feat(server): search blob names from indexer (#12822)

#### PR Dependency Tree


* **PR #12822** 👈

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**
- Added the ability to search for blob names by their IDs within a
workspace.
- **Tests**
- Introduced new test cases and snapshot tests to validate searching
blob names and reading filenames from document snapshots.
<!-- end of auto-generated comment: release notes by coderabbit.ai -->

packages/backend/server/src/__tests__/mocks/doc-snapshot.mock.ts

@@ -12,6 +12,7 @@ export type MockDocSnapshotInput = {
   docId?: string;
   blob?: Uint8Array;
   updatedAt?: Date;
+  snapshotFile?: string;
 };
 
 export type MockedDocSnapshot = Snapshot;
@@ -23,7 +24,10 @@ export class MockDocSnapshot extends Mocker<
   override async create(input: MockDocSnapshotInput) {
     if (!input.blob) {
       const snapshot = await readFile(
-        path.join(import.meta.dirname, '../__fixtures__/test-doc.snapshot.bin')
+        path.join(
+          import.meta.dirname,
+          `../__fixtures__/${input.snapshotFile ?? 'test-doc.snapshot.bin'}`
+        )
       );
       input.blob = snapshot;
     }

packages/backend/server/src/core/utils/__tests__/__snapshots__/blocksute.spec.ts.md

@@ -646,6 +646,93 @@ Generated by [AVA](https://avajs.dev).
       title: 'Write, Draw, Plan all at Once.',
     }
 
+## can read blob filename from doc snapshot
+
+> Snapshot 1
+
+    {
+      blocks: [
+        {
+          additional: {
+            displayMode: 'edgeless',
+          },
+          blockId: '4YHKIhPzAK',
+          content: 'index file name',
+          docId: 'doc-0',
+          flavour: 'affine:page',
+          yblock: {
+            'prop:title': 'index file name',
+            'sys:children': [
+              'WypcCGdupE',
+              'hZ1-cdLW5e',
+            ],
+            'sys:flavour': 'affine:page',
+            'sys:id': '4YHKIhPzAK',
+            'sys:version': 2,
+          },
+        },
+        {
+          additional: {
+            displayMode: 'edgeless',
+          },
+          blockId: 'WypcCGdupE',
+          content: [],
+          docId: 'doc-0',
+          flavour: 'affine:surface',
+          parentBlockId: '4YHKIhPzAK',
+          parentFlavour: 'affine:page',
+          yblock: {
+            'prop:elements': {
+              type: '$blocksuite:internal:native$',
+              value: {},
+            },
+            'sys:children': [],
+            'sys:flavour': 'affine:surface',
+            'sys:id': 'WypcCGdupE',
+            'sys:version': 5,
+          },
+        },
+        {
+          additional: {
+            displayMode: 'page',
+            noteBlockId: 'hZ1-cdLW5e',
+          },
+          blob: [
+            'ldZMrM4PDlsNG4Q4YvCsz623h6TKu4qI9_FpTqIypfw=',
+          ],
+          blockId: 'tfz1yFZdnn',
+          content: 'test file name here.txt',
+          docId: 'doc-0',
+          flavour: 'affine:attachment',
+          parentBlockId: 'hZ1-cdLW5e',
+          parentFlavour: 'affine:note',
+          yblock: {
+            'prop:embed': false,
+            'prop:footnoteIdentifier': null,
+            'prop:index': 'a0',
+            'prop:lockedBySelf': false,
+            'prop:meta:createdAt': 1750036953927,
+            'prop:meta:createdBy': '46ce597c-098a-4c61-a106-ce79827ec1de',
+            'prop:meta:updatedAt': 1750036953928,
+            'prop:meta:updatedBy': '46ce597c-098a-4c61-a106-ce79827ec1de',
+            'prop:name': 'test file name here.txt',
+            'prop:rotate': 0,
+            'prop:size': 3,
+            'prop:sourceId': 'ldZMrM4PDlsNG4Q4YvCsz623h6TKu4qI9_FpTqIypfw=',
+            'prop:style': 'horizontalThin',
+            'prop:type': 'text/plain',
+            'prop:xywh': '[0,0,0,0]',
+            'sys:children': [],
+            'sys:flavour': 'affine:attachment',
+            'sys:id': 'tfz1yFZdnn',
+            'sys:version': 1,
+          },
+        },
+      ],
+      summary: '',
+      title: 'index file name',
+    }
+
 ## can read all blocks from doc snapshot without workspace snapshot
 
 > Snapshot 1

packages/backend/server/src/core/utils/__tests__/blocksute.spec.ts

@@ -56,6 +56,23 @@ test('can read all blocks from doc snapshot', async t => {
   });
 });
 
+test('can read blob filename from doc snapshot', async t => {
+  const docSnapshot = await module.create(Mockers.DocSnapshot, {
+    workspaceId: workspace.id,
+    user: owner,
+    snapshotFile: 'test-doc-with-blob.snapshot.bin',
+  });
+
+  const result = await readAllBlocksFromDocSnapshot(
+    workspace.id,
+    'doc-0',
+    docSnapshot.blob
+  );
+
+  // NOTE: avoid snapshot result directly, because it will cause hanging
+  t.snapshot(JSON.parse(JSON.stringify(result)));
+});
+
 test('can read all blocks from doc snapshot without workspace snapshot', async t => {
   const doc = await models.doc.get(workspace.id, docSnapshot.id);
   t.truthy(doc);

packages/backend/server/src/plugins/indexer/__tests__/__snapshots__/service.spec.ts.md

