feat(server): add search docs by keyword gql api (#12866)

close AI-220









#### PR Dependency Tree


* **PR #12866** 👈

This tree was auto-generated by
[Charcoal](https://github.com/danerwilliams/charcoal)

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

## Summary by CodeRabbit

- **New Features**
- Introduced a new document search capability, allowing users to search
for documents by keyword within a workspace.
- Search results include document details such as title, highlights,
creation and update timestamps, and creator/updater information.
  - Added support for limiting the number of search results returned.
- **Tests**
- Added comprehensive end-to-end and snapshot tests to ensure accuracy
and access control for the new search functionality.

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

packages/backend/server/src/__tests__/e2e/indexer/__snapshots__/search-docs.spec.ts.md

@@ -0,0 +1,51 @@
+# Snapshot report for `src/__tests__/e2e/indexer/search-docs.spec.ts`
+
+The actual snapshot is saved in `search-docs.spec.ts.snap`.
+
+Generated by [AVA](https://avajs.dev).
+
+## should search docs by keyword
+
+> Snapshot 1
+
+    [
+      {
+        blockId: 'block-0',
+        createdAt: '2025-04-22T00:00:00.000Z',
+        docId: 'doc-0',
+        highlight: 'test1 <b>hello</b>',
+        title: '',
+        updatedAt: '2025-04-22T00:00:00.000Z',
+      },
+      {
+        blockId: 'block-2',
+        createdAt: '2025-03-22T00:00:00.000Z',
+        docId: 'doc-2',
+        highlight: 'test3 <b>hello</b>',
+        title: '',
+        updatedAt: '2025-03-22T03:00:01.000Z',
+      },
+      {
+        blockId: 'block-1',
+        createdAt: '2021-04-22T00:00:00.000Z',
+        docId: 'doc-1',
+        highlight: 'test2 <b>hello</b>',
+        title: '',
+        updatedAt: '2021-04-22T00:00:00.000Z',
+      },
+    ]
+
+## should search docs by keyword with limit 1
+
+> Snapshot 1
+
+    [
+      {
+        blockId: 'block-0',
+        createdAt: '2025-04-22T00:00:00.000Z',
+        docId: 'doc-0',
+        highlight: 'test1 <b>hello</b>',
+        title: '',
+        updatedAt: '2025-04-22T00:00:00.000Z',
+      },
+    ]

