refactor: 重构 inference 模块架构，引入设计模式并分组文件

- 新增 protocol.py 协议层，Template Method 模式消除流/非流分支 45% 重复
- SSEBuilder 统一 SSE 构造，StopChecker 独立 stop_sequence 检测
- AnthropicHandler 追踪已产出文本，修复 stop 时重复 delta
- server.py 路由从约 100 行缩减至 3 行
- 拆分为 core/（cache/executor/scheduler/task）和 api/（protocol/server）
- 外部保持二级导入路径（from astrai.inference import Name）
- 删除所有分隔线注释，代码按语义自然分组

tests/inference/conftest.py

@@ -5,7 +5,7 @@ from unittest.mock import MagicMock
 import pytest
 from fastapi.testclient import TestClient
 
-from astrai.inference.server import app
+from astrai.inference import app
 
 
 @pytest.fixture

tests/inference/test_cache.py

@@ -2,7 +2,7 @@
 
 import torch
 
-from astrai.inference.cache import (
+from astrai.inference import (
     PagedCache,
     PagePool,
     PrefixCache,

tests/inference/test_engine.py

@@ -3,8 +3,8 @@
 import threading
 from unittest.mock import MagicMock, patch
 
+from astrai.inference import STOP
 from astrai.inference.engine import GenerateResult
-from astrai.inference.task import STOP
 
 
 def test_result_append_single():

tests/inference/test_scheduler.py

@@ -6,7 +6,7 @@ from unittest.mock import MagicMock, patch
 import pytest
 import torch
 
-from astrai.inference.scheduler import InferenceScheduler
+from astrai.inference import InferenceScheduler
 
 
 @pytest.fixture
@@ -36,8 +36,8 @@ def test_scheduler_concurrent_add_task(mock_model_and_tokenizer):
     """Test concurrent add_task operations."""
     mock_model, mock_tokenizer = mock_model_and_tokenizer
 
-    with patch("astrai.inference.scheduler.AutoModel"):
-        with patch("astrai.inference.scheduler.AutoTokenizer"):
+    with patch("astrai.inference.core.scheduler.AutoModel"):
+        with patch("astrai.inference.core.scheduler.AutoTokenizer"):
             scheduler = InferenceScheduler(
                 model=mock_model,
                 tokenizer=mock_