@lobehub/chat 1.127.1 → 1.127.2

package/CHANGELOG.md

@@ -2,6 +2,31 @@
 
 # Changelog
 
+### [Version 1.127.2](https://github.com/lobehub/lobe-chat/compare/v1.127.1...v1.127.2)
+
+<sup>Released on **2025-09-11**</sup>
+
+#### 🐛 Bug Fixes
+
+- **misc**: Delete files should delete chunks、embedings、fileChunk.
+
+<br/>
+
+<details>
+<summary><kbd>Improvements and Fixes</kbd></summary>
+
+#### What's fixed
+
+- **misc**: Delete files should delete chunks、embedings、fileChunk, closes [#9196](https://github.com/lobehub/lobe-chat/issues/9196) ([4ee1d29](https://github.com/lobehub/lobe-chat/commit/4ee1d29))
+
+</details>
+
+<div align="right">
+
+[![](https://img.shields.io/badge/-BACK_TO_TOP-151515?style=flat-square)](#readme-top)
+
+</div>
+
 ### [Version 1.127.1](https://github.com/lobehub/lobe-chat/compare/v1.127.0...v1.127.1)
 
 <sup>Released on **2025-09-11**</sup>

package/changelog/v1.json

@@ -1,4 +1,13 @@
 [
+  {
+    "children": {
+      "fixes": [
+        "Delete files should delete chunks、embedings、fileChunk."
+      ]
+    },
+    "date": "2025-09-11",
+    "version": "1.127.2"
+  },
   {
     "children": {
       "fixes": [

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@lobehub/chat",
-  "version": "1.127.1",
+  "version": "1.127.2",
   "description": "Lobe Chat - an open-source, high-performance chatbot framework that supports speech synthesis, multimodal, and extensible Function Call plugin system. Supports one-click free deployment of your private ChatGPT/LLM web application.",
   "keywords": [
     "framework",

package/packages/database/src/models/__tests__/file.test.ts

@@ -4,7 +4,7 @@ import { afterEach, beforeEach, describe, expect, it, vi } from 'vitest';
 
 import { FilesTabs, SortType } from '@/types/files';
 
-import { files, globalFiles, knowledgeBaseFiles, knowledgeBases, users } from '../../schemas';
+import { chunks, embeddings, fileChunks, files, globalFiles, knowledgeBaseFiles, knowledgeBases, users } from '../../schemas';
 import { LobeChatDatabase } from '../../type';
 import { FileModel } from '../file';
 import { getTestDB } from './_util';
@@ -1058,15 +1058,8 @@ describe('FileModel', () => {
       // Create many chunks for this file to trigger batch processing
       // Note: This is a simplified test since we can't easily create 3000+ chunks
       // But it will still exercise the batch deletion code path
-      const chunkData = Array.from({ length: 10 }, (_, i) => ({
-        id: `chunk-${i}`,
-        text: `chunk content ${i}`,
-        index: i,
-        type: 'text' as const,
-        userId,
-      }));
 
-      // Insert chunks (this might need to be done through proper API)
+      // Insert chunks (this might need to be done through proper API)  
       // For testing purposes, we'll delete the file which should trigger the batch deletion
       await fileModel.delete(fileId, true);
 