packages/backend/server/src/__tests__/e2e/indexer/search-docs.spec.ts

@@ -0,0 +1,182 @@
+import { indexerSearchDocsQuery, SearchTable } from '@affine/graphql';
+import { omit } from 'lodash-es';
+
+import { IndexerService } from '../../../plugins/indexer/service';
+import { Mockers } from '../../mocks';
+import { app, e2e } from '../test';
+
+e2e('should search docs by keyword', async t => {
+  const owner = await app.signup();
+
+  const workspace = await app.create(Mockers.Workspace, {
+    owner,
+  });
+
+  const indexerService = app.get(IndexerService);
+
+  await indexerService.write(
+    SearchTable.block,
+    [
+      {
+        docId: 'doc-0',
+        workspaceId: workspace.id,
+        content: 'test1 hello',
+        flavour: 'markdown',
+        blockId: 'block-0',
+        createdByUserId: owner.id,
+        updatedByUserId: owner.id,
+        createdAt: new Date('2025-04-22T00:00:00.000Z'),
+        updatedAt: new Date('2025-04-22T00:00:00.000Z'),
+      },
+      {
+        docId: 'doc-1',
+        workspaceId: workspace.id,
+        content: 'test2 hello',
+        flavour: 'markdown',
+        blockId: 'block-1',
+        refDocId: ['doc-0'],
+        ref: ['{"foo": "bar1"}'],
+        createdByUserId: owner.id,
+        updatedByUserId: owner.id,
+        createdAt: new Date('2021-04-22T00:00:00.000Z'),
+        updatedAt: new Date('2021-04-22T00:00:00.000Z'),
+      },
+      {
+        docId: 'doc-2',
+        workspaceId: workspace.id,
+        content: 'test3 hello',
+        flavour: 'markdown',
+        blockId: 'block-2',
+        refDocId: ['doc-0', 'doc-2'],
+        ref: ['{"foo": "bar1"}', '{"foo": "bar3"}'],
+        createdByUserId: owner.id,
+        updatedByUserId: owner.id,
+        createdAt: new Date('2025-03-22T00:00:00.000Z'),
+        updatedAt: new Date('2025-03-22T03:00:01.000Z'),
+      },
+    ],
+    {
+      refresh: true,
+    }
+  );
+
+  const result = await app.gql({
+    query: indexerSearchDocsQuery,
+    variables: {
+      id: workspace.id,
+      input: {
+        keyword: 'hello',
+      },
+    },
+  });
+
+  t.is(result.workspace.searchDocs.length, 3);
+  t.snapshot(
+    result.workspace.searchDocs.map(doc =>
+      omit(doc, 'createdByUser', 'updatedByUser')
+    )
+  );
+});
+
+e2e('should search docs by keyword with limit 1', async t => {
+  const owner = await app.signup();
+
+  const workspace = await app.create(Mockers.Workspace, {
+    owner,
+  });
+
+  const indexerService = app.get(IndexerService);
+
+  await indexerService.write(
+    SearchTable.block,
+    [
+      {
+        docId: 'doc-0',
+        workspaceId: workspace.id,
+        content: 'test1 hello',
+        flavour: 'markdown',
+        blockId: 'block-0',
+        createdByUserId: owner.id,
+        updatedByUserId: owner.id,
+        createdAt: new Date('2025-04-22T00:00:00.000Z'),
+        updatedAt: new Date('2025-04-22T00:00:00.000Z'),
+      },
+      {
+        docId: 'doc-1',
+        workspaceId: workspace.id,
+        content: 'test2 hello',
+        flavour: 'markdown',
+        blockId: 'block-1',
+        refDocId: ['doc-0'],
+        ref: ['{"foo": "bar1"}'],
+        createdByUserId: owner.id,
+        updatedByUserId: owner.id,
+        createdAt: new Date('2021-04-22T00:00:00.000Z'),
+        updatedAt: new Date('2021-04-22T00:00:00.000Z'),
+      },
+      {
+        docId: 'doc-2',
+        workspaceId: workspace.id,
+        content: 'test3 hello',
+        flavour: 'markdown',
+        blockId: 'block-2',
+        refDocId: ['doc-0', 'doc-2'],
+        ref: ['{"foo": "bar1"}', '{"foo": "bar3"}'],
+        createdByUserId: owner.id,
+        updatedByUserId: owner.id,
+        createdAt: new Date('2025-03-22T00:00:00.000Z'),
+        updatedAt: new Date('2025-03-22T03:00:01.000Z'),
+      },
+    ],
+    {
+      refresh: true,
+    }
+  );
+
+  const result = await app.gql({
+    query: indexerSearchDocsQuery,
+    variables: {
+      id: workspace.id,
+      input: {
+        keyword: 'hello',
+        limit: 1,
+      },
+    },
+  });
+
+  t.is(result.workspace.searchDocs.length, 1);
+  t.snapshot(
+    result.workspace.searchDocs.map(doc =>
+      omit(doc, 'createdByUser', 'updatedByUser')
+    )
+  );
+});
+
+e2e(
+  'should search docs by keyword failed when workspace is no permission',
+  async t => {
+    const owner = await app.signup();
+
+    const workspace = await app.create(Mockers.Workspace, {
+      owner,
+    });
+
+    // signup another user
+    await app.signup();
+
+    await t.throwsAsync(
+      app.gql({
+        query: indexerSearchDocsQuery,
+        variables: {
+          id: workspace.id,
+          input: {
+            keyword: 'hello',
+          },
+        },
+      }),
+      {
+        message: /You do not have permission to access Space/,
+      }
+    );
+  }
+);

packages/backend/server/src/plugins/indexer/resolver.ts

@@ -10,6 +10,8 @@ import { IndexerService, SearchNodeWithMeta } from './service';
 import {
   AggregateInput,
   AggregateResultObjectType,
+  SearchDocObjectType,
+  SearchDocsInput,
   SearchInput,
   SearchQueryOccur,
   SearchQueryType,
@@ -86,6 +88,29 @@ export class IndexerResolver {
     };
   }
 