@@ -494,3 +494,30 @@ Generated by [AVA](https://avajs.dev).
         ],
       },
     ]
+
+## should search blob names from doc snapshot work
+
+> Snapshot 1
+
+    Map {
+      'ldZMrM4PDlsNG4Q4YvCsz623h6TKu4qI9_FpTqIypfw=' => 'test file name here.txt',
+    }
+
+## should search blob names work
+
+> Snapshot 1
+
+    [
+      [
+        'blob1',
+        'blob1 name.txt',
+      ],
+      [
+        'blob2',
+        'blob2 name.md',
+      ],
+      [
+        'blob3',
+        'blob3 name.docx',
+      ],
+    ]

packages/backend/server/src/plugins/indexer/__tests__/service.spec.ts

@@ -2113,3 +2113,103 @@ test('should index doc work', async t => {
   t.is(module.event.count('doc.indexer.updated'), count + 1);
 });
 // #endregion
+
+// #region searchBlobNames()
+
+test('should search blob names from doc snapshot work', async t => {
+  const docSnapshot = await module.create(Mockers.DocSnapshot, {
+    workspaceId: workspace.id,
+    user,
+    snapshotFile: 'test-doc-with-blob.snapshot.bin',
+  });
+
+  await indexerService.indexDoc(workspace.id, docSnapshot.id, {
+    refresh: true,
+  });
+
+  const blobNameMap = await indexerService.searchBlobNames(workspace.id, [
+    'ldZMrM4PDlsNG4Q4YvCsz623h6TKu4qI9_FpTqIypfw=',
+  ]);
+
+  t.snapshot(blobNameMap);
+});
+
+test('should search blob names work', async t => {
+  const workspaceId = randomUUID();
+  const blobId1 = 'blob1';
+  const blobId2 = 'blob2';
+  const blobId3 = 'blob3';
+  const blobId4 = 'blob4';
+
+  await indexerService.write(
+    SearchTable.block,
+    [
+      {
+        workspaceId,
+        blob: blobId1,
+        content: 'blob1 name.txt',
+        flavour: 'affine:attachment',
+        docId: randomUUID(),
+        blockId: randomUUID(),
+        createdByUserId: user.id,
+        updatedByUserId: user.id,
+        createdAt: new Date(),
+        updatedAt: new Date(),
+      },
+      {
+        workspaceId,
+        blob: blobId2,
+        content: 'blob2 name.md',
+        flavour: 'affine:attachment',
+        docId: randomUUID(),
+        blockId: randomUUID(),
+        createdByUserId: user.id,
+        updatedByUserId: user.id,
+        createdAt: new Date(),
+        updatedAt: new Date(),
+      },
+      {
+        workspaceId,
+        blob: blobId3,
+        content: 'blob3 name.docx',
+        flavour: 'affine:attachment',
+        docId: randomUUID(),
+        blockId: randomUUID(),
+        createdByUserId: user.id,
+        updatedByUserId: user.id,
+        createdAt: new Date(),
+        updatedAt: new Date(),
+      },
+      // no attachment
+      {
+        workspaceId,
+        blob: blobId3,
+        content: 'mock blob3 content',
+        flavour: 'affine:page',
+        docId: randomUUID(),
+        blockId: randomUUID(),
+        createdByUserId: user.id,
+        updatedByUserId: user.id,
+        createdAt: new Date(),
+        updatedAt: new Date(),
+      },
+    ],
+    {
+      refresh: true,
+    }
+  );
+
+  const blobNameMap = await indexerService.searchBlobNames(workspaceId, [
+    blobId1,
+    blobId2,
+    blobId3,
+    blobId4,
+  ]);
+
+  t.is(blobNameMap.size, 3);
+  t.snapshot(
+    Array.from(blobNameMap.entries()).sort((a, b) => a[0].localeCompare(b[0]))
+  );
+});
+
+// #endregion

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

@@ -387,6 +387,52 @@ export class IndexerService {
     await searchProvider.deleteByQuery(table, dsl, options);
   }
 
+  async searchBlobNames(workspaceId: string, blobIds: string[]) {
+    const result = await this.search({
+      table: SearchTable.block,
+      query: {
+        type: SearchQueryType.boolean,
+        occur: SearchQueryOccur.must,
+        queries: [
+          {
+            type: SearchQueryType.match,
+            field: 'workspaceId',
+            match: workspaceId,
+          },
+          {
+            type: SearchQueryType.match,
+            field: 'flavour',
+            match: 'affine:attachment',
+          },
+          {
+            type: SearchQueryType.boolean,
+            occur: SearchQueryOccur.should,
+            queries: blobIds.map(blobId => ({
+              type: SearchQueryType.match,
+              field: 'blob',
+              match: blobId,
+            })),
+          },
+        ],
+      },
+      options: {
+        fields: ['blob', 'content'],
+        pagination: {
+          limit: 10000,
+        },
+      },
+    });
+    const blobNameMap = new Map<string, string>();
+    for (const node of result.nodes) {
+      const blobId = node.fields.blob[0] as string;
+      const content = node.fields.content[0] as string;
+      if (blobId && content) {
+        blobNameMap.set(blobId, content);
+      }
+    }
+    return blobNameMap;
+  }
+
   #formatSearchNodes(nodes: SearchNode[]) {
     return nodes.map(node => ({
       ...node,