From bb23668f3cdc4d88c3157076698a6aa7e10d073a Mon Sep 17 00:00:00 2001
From: Kevin Tam <kevin@glorat.net>
Date: Tue, 25 Jun 2024 09:36:58 +0800
Subject: [PATCH] fix: Document should handle empty pageContent (fixes #5884)

---
 langchain-core/src/documents/document.ts  | 8 ++++----
 langchain-core/src/tests/document.test.ts | 7 +++++++
 2 files changed, 11 insertions(+), 4 deletions(-)
 create mode 100644 langchain-core/src/tests/document.test.ts

diff --git a/langchain-core/src/documents/document.ts b/langchain-core/src/documents/document.ts
index 5979bbf63756..1d27d8002a20 100644
--- a/langchain-core/src/documents/document.ts
+++ b/langchain-core/src/documents/document.ts
@@ -20,8 +20,8 @@ export interface DocumentInterface<
  * Interface for interacting with a document.
  */
 export class Document<
-  // eslint-disable-next-line @typescript-eslint/no-explicit-any
-  Metadata extends Record<string, any> = Record<string, any>
+    // eslint-disable-next-line @typescript-eslint/no-explicit-any
+    Metadata extends Record<string, any> = Record<string, any>
 > implements DocumentInput, DocumentInterface
 {
   pageContent: string;
@@ -29,9 +29,9 @@ export class Document<
   metadata: Metadata;
 
   constructor(fields: DocumentInput<Metadata>) {
-    this.pageContent = fields.pageContent
+    this.pageContent = fields.pageContent !== undefined
       ? fields.pageContent.toString()
-      : this.pageContent;
+      : "";
     this.metadata = fields.metadata ?? ({} as Metadata);
   }
 }
diff --git a/langchain-core/src/tests/document.test.ts b/langchain-core/src/tests/document.test.ts
new file mode 100644
index 000000000000..e839621b2b4f
--- /dev/null
+++ b/langchain-core/src/tests/document.test.ts
@@ -0,0 +1,7 @@
+import {test, expect} from '@jest/globals'
+import {Document} from '../documents/document.js'
+
+test('Document should handle empty pageContent', () => {
+  const doc = new Document({pageContent: ''})
+  expect(doc.pageContent).toEqual('')
+})