@lobehub/lobehub 2.0.0-next.110 → 2.0.0-next.111

package/CHANGELOG.md

@@ -2,6 +2,32 @@
 
 # Changelog
 
+## [Version 2.0.0-next.111](https://github.com/lobehub/lobe-chat/compare/v2.0.0-next.110...v2.0.0-next.111)
+
+<sup>Released on **2025-11-24**</sup>
+
+#### 🐛 Bug Fixes
+
+- **misc**: Fix db migration snapshot not align with db schema, Separate agent file injection from knowledge base RAG search.
+
+<br/>
+
+<details>
+<summary><kbd>Improvements and Fixes</kbd></summary>
+
+#### What's fixed
+
+- **misc**: Fix db migration snapshot not align with db schema, closes [#10399](https://github.com/lobehub/lobe-chat/issues/10399) ([760105a](https://github.com/lobehub/lobe-chat/commit/760105a))
+- **misc**: Separate agent file injection from knowledge base RAG search, closes [#10398](https://github.com/lobehub/lobe-chat/issues/10398) ([e1c813a](https://github.com/lobehub/lobe-chat/commit/e1c813a))
+
+</details>
+
+<div align="right">
+
+[![](https://img.shields.io/badge/-BACK_TO_TOP-151515?style=flat-square)](#readme-top)
+
+</div>
+
 ## [Version 2.0.0-next.110](https://github.com/lobehub/lobe-chat/compare/v2.0.0-next.109...v2.0.0-next.110)
 
 <sup>Released on **2025-11-24**</sup>

package/changelog/v1.json

@@ -1,4 +1,13 @@
 [
+  {
+    "children": {
+      "fixes": [
+        "Fix db migration snapshot not align with db schema, Separate agent file injection from knowledge base RAG search."
+      ]
+    },
+    "date": "2025-11-24",
+    "version": "2.0.0-next.111"
+  },
   {
     "children": {
       "improvements": [

package/docs/development/database-schema.dbml

@@ -187,6 +187,7 @@ table documents {
   user_id text [not null]
   client_id text
   editor_data jsonb
+  slug varchar(255)
   accessed_at "timestamp with time zone" [not null, default: `now()`]
   created_at "timestamp with time zone" [not null, default: `now()`]
   updated_at "timestamp with time zone" [not null, default: `now()`]
@@ -197,6 +198,7 @@ table documents {
     file_id [name: 'documents_file_id_idx']
     parent_id [name: 'documents_parent_id_idx']
     (client_id, user_id) [name: 'documents_client_id_user_id_unique', unique]
+    (slug, user_id) [name: 'documents_slug_user_id_unique', unique]
   }
 }
 
@@ -214,7 +216,6 @@ table files {
   metadata jsonb
   chunk_task_id uuid
   embedding_task_id uuid
-  slug text [unique]
   accessed_at "timestamp with time zone" [not null, default: `now()`]
   created_at "timestamp with time zone" [not null, default: `now()`]
   updated_at "timestamp with time zone" [not null, default: `now()`]

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@lobehub/lobehub",
-  "version": "2.0.0-next.110",
+  "version": "2.0.0-next.111",
   "description": "LobeHub - an open-source,comprehensive AI Agent 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/context-engine/src/index.ts

@@ -10,7 +10,7 @@ export type { ContextEngineConfig } from './pipeline';
 export { ContextEngine } from './pipeline';
 
 // Context Providers
-export { HistorySummaryProvider, SystemRoleInjector, ToolSystemRoleProvider } from './providers';
+export * from './providers';
 