@@ -1077,4 +1070,205 @@ describe('FileModel', () => {
       expect(deletedFile).toBeUndefined();
     });
   });
+
+  describe('deleteFileChunks error handling', () => {
+    let consoleWarnSpy: any;
+
+    beforeEach(() => {
+      consoleWarnSpy = vi.spyOn(console, 'warn').mockImplementation(() => {});
+    });
+
+    afterEach(() => {
+      consoleWarnSpy.mockRestore();
+    });
+
+    it('should delete file even when chunks deletion fails', async () => {
+      // 创建测试文件
+      const testFile = {
+        name: 'error-test-file.txt',
+        url: 'https://example.com/error-test-file.txt',
+        size: 100,
+        fileType: 'text/plain',
+        fileHash: 'error-test-hash',
+      };
+
+      const { id: fileId } = await fileModel.create(testFile, true);
+
+      // 创建一些测试数据来模拟chunks关联
+      const chunkId1 = '550e8400-e29b-41d4-a716-446655440001';
+      const chunkId2 = '550e8400-e29b-41d4-a716-446655440002';
+
+      // 插入chunks
+      await serverDB.insert(chunks).values([
+        { id: chunkId1, text: 'chunk 1', userId, type: 'text' },
+        { id: chunkId2, text: 'chunk 2', userId, type: 'text' },
+      ]);
+
+      // 插入fileChunks关联
+      await serverDB.insert(fileChunks).values([
+        { fileId, chunkId: chunkId1, userId },
+        { fileId, chunkId: chunkId2, userId },
+      ]);
+
+      // 插入embeddings (1024维向量)
+      const testEmbedding = new Array(1024).fill(0.1);
+      await serverDB.insert(embeddings).values([
+        { chunkId: chunkId1, embeddings: testEmbedding, model: 'test-model', userId },
+      ]);
+
+      // 跳过 documentChunks 测试，因为需要先创建 documents 记录
+
+      // 删除文件，应该会清理所有相关数据
+      const result = await fileModel.delete(fileId, true);
+
+      // 验证文件被删除
+      const deletedFile = await serverDB.query.files.findFirst({
+        where: eq(files.id, fileId),
+      });
+      expect(deletedFile).toBeUndefined();
+
+      // 验证chunks被删除
+      const remainingChunks = await serverDB.query.chunks.findMany({
+        where: inArray(chunks.id, [chunkId1, chunkId2]),
+      });
+      expect(remainingChunks).toHaveLength(0);
+
+      // 验证embeddings被删除
+      const remainingEmbeddings = await serverDB.query.embeddings.findMany({
+        where: inArray(embeddings.chunkId, [chunkId1, chunkId2]),
+      });
+      expect(remainingEmbeddings).toHaveLength(0);
+
+      // 验证fileChunks被删除
+      const remainingFileChunks = await serverDB.query.fileChunks.findMany({
+        where: eq(fileChunks.fileId, fileId),
+      });
+      expect(remainingFileChunks).toHaveLength(0);
+
+      expect(result).toBeDefined();
+    });
+
+    it('should successfully delete file with all related chunks and embeddings', async () => {
+      // 简化测试：只验证正常的完整删除流程（移除知识库保护后）
+      const testFile = {
+        name: 'complete-deletion-test.txt',
+        url: 'https://example.com/complete-deletion-test.txt',
+        size: 100,
+        fileType: 'text/plain',
+        fileHash: 'complete-deletion-hash',
+      };
+
+      const { id: fileId } = await fileModel.create(testFile, true);
+
+      const chunkId = '550e8400-e29b-41d4-a716-446655440003';
+
+      // 插入chunk
+      await serverDB.insert(chunks).values([
+        { id: chunkId, text: 'complete test chunk', userId, type: 'text' },
+      ]);
+
+      // 插入fileChunks关联
+      await serverDB.insert(fileChunks).values([
+        { fileId, chunkId, userId },
+      ]);
+
+      // 插入embeddings
+      const testEmbedding = new Array(1024).fill(0.1);
+      await serverDB.insert(embeddings).values([
+        { chunkId, embeddings: testEmbedding, model: 'test-model', userId },
+      ]);
+
+      // 删除文件
+      await fileModel.delete(fileId, true);
+
+      // 验证文件被删除
+      const deletedFile = await serverDB.query.files.findFirst({
+        where: eq(files.id, fileId),
+      });
+      expect(deletedFile).toBeUndefined();
+
+      // 验证chunks被删除
+      const remainingChunks = await serverDB.query.chunks.findMany({
+        where: eq(chunks.id, chunkId),
+      });
+      expect(remainingChunks).toHaveLength(0);
+
+      // 验证embeddings被删除
+      const remainingEmbeddings = await serverDB.query.embeddings.findMany({
+        where: eq(embeddings.chunkId, chunkId),
+      });
+      expect(remainingEmbeddings).toHaveLength(0);
+
+      // 验证fileChunks被删除
+      const remainingFileChunks = await serverDB.query.fileChunks.findMany({
+        where: eq(fileChunks.fileId, fileId),
+      });
+      expect(remainingFileChunks).toHaveLength(0);
+    });
+
+
+    it('should delete files that are in knowledge bases (removed protection)', async () => {
+      // 测试修复后的逻辑：知识库中的文件也应该被删除
+      const testFile = {
+        name: 'knowledge-base-file.txt',
+        url: 'https://example.com/knowledge-base-file.txt',
+        size: 100,
+        fileType: 'text/plain',
+        fileHash: 'kb-file-hash',
+        knowledgeBaseId: 'kb1',
+      };
+
+      const { id: fileId } = await fileModel.create(testFile, true);
+
+      const chunkId = '550e8400-e29b-41d4-a716-446655440007';
+
+      // 插入chunk和关联数据
+      await serverDB.insert(chunks).values([
+        { id: chunkId, text: 'knowledge base chunk', userId, type: 'text' },
+      ]);
+
+      await serverDB.insert(fileChunks).values([
+        { fileId, chunkId, userId },
+      ]);
+
+      // 插入embeddings (1024维向量)
+      const testEmbedding = new Array(1024).fill(0.1);
+      await serverDB.insert(embeddings).values([
+        { chunkId, embeddings: testEmbedding, model: 'test-model', userId },
+      ]);
+
+      // 验证文件确实在知识库中
+      const kbFile = await serverDB.query.knowledgeBaseFiles.findFirst({
+        where: eq(knowledgeBaseFiles.fileId, fileId),
+      });
+      expect(kbFile).toBeDefined();
+
+      // 删除文件
+      await fileModel.delete(fileId, true);
+
+      // 验证知识库中的文件也被完全删除
+      const deletedFile = await serverDB.query.files.findFirst({
+        where: eq(files.id, fileId),
+      });
+      expect(deletedFile).toBeUndefined();
+
+      // 验证chunks被删除（这是修复的核心：之前知识库文件的chunks不会被删除）
+      const remainingChunks = await serverDB.query.chunks.findMany({
+        where: eq(chunks.id, chunkId),
+      });
+      expect(remainingChunks).toHaveLength(0);
+
+      // 验证embeddings被删除
+      const remainingEmbeddings = await serverDB.query.embeddings.findMany({
+        where: eq(embeddings.chunkId, chunkId),
+      });
+      expect(remainingEmbeddings).toHaveLength(0);
+
+      // 验证fileChunks被删除
+      const remainingFileChunks = await serverDB.query.fileChunks.findMany({
+        where: eq(fileChunks.fileId, fileId),
+      });
+      expect(remainingFileChunks).toHaveLength(0);
+    });
+  });
 });