+  @ResolveField(() => [SearchDocObjectType], {
+    description: 'Search docs by keyword',
+  })
+  async searchDocs(
+    @CurrentUser() me: UserType,
+    @Parent() workspace: WorkspaceType,
+    @Args('input') input: SearchDocsInput
+  ): Promise<SearchDocObjectType[]> {
+    const docs = await this.indexer.searchDocsByKeyword(
+      workspace.id,
+      input.keyword,
+      {
+        limit: input.limit,
+      }
+    );
+
+    const needs = await this.ac
+      .user(me.id)
+      .workspace(workspace.id)
+      .docs(docs, 'Doc.Read');
+    return needs;
+  }
+
   #addWorkspaceFilter(
     workspace: WorkspaceType,
     input: SearchInput | AggregateInput

packages/backend/server/src/plugins/indexer/types.ts

@@ -9,6 +9,7 @@ import {
 } from '@nestjs/graphql';
 import { GraphQLJSONObject } from 'graphql-scalars';
 
+import { PublicUserType } from '../../core/user';
 import { PublicUser } from '../../models';
 import { SearchTable } from './tables';
 
@@ -171,6 +172,18 @@ export class AggregateInput {
   options!: AggregateOptions;
 }
 
+@InputType()
+export class SearchDocsInput {
+  @Field(() => String)
+  keyword!: string;
+
+  @Field({
+    nullable: true,
+    description: 'Limit the number of docs to return, default is 20',
+  })
+  limit?: number;
+}
+
 @ObjectType()
 export class BlockObjectType {
   @Field(() => [String], { nullable: true })
@@ -320,3 +333,30 @@ export class AggregateResultObjectType {
   @Field(() => SearchResultPagination)
   pagination!: SearchResultPagination;
 }
+
+@ObjectType()
+export class SearchDocObjectType implements Partial<SearchDoc> {
+  @Field(() => String)
+  docId!: string;
+
+  @Field(() => String)
+  title!: string;
+
+  @Field(() => String)
+  blockId!: string;
+
+  @Field(() => String)
+  highlight!: string;
+
+  @Field(() => Date)
+  createdAt!: Date;
+
+  @Field(() => Date)
+  updatedAt!: Date;
+
+  @Field(() => PublicUserType, { nullable: true })
+  createdByUser?: PublicUserType;
+
+  @Field(() => PublicUserType, { nullable: true })
+  updatedByUser?: PublicUserType;
+}

packages/backend/server/src/schema.gql

@@ -1502,6 +1502,24 @@ type SameSubscriptionRecurringDataType {
   recurring: String!
 }
 
+type SearchDocObjectType {
+  blockId: String!
+  createdAt: DateTime!
+  createdByUser: PublicUserType
+  docId: String!
+  highlight: String!
+  title: String!
+  updatedAt: DateTime!
+  updatedByUser: PublicUserType
+}
+
+input SearchDocsInput {
+  keyword: String!
+
+  """Limit the number of docs to return, default is 20"""
+  limit: Int
+}
+
 input SearchHighlight {
   before: String!
   end: String!
@@ -2074,6 +2092,9 @@ type WorkspaceType {
   """Search a specific table"""
   search(input: SearchInput!): SearchResultObjectType!
 
+  """Search docs by keyword"""
+  searchDocs(input: SearchDocsInput!): [SearchDocObjectType!]!
+
   """The team subscription of the workspace, if exists."""
   subscription: SubscriptionType