 // Processors
 export {

package/packages/context-engine/src/providers/KnowledgeInjector.ts

@@ -0,0 +1,78 @@
+import { promptAgentKnowledge } from '@lobechat/prompts';
+import type { FileContent, KnowledgeBaseInfo } from '@lobechat/prompts';
+import debug from 'debug';
+
+import { BaseProvider } from '../base/BaseProvider';
+import type { PipelineContext, ProcessorOptions } from '../types';
+
+const log = debug('context-engine:provider:KnowledgeInjector');
+
+export interface KnowledgeInjectorConfig {
+  /** File contents to inject */
+  fileContents?: FileContent[];
+  /** Knowledge bases to inject */
+  knowledgeBases?: KnowledgeBaseInfo[];
+}
+
+/**
+ * Knowledge Injector
+ * Responsible for injecting agent's knowledge (files and knowledge bases) into context
+ */
+export class KnowledgeInjector extends BaseProvider {
+  readonly name = 'KnowledgeInjector';
+
+  constructor(
+    private config: KnowledgeInjectorConfig,
+    options: ProcessorOptions = {},
+  ) {
+    super(options);
+  }
+
+  protected async doProcess(context: PipelineContext): Promise<PipelineContext> {
+    const clonedContext = this.cloneContext(context);
+
+    const fileContents = this.config.fileContents || [];
+    const knowledgeBases = this.config.knowledgeBases || [];
+
+    // Generate unified knowledge prompt
+    const formattedContent = promptAgentKnowledge({ fileContents, knowledgeBases });
+
+    // Skip injection if no knowledge at all
+    if (!formattedContent) {
+      log('No knowledge to inject');
+      return this.markAsExecuted(clonedContext);
+    }
+
+    // Find the first user message index
+    const firstUserIndex = clonedContext.messages.findIndex((msg) => msg.role === 'user');
+
+    if (firstUserIndex === -1) {
+      log('No user messages found, skipping injection');
+      return this.markAsExecuted(clonedContext);
+    }
+
+    // Insert a new user message with knowledge before the first user message
+    // Mark it as application-level system injection
+    const knowledgeMessage = {
+      content: formattedContent,
+      createdAt: Date.now(),
+      id: `knowledge-${Date.now()}`,
+      meta: { injectType: 'knowledge', systemInjection: true },
+      role: 'user' as const,
+      updatedAt: Date.now(),
+    };
+
+    clonedContext.messages.splice(firstUserIndex, 0, knowledgeMessage);
+
+    // Update metadata
+    clonedContext.metadata.knowledgeInjected = true;
+    clonedContext.metadata.filesCount = fileContents.length;
+    clonedContext.metadata.knowledgeBasesCount = knowledgeBases.length;
+
+    log(
+      `Agent knowledge injected as new user message: ${fileContents.length} file(s), ${knowledgeBases.length} knowledge base(s)`,
+    );
+
+    return this.markAsExecuted(clonedContext);
+  }
+}

package/packages/context-engine/src/providers/index.ts

@@ -1,9 +1,11 @@
 // Context Provider exports
 export { HistorySummaryProvider } from './HistorySummary';
+export { KnowledgeInjector } from './KnowledgeInjector';
 export { SystemRoleInjector } from './SystemRoleInjector';
 export { ToolSystemRoleProvider } from './ToolSystemRole';
 
 // Re-export types
 export type { HistorySummaryConfig } from './HistorySummary';
+export type { KnowledgeInjectorConfig } from './KnowledgeInjector';
 export type { SystemRoleInjectorConfig } from './SystemRoleInjector';
 export type { ToolSystemRoleConfig } from './ToolSystemRole';

package/packages/database/migrations/0047_add_slug_document.sql

@@ -1,6 +1,2 @@
 ALTER TABLE "documents" ADD COLUMN IF NOT EXISTS "slug" varchar(255);--> statement-breakpoint
-DO $$ BEGIN
- CREATE UNIQUE INDEX IF NOT EXISTS "documents_slug_user_id_unique" ON "documents" ("slug","user_id") WHERE "slug" IS NOT NULL;
-EXCEPTION
- WHEN duplicate_object THEN null;
-END $$;
+CREATE UNIQUE INDEX IF NOT EXISTS "documents_slug_user_id_unique" ON "documents" USING btree ("slug","user_id") WHERE "documents"."slug" is not null;

package/packages/database/migrations/meta/0047_snapshot.json

@@ -6,7 +6,7 @@
   },
   "dialect": "postgresql",
   "enums": {},
-  "id": "8a423b68-c609-44e1-bd48-f2ed62fec5ef",
+  "id": "20a4c30c-04f0-4993-b493-0ae6a3144f9e",
   "policies": {},
   "prevId": "abe0cb47-a970-4ec0-a29f-1afa9bd2fe02",
   "roles": {},
@@ -1259,6 +1259,12 @@
           "primaryKey": false,
           "notNull": false
         },
+        "slug": {
+          "name": "slug",
+          "type": "varchar(255)",
+          "primaryKey": false,
+          "notNull": false
+        },
         "accessed_at": {
           "name": "accessed_at",
           "type": "timestamp with time zone",
@@ -1362,6 +1368,28 @@
           "concurrently": false,
           "method": "btree",
           "with": {}
+        },
+        "documents_slug_user_id_unique": {
+          "name": "documents_slug_user_id_unique",
+          "columns": [
+            {
+              "expression": "slug",
+              "isExpression": false,
+              "asc": true,
+              "nulls": "last"
+            },
+            {
+              "expression": "user_id",
+              "isExpression": false,
+              "asc": true,
+              "nulls": "last"
+            }
+          ],
+          "isUnique": true,
+          "where": "\"documents\".\"slug\" is not null",
+          "concurrently": false,
+          "method": "btree",
+          "with": {}
         }
       },
       "foreignKeys": {
@@ -1481,12 +1509,6 @@
           "primaryKey": false,
           "notNull": false
         },
-        "slug": {
-          "name": "slug",
-          "type": "text",
-          "primaryKey": false,
-          "notNull": false
-        },
         "accessed_at": {
           "name": "accessed_at",
           "type": "timestamp with time zone",
@@ -1610,13 +1632,7 @@
         }
       },
       "compositePrimaryKeys": {},
-      "uniqueConstraints": {
-        "files_slug_unique": {
-          "name": "files_slug_unique",
-          "nullsNotDistinct": false,
-          "columns": ["slug"]
-        }
-      },
+      "uniqueConstraints": {},
       "policies": {},
       "checkConstraints": {},
       "isRLSEnabled": false