package/packages/database/src/models/file.ts

@@ -1,7 +1,6 @@
 import { and, asc, count, desc, eq, ilike, inArray, like, notExists, or, sum } from 'drizzle-orm';
 import type { PgTransaction } from 'drizzle-orm/pg-core';
 
-import { LobeChatDatabase, Transaction } from '../type';
 import { FilesTabs, QueryFileListParams, SortType } from '@/types/files';
 
 import {
@@ -9,12 +8,14 @@ import {
   NewFile,
   NewGlobalFile,
   chunks,
+  documentChunks,
   embeddings,
   fileChunks,
   files,
   globalFiles,
   knowledgeBaseFiles,
 } from '../schemas';
+import { LobeChatDatabase, Transaction } from '../type';
 
 export class FileModel {
   private readonly userId: string;
@@ -326,30 +327,19 @@ export class FileModel {
   private deleteFileChunks = async (trx: PgTransaction<any>, fileIds: string[]) => {
     if (fileIds.length === 0) return;
 
-    // 直接使用 JOIN 优化查询，减少数据传输量
+    // 获取要删除的文件相关的所有 chunk IDs（移除知识库保护逻辑）
     const relatedChunks = await trx
       .select({ chunkId: fileChunks.chunkId })
       .from(fileChunks)
-      .where(
-        and(
-          inArray(fileChunks.fileId, fileIds),
-          // 确保只查询有效的 chunkId
-          notExists(
-            trx
-              .select()
-              .from(knowledgeBaseFiles)
-              .where(eq(knowledgeBaseFiles.fileId, fileChunks.fileId)),
-          ),
-        ),
-      );
+      .where(inArray(fileChunks.fileId, fileIds));
 
     const chunkIds = relatedChunks.map((c) => c.chunkId).filter(Boolean) as string[];
 
     if (chunkIds.length === 0) return;
 
     // 批量处理配置
-    const BATCH_SIZE = 1000; // 增加批处理量
-    const MAX_CONCURRENT_BATCHES = 3; // 最大并行批次数
+    const BATCH_SIZE = 1000;
+    const MAX_CONCURRENT_BATCHES = 3;
 
     // 分批并行处理
     for (let i = 0; i < chunkIds.length; i += BATCH_SIZE * MAX_CONCURRENT_BATCHES) {
@@ -363,12 +353,31 @@ export class FileModel {
         const batchChunkIds = chunkIds.slice(startIdx, startIdx + BATCH_SIZE);
         if (batchChunkIds.length === 0) continue;
 
-        // 为每个批次创建一个删除任务
+        // 按正确的删除顺序处理每个批次，失败不阻止流程
         const batchPromise = (async () => {
-          // 先删除嵌入向量
-          await trx.delete(embeddings).where(inArray(embeddings.chunkId, batchChunkIds));
-          // 再删除块
-          await trx.delete(chunks).where(inArray(chunks.id, batchChunkIds));
+          // 1. 删除 embeddings (最顶层，有外键依赖)
+          try {
+            await trx.delete(embeddings).where(inArray(embeddings.chunkId, batchChunkIds));
+          } catch (e) {
+            // 静默处理，不阻止删除流程
+            console.warn('Failed to delete embeddings:', e);
+          }
+
+          // 2. 删除 documentChunks 关联 (如果存在)
+          try {
+            await trx.delete(documentChunks).where(inArray(documentChunks.chunkId, batchChunkIds));
+          } catch (e) {
+            // 静默处理，不阻止删除流程
+            console.warn('Failed to delete documentChunks:', e);
+          }
+
+          // 3. 删除 chunks (核心数据)
+          try {
+            await trx.delete(chunks).where(inArray(chunks.id, batchChunkIds));
+          } catch (e) {
+            // 静默处理，不阻止删除流程
+            console.warn('Failed to delete chunks:', e);
+          }
         })();
 
         batchPromises.push(batchPromise);
@@ -378,6 +387,14 @@ export class FileModel {
       await Promise.all(batchPromises);
     }
 
+    // 4. 最后删除 fileChunks 关联表记录
+    try {
+      await trx.delete(fileChunks).where(inArray(fileChunks.fileId, fileIds));
+    } catch (e) {
+      // 静默处理，不阻止删除流程
+      console.warn('Failed to delete fileChunks:', e);
+    }
+
     return chunkIds;
   };
 }