fix(server): frequent embedding (#13475)

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

- New Features
- Smarter embedding pipeline skips re-embedding when content hasn’t
changed; added content sanitization for embeddings and workspace content
retrieval.
- Bug Fixes
- Re-embedding now requires both a document update and the last
embedding being older than 10 minutes, reducing unnecessary work.
- Refactor
- Consolidated embedding preprocessing and moved sanitization utilities
into shared models; upserts now refresh stored content.
- Tests
- Expanded snapshot-based tests covering multiple time/age scenarios for
embedding decision logic.
<!-- end of auto-generated comment: release notes by coderabbit.ai -->

packages/backend/server/src/__tests__/models/__snapshots__/copilot-workspace.spec.ts.md

@@ -101,6 +101,28 @@ Generated by [AVA](https://avajs.dev).
 
     0
 
+## should check need to be embedded
+
+> document with no embedding should need embedding
+
+    true
+
+> document with recent embedding should not need embedding
+
+    false
+
+> document updated after embedding and older-than-10m should need embedding
+
+    true
+
+> should not need embedding when only 10-minute window passed without updates
+
+    false
+
+> should need embedding when doc updated and last embedding older than 10 minutes
+
+    true
+
 ## should filter outdated doc id style in embedding status
 
 > should include modern doc format

packages/backend/server/src/__tests__/models/copilot-workspace.spec.ts

@@ -293,7 +293,10 @@ test('should check need to be embedded', async t => {
       workspace.id,
       docId
     );
-    t.true(needsEmbedding, 'document with no embedding should need embedding');
+    t.snapshot(
+      needsEmbedding,
+      'document with no embedding should need embedding'
+    );
   }
 
   {
@@ -313,7 +316,7 @@ test('should check need to be embedded', async t => {
       workspace.id,
       docId
     );
-    t.false(
+    t.snapshot(
       needsEmbedding,
       'document with recent embedding should not need embedding'
     );
@@ -328,15 +331,83 @@ test('should check need to be embedded', async t => {
       editorId: user.id,
     });
 
+    // simulate an old embedding
+    const oldEmbeddingTime = new Date(Date.now() - 25 * 60 * 1000);
+    await t.context.db.aiWorkspaceEmbedding.updateMany({
+      where: { workspaceId: workspace.id, docId },
+      data: { updatedAt: oldEmbeddingTime },
+    });
+
     let needsEmbedding = await t.context.copilotWorkspace.checkDocNeedEmbedded(
       workspace.id,
       docId
     );
-    t.true(
+    t.snapshot(
       needsEmbedding,
-      'document updated after embedding should need embedding'
+      'document updated after embedding and older-than-10m should need embedding'
     );
   }
+
+  {
+    // only time passed (>10m since last embedding) but no doc updates => should NOT re-embed
+    const baseNow = Date.now();
+    const docId2 = randomUUID();
+    const t0 = baseNow - 30 * 60 * 1000; // snapshot updated 30 minutes ago
+    const t1 = baseNow - 25 * 60 * 1000; // embedding updated 25 minutes ago
+
+    await t.context.doc.upsert({
+      spaceId: workspace.id,
+      docId: docId2,
+      blob: Uint8Array.from([1, 2, 3]),
+      timestamp: t0,
+      editorId: user.id,
+    });
+
+    await t.context.copilotContext.insertWorkspaceEmbedding(
+      workspace.id,
+      docId2,
+      [
+        {
+          index: 0,
+          content: 'content2',
+          embedding: Array.from({ length: 1024 }, () => 1),
+        },
+      ]
+    );
+
+    await t.context.db.aiWorkspaceEmbedding.updateMany({
+      where: { workspaceId: workspace.id, docId: docId2 },
+      data: { updatedAt: new Date(t1) },
+    });
+
+    let needsEmbedding = await t.context.copilotWorkspace.checkDocNeedEmbedded(
+      workspace.id,
+      docId2
+    );
+    t.snapshot(
+      needsEmbedding,
+      'should not need embedding when only 10-minute window passed without updates'
+    );
+
+    const t2 = baseNow - 5 * 60 * 1000; // doc updated 5 minutes ago
+    await t.context.doc.upsert({
+      spaceId: workspace.id,
+      docId: docId2,
+      blob: Uint8Array.from([7, 8, 9]),
+      timestamp: t2,
+      editorId: user.id,
+    });
+
+    needsEmbedding = await t.context.copilotWorkspace.checkDocNeedEmbedded(
+      workspace.id,
+      docId2
+    );
+    t.snapshot(
+      needsEmbedding,
+      'should need embedding when doc updated and last embedding older than 10 minutes'
+    );
+  }
+  // --- new cases end ---
 });
 
 test('should check embedding table', async t => {