package/packages/database/migrations/meta/_journal.json

@@ -333,7 +333,7 @@
     {
       "idx": 47,
       "version": "7",
-      "when": 1763535087148,
+      "when": 1763987922211,
       "tag": "0047_add_slug_document",
       "breakpoints": true
     }

package/packages/database/src/core/migrations.json

@@ -792,10 +792,10 @@
   {
     "sql": [
       "ALTER TABLE \"documents\" ADD COLUMN IF NOT EXISTS \"slug\" varchar(255);",
-      "\nDO $$ BEGIN\n CREATE UNIQUE INDEX IF NOT EXISTS \"documents_slug_user_id_unique\" ON \"documents\" (\"slug\",\"user_id\") WHERE \"slug\" IS NOT NULL;\nEXCEPTION\n WHEN duplicate_object THEN null;\nEND $$;"
+      "\nCREATE UNIQUE INDEX IF NOT EXISTS \"documents_slug_user_id_unique\" ON \"documents\" USING btree (\"slug\",\"user_id\") WHERE \"documents\".\"slug\" is not null;\n"
     ],
     "bps": true,
-    "folderMillis": 1763535087148,
-    "hash": "57a82ce14d96ffa9f140ff63f00af994e91a74703f4d2378286e36be259f117b"
+    "folderMillis": 1763987922211,
+    "hash": "f823b521f4d25e5dc5ab238b372727d2d2d7f0aed27b5eabc8a9608ce4e50568"
   }
 ]

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

@@ -2,17 +2,18 @@
 import { eq } from 'drizzle-orm';
 import { afterEach, beforeEach, describe, expect, it } from 'vitest';
 
-import { LobeChatDatabase } from '../../type';
 import {
   agents,
   agentsFiles,
   agentsKnowledgeBases,
   agentsToSessions,
+  documents,
   files,
   knowledgeBases,
   sessions,
   users,
 } from '../../schemas';
+import { LobeChatDatabase } from '../../type';
 import { AgentModel } from '../agent';
 import { getTestDB } from './_util';
 
@@ -69,6 +70,76 @@ describe('AgentModel', () => {
       expect(result.knowledgeBases).toHaveLength(1);
       expect(result.files).toHaveLength(1);
     });
+
+    it('should fetch and include document content for enabled files', async () => {
+      const agentId = 'test-agent-with-docs';
+      await serverDB.insert(agents).values({ id: agentId, userId });
+      await serverDB.insert(agentsFiles).values({ agentId, fileId: '1', userId, enabled: true });
+      await serverDB.insert(documents).values({
+        id: 'doc1',
+        fileId: '1',
+        userId,
+        content: 'This is document content',
+        fileType: 'application/pdf',
+        totalCharCount: 100,
+        totalLineCount: 10,
+        sourceType: 'file',
+        source: 'document.pdf',
+      });
+
+      const result = await agentModel.getAgentConfigById(agentId);
+
+      expect(result).toBeDefined();
+      expect(result.files).toHaveLength(1);
+      expect(result.files[0].content).toBe('This is document content');
+      expect(result.files[0].enabled).toBe(true);
+    });
+
+    it('should not include content for disabled files', async () => {
+      const agentId = 'test-agent-disabled-file';
+      await serverDB.insert(agents).values({ id: agentId, userId });
+      await serverDB.insert(agentsFiles).values({ agentId, fileId: '1', userId, enabled: false });
+      await serverDB.insert(documents).values({
+        id: 'doc2',
+        fileId: '1',
+        userId,
+        content: 'This should not be included',
+        fileType: 'application/pdf',
+        totalCharCount: 100,
+        totalLineCount: 10,
+        sourceType: 'file',
+        source: 'document.pdf',
+      });
+
+      const result = await agentModel.getAgentConfigById(agentId);
+
+      expect(result).toBeDefined();
+      expect(result.files).toHaveLength(1);
+      expect(result.files[0].content).toBeUndefined();
+      expect(result.files[0].enabled).toBe(false);
+    });
+
+    it('should handle files without documents', async () => {
+      const agentId = 'test-agent-no-docs';
+      await serverDB.insert(agents).values({ id: agentId, userId });
+      await serverDB.insert(agentsFiles).values({ agentId, fileId: '2', userId, enabled: true });
+
+      const result = await agentModel.getAgentConfigById(agentId);
+
+      expect(result).toBeDefined();
+      expect(result.files).toHaveLength(1);
+      expect(result.files[0].content).toBeUndefined();
+    });
+
+    it('should handle agent with no files', async () => {
+      const agentId = 'test-agent-no-files';
+      await serverDB.insert(agents).values({ id: agentId, userId });
+
+      const result = await agentModel.getAgentConfigById(agentId);
+
+      expect(result).toBeDefined();
+      expect(result.files).toHaveLength(0);
+    });
   });
 
   describe('findBySessionId', () => {
@@ -84,10 +155,16 @@ describe('AgentModel', () => {
       expect(result).toBeDefined();
       expect(result?.id).toBe(agentId);
     });
+
+    it('should return undefined when session is not found', async () => {
+      const result = await agentModel.findBySessionId('non-existent-session');
+
+      expect(result).toBeUndefined();
+    });
   });
 
   describe('createAgentKnowledgeBase', () => {
-    it('should create a new agent knowledge base association', async () => {
+    it('should create a new agent knowledge base association with enabled=true by default', async () => {
       const agent = await serverDB
         .insert(agents)
         .values({ userId })
@@ -107,6 +184,27 @@ describe('AgentModel', () => {
         enabled: true,
       });
     });
+
+    it('should create a new agent knowledge base association with enabled=false', async () => {
+      const agent = await serverDB
+        .insert(agents)
+        .values({ userId })
+        .returning()
+        .then((res) => res[0]);
+
+      await agentModel.createAgentKnowledgeBase(agent.id, knowledgeBase.id, false);
+
+      const result = await serverDB.query.agentsKnowledgeBases.findFirst({
+        where: eq(agentsKnowledgeBases.agentId, agent.id),
+      });
+
+      expect(result).toMatchObject({
+        agentId: agent.id,
+        knowledgeBaseId: knowledgeBase.id,
+        userId,
+        enabled: false,
+      });
+    });
   });
 
   describe('deleteAgentKnowledgeBase', () => {
@@ -153,7 +251,7 @@ describe('AgentModel', () => {
   });
 
   describe('createAgentFiles', () => {
-    it('should create new agent file associations', async () => {
+    it('should create new agent file associations with enabled=true by default', async () => {
       const agent = await serverDB
         .insert(agents)
         .values({ userId })
@@ -174,6 +272,77 @@ describe('AgentModel', () => {
         ]),
       );
     });
+
+    it('should create new agent file associations with enabled=false', async () => {
+      const agent = await serverDB
+        .insert(agents)
+        .values({ userId })
+        .returning()
+        .then((res) => res[0]);
+
+      await agentModel.createAgentFiles(agent.id, ['1'], false);
+
+      const results = await serverDB.query.agentsFiles.findMany({
+        where: eq(agentsFiles.agentId, agent.id),
+      });
+
+      expect(results).toHaveLength(1);
+      expect(results[0]).toMatchObject({
+        agentId: agent.id,
+        fileId: '1',
+        userId,
+        enabled: false,
+      });
+    });
+
+    it('should skip files that already exist', async () => {
+      const agent = await serverDB
+        .insert(agents)
+        .values({ userId })
+        .returning()
+        .then((res) => res[0]);
+
+      // First insert
+      await serverDB.insert(agentsFiles).values({ agentId: agent.id, fileId: '1', userId });
+
+      // Try to insert the same file again
+      await agentModel.createAgentFiles(agent.id, ['1', '2']);
+
+      const results = await serverDB.query.agentsFiles.findMany({
+        where: eq(agentsFiles.agentId, agent.id),
+      });
+
+      // Should only have 2 files (1 existing + 1 new), not 3
+      expect(results).toHaveLength(2);
+      expect(results.map((r) => r.fileId).sort()).toEqual(['1', '2']);
+    });
+
+    it('should return early when all files already exist', async () => {
+      const agent = await serverDB
+        .insert(agents)
+        .values({ userId })
+        .returning()
+        .then((res) => res[0]);
+
+      // First insert
+      await serverDB.insert(agentsFiles).values([
+        { agentId: agent.id, fileId: '1', userId },
+        { agentId: agent.id, fileId: '2', userId },
+      ]);
+
+      // Try to insert the same files again
+      const result = await agentModel.createAgentFiles(agent.id, ['1', '2']);
+
+      // Should return undefined (early return)
+      expect(result).toBeUndefined();
+
+      const results = await serverDB.query.agentsFiles.findMany({
+        where: eq(agentsFiles.agentId, agent.id),
+      });
+
+      // Should still only have 2 files
+      expect(results).toHaveLength(2);
+    });
   });
 
   describe('deleteAgentFile', () => {