrowTable](functions/makeArrowTable.md)
 - [packBits](functions/packBits.md)

nodejs/lancedb/query.ts

@@ -20,35 +20,25 @@ import {
 } from "./native";
 import { Reranker } from "./rerankers";
 
-export class RecordBatchIterator implements AsyncIterator<RecordBatch> {
-  private promisedInner?: Promise<NativeBatchIterator>;
-  private inner?: NativeBatchIterator;
+export async function* RecordBatchIterator(
+  promisedInner: Promise<NativeBatchIterator>,
+) {
+  const inner = await promisedInner;
 
-  constructor(promise?: Promise<NativeBatchIterator>) {
-    // TODO: check promise reliably so we dont need to pass two arguments.
-    this.promisedInner = promise;
+  if (inner === undefined) {
+    throw new Error("Invalid iterator state");
   }
 
-  // biome-ignore lint/suspicious/noExplicitAny: skip
-  async next(): Promise<IteratorResult<RecordBatch<any>>> {
-    if (this.inner === undefined) {
-      this.inner = await this.promisedInner;
-    }
-    if (this.inner === undefined) {
-      throw new Error("Invalid iterator state state");
-    }
-    const n = await this.inner.next();
-    if (n == null) {
-      return Promise.resolve({ done: true, value: null });
-    }
-    const tbl = tableFromIPC(n);
-    if (tbl.batches.length != 1) {
+  for (let buffer = await inner.next(); buffer; buffer = await inner.next()) {
+    const { batches } = tableFromIPC(buffer);
+
+    if (batches.length !== 1) {
       throw new Error("Expected only one batch");
     }
-    return Promise.resolve({ done: false, value: tbl.batches[0] });
+
+    yield batches[0];
   }
 }
-/* eslint-enable */
 
 class RecordBatchIterable<
   NativeQueryType extends NativeQuery | NativeVectorQuery | NativeTakeQuery,
@@ -64,7 +54,7 @@ class RecordBatchIterable<
 
   // biome-ignore lint/suspicious/noExplicitAny: skip
   [Symbol.asyncIterator](): AsyncIterator<RecordBatch<any>, any, undefined> {
-    return new RecordBatchIterator(
+    return RecordBatchIterator(
       this.inner.execute(this.options?.maxBatchLength, this.options?.timeoutMs),
     );
   }
@@ -231,10 +221,8 @@ export class QueryBase<
    * single query)
    *
    */
-  protected execute(
-    options?: Partial<QueryExecutionOptions>,
-  ): RecordBatchIterator {
-    return new RecordBatchIterator(this.nativeExecute(options));
+  protected execute(options?: Partial<QueryExecutionOptions>) {
+    return RecordBatchIterator(this.nativeExecute(options));
   }
 
   /**
@@ -242,8 +230,7 @@ export class QueryBase<
    */
   // biome-ignore lint/suspicious/noExplicitAny: skip
   [Symbol.asyncIterator](): AsyncIterator<RecordBatch<any>> {
-    const promise = this.nativeExecute();
-    return new RecordBatchIterator(promise);
+    return RecordBatchIterator(this.nativeExecute());
   }
 
   /** Collect the results as an Arrow @see {@link ArrowTable}. */