feat(avatar): show multi-file knowledge upload progress

digital-avatar-app/backend/tests/test_embeddings.py

@@ -49,6 +49,7 @@ class RemoteEmbeddingTests(unittest.TestCase):
         texts = [f"chunk-{index}" for index in range(14)]
         batch_sizes = []
         requested_urls = []
+        progress_updates = []
 
         def fake_urlopen(request, timeout):
             self.assertEqual(timeout, 30)
@@ -68,7 +69,10 @@ class RemoteEmbeddingTests(unittest.TestCase):
             "EMBEDDING_MODEL": "text-embedding-v4",
             "EMBEDDING_BATCH_SIZE": "10",
         }), patch("embeddings.urllib.request.urlopen", side_effect=fake_urlopen):
-            result = embeddings.embed(texts)
+            result = embeddings.embed(
+                texts,
+                on_progress=lambda completed, total: progress_updates.append((completed, total)),
+            )
 
         self.assertEqual(batch_sizes, [10, 4])
         self.assertEqual(requested_urls, [
@@ -76,6 +80,7 @@ class RemoteEmbeddingTests(unittest.TestCase):
             "https://embedding.example/v1/embeddings",
         ])
         self.assertEqual(result, [[float(index)] for index in range(14)])
+        self.assertEqual(progress_updates, [(10, 14), (14, 14)])
 
     def test_full_embedding_endpoint_is_not_modified(self):
         self.assertEqual(

digital-avatar-app/backend/tests/test_knowledge_storage.py

@@ -116,6 +116,8 @@ def test_background_vectorizer_commits_ready_document_and_chunks_together(
         assert stored.status == "ready"
         assert stored.vectorized is True
         assert stored.chunk_count == 1
+        assert stored.index_stage == "ready"
+        assert stored.index_progress == 100
         assert db.query(KnowledgeChunk).filter(KnowledgeChunk.doc_id == stored.id).count() == 1
         db.query(KnowledgeChunk).filter(KnowledgeChunk.doc_id == stored.id).delete()
         db.delete(stored)