feat: add voice studio prototype flow

backend/tests/test_voice_sessions.py

@@ -199,3 +199,147 @@ async def test_voice_session_abandon_blocks_future_turns(
             assert response.status_code == 409
     finally:
         app.dependency_overrides.clear()
+
+
+async def test_voice_session_uploaded_audio_turn_uses_demo_transcript_hint(
+    db_session,
+    auth_token,
+):
+    async def override_get_db():
+        yield db_session
+
+    app.dependency_overrides[get_db] = override_get_db
+
+    with (
+        patch(
+            "app.services.voice_session_service.generate_story_content",
+            new_callable=AsyncMock,
+        ) as mock_generate,
+        patch(
+            "app.services.voice_session_service.text_to_speech",
+            new_callable=AsyncMock,
+        ) as mock_tts,
+    ):
+        mock_generate.return_value = StoryOutput(
+            mode="generated",
+            title="小鲸鱼找朋友",
+            story_text="小鲸鱼在海面上遇见了一只会唱歌的海鸥。",
+            cover_prompt_suggestion="温暖儿童绘本封面，小鲸鱼和海鸥",
+        )
+        mock_tts.return_value = b"fake-upload-audio"
+
+        transport = ASGITransport(app=app)
+        try:
+            async with AsyncClient(transport=transport, base_url="http://test") as client:
+                client.cookies.set("access_token", auth_token)
+
+                response = await client.post("/api/voice-sessions", json={})
+                assert response.status_code == 201
+                session_id = response.json()["id"]
+
+                response = await client.post(
+                    f"/api/voice-sessions/{session_id}/turns",
+                    files={
+                        "audio_file": ("turn.webm", b"fake-webm-audio", "audio/webm"),
+                    },
+                    data={
+                        "duration_ms": "3200",
+                        "transcript_hint": "我想听一个小鲸鱼找朋友的故事",
+                    },
+                )
+                assert response.status_code == 202
+                turn_data = response.json()
+                assert turn_data["status"] == "audio_ready"
+                assert turn_data["transcription_provider"] == "demo"
+
+                turn_id = turn_data["turn_id"]
+                response = await client.get(
+                    f"/api/voice-sessions/{session_id}/turns/{turn_id}"
+                )
+                assert response.status_code == 200
+                detail = response.json()
+                assert detail["user_audio_ready"] is True
+                assert detail["user_audio_url"].endswith("/user-audio")
+                assert detail["transcription_provider"] == "demo"
+                assert detail["assistant_audio_ready"] is True
+
+                response = await client.get(detail["user_audio_url"])
+                assert response.status_code == 200
+                assert response.content == b"fake-webm-audio"
+                assert response.headers["content-type"] == "audio/webm"
+        finally:
+            app.dependency_overrides.clear()
+
+
+async def test_voice_session_list_orders_recent_sessions_first(
+    db_session,
+    auth_token,
+):
+    async def override_get_db():
+        yield db_session
+
+    app.dependency_overrides[get_db] = override_get_db
+
+    with (
+        patch(
+            "app.services.voice_session_service.generate_story_content",
+            new_callable=AsyncMock,
+        ) as mock_generate,
+        patch(
+            "app.services.voice_session_service.text_to_speech",
+            new_callable=AsyncMock,
+        ) as mock_tts,
+    ):
+        mock_generate.side_effect = [
+            StoryOutput(
+                mode="generated",
+                title="第一场冒险",
+                story_text="第一段故事。",
+                cover_prompt_suggestion="封面一",
+            ),
+            StoryOutput(
+                mode="generated",
+                title="第二场冒险",
+                story_text="第二段故事。",
+                cover_prompt_suggestion="封面二",
+            ),
+        ]
+        mock_tts.side_effect = [b"audio-1", b"audio-2"]
+
+        transport = ASGITransport(app=app)
+        try:
+            async with AsyncClient(transport=transport, base_url="http://test") as client:
+                client.cookies.set("access_token", auth_token)
+
+                response = await client.post("/api/voice-sessions", json={})
+                first_session_id = response.json()["id"]
+                await client.post(
+                    f"/api/voice-sessions/{first_session_id}/turns/fallback",
+                    json={"transcript_text": "第一个故事"},
+                )
+
+                response = await client.post("/api/voice-sessions", json={})
+                second_session_id = response.json()["id"]
+                await client.post(
+                    f"/api/voice-sessions/{second_session_id}/turns/fallback",
+                    json={"transcript_text": "第二个故事"},
+                )
+
+                response = await client.get("/api/voice-sessions?limit=8")
+                assert response.status_code == 200
+                sessions = response.json()
+                assert len(sessions) >= 2
+                assert sessions[0]["id"] == second_session_id
+                assert sessions[1]["id"] == first_session_id
+                assert sessions[0]["total_turns"] == 1
+                assert sessions[0]["last_turn_status"] == "audio_ready"
+
+                response = await client.get("/api/voice-sessions?active_only=true")
+                assert response.status_code == 200
+                active_sessions = response.json()
+                assert {item["id"] for item in active_sessions} >= {
+                    first_session_id,
+                    second_session_id,
+                }
+        finally:
+            app.dependency_overrides.clear()