chore(server): improve semantic search (#12838)

<!-- This is an auto-generated comment: release notes by coderabbit.ai
-->

## Summary by CodeRabbit

- **New Features**
- Updated document-related tools in chat prompts, including new options
for document keyword and semantic search.

- **Refactor**
- Renamed and reorganized document search tools for improved clarity in
tool selection and results display.

- **Bug Fixes**
- Ensured tool identifiers and results are consistent across chat and
search features.

<!-- end of auto-generated comment: release notes by coderabbit.ai -->

packages/backend/server/src/plugins/copilot/prompt/prompts.ts

@@ -1792,11 +1792,11 @@ Below is the user's query. Please respond in the user's preferred language witho
   ],
   config: {
     tools: [
-      'readDoc',
-      'editDoc',
+      'docRead',
+      'docEdit',
+      'docKeywordSearch',
+      'docSemanticSearch',
       'webSearch',
-      'keywordSearch',
-      'semanticSearch',
     ],
   },
 };

packages/backend/server/src/plugins/copilot/providers/provider.ts

@@ -13,9 +13,9 @@ import { AccessController } from '../../../core/permission';
 import { CopilotContextService } from '../context';
 import {
   buildDocSearchGetter,
+  createDocSemanticSearchTool,
   createExaCrawlTool,
   createExaSearchTool,
-  createSemanticSearchTool,
 } from '../tools';
 import { CopilotProviderFactory } from './factory';
 import {
@@ -131,20 +131,21 @@ export abstract class CopilotProvider<C = any> {
           continue;
         }
         switch (tool) {
-          case 'webSearch': {
-            tools.web_search_exa = createExaSearchTool(this.AFFiNEConfig);
-            tools.web_crawl_exa = createExaCrawlTool(this.AFFiNEConfig);
-            break;
-          }
-          case 'semanticSearch': {
+          case 'docSemanticSearch': {
             const ac = this.moduleRef.get(AccessController, { strict: false });
             const context = this.moduleRef.get(CopilotContextService, {
               strict: false,
             });
             const searchDocs = buildDocSearchGetter(ac, context);
-            tools.semantic_search = createSemanticSearchTool(
+            tools.doc_semantic_search = createDocSemanticSearchTool(
               searchDocs.bind(null, options)
             );
+            break;
+          }
+          case 'webSearch': {
+            tools.web_search_exa = createExaSearchTool(this.AFFiNEConfig);
+            tools.web_crawl_exa = createExaCrawlTool(this.AFFiNEConfig);
+            break;
           }
         }
       }

packages/backend/server/src/plugins/copilot/providers/types.ts

@@ -60,14 +60,14 @@ export const PromptConfigStrictSchema = z.object({
   tools: z
     .enum([
       // work with morph
-      'editDoc',
+      'docEdit',
+      // work with indexer
+      'docRead',
+      'docKeywordSearch',
+      // work with embeddings
+      'docSemanticSearch',
       // work with exa/model internal tools
       'webSearch',
-      // work with indexer
-      'readDoc',
-      'keywordSearch',
-      // work with embeddings
-      'semanticSearch',
     ])
     .array()
     .nullable()

packages/backend/server/src/plugins/copilot/providers/utils.ts

@@ -10,9 +10,9 @@ import {
 import { ZodType } from 'zod';
 
 import {
+  createDocSemanticSearchTool,
   createExaCrawlTool,
   createExaSearchTool,
-  createSemanticSearchTool,
 } from '../tools';
 import { PromptMessage } from './types';
 
@@ -380,9 +380,9 @@ export class CitationParser {
 }
 
 export interface CustomAITools extends ToolSet {
+  doc_semantic_search: ReturnType<typeof createDocSemanticSearchTool>;
   web_search_exa: ReturnType<typeof createExaSearchTool>;
   web_crawl_exa: ReturnType<typeof createExaCrawlTool>;
-  semantic_search: ReturnType<typeof createSemanticSearchTool>;
 }
 
 type ChunkType = TextStreamPart<CustomAITools>['type'];
@@ -429,7 +429,7 @@ export class TextStreamParser {
       case 'tool-result': {
         result = this.addPrefix(result);
         switch (chunk.toolName) {
-          case 'semantic_search': {
+          case 'doc_semantic_search': {
             if (Array.isArray(chunk.result)) {
               result += `\nFound ${chunk.result.length} document${chunk.result.length !== 1 ? 's' : ''} related to “${chunk.args.query}”.\n`;
             }

packages/backend/server/src/plugins/copilot/tools/doc-semantic-search.ts

@@ -20,12 +20,21 @@ export const buildDocSearchGetter = (
       .can('Workspace.Read');
     if (!canAccess) return undefined;
     const chunks = await context.matchWorkspaceAll(options.workspace, query);
-    return chunks || undefined;
+    const docChunks = await ac
+      .user(options.user)
+      .workspace(options.workspace)
+      .docs(
+        chunks.filter(c => 'docId' in c),
+        'Doc.Read'
+      );
+    const fileChunks = chunks.filter(c => 'fileId' in c);
+    if (!docChunks.length && !fileChunks.length) return undefined;
+    return [...fileChunks, ...docChunks];
   };
   return searchDocs;
 };
 
-export const createSemanticSearchTool = (
+export const createDocSemanticSearchTool = (
   searchDocs: (query: string) => Promise<ChunkSimilarity[] | undefined>
 ) => {
   return tool({

packages/backend/server/src/plugins/copilot/tools/index.ts

@@ -1,2 +1,2 @@
-export * from './semantic-search';
+export * from './doc-semantic-search';
 export * from './web-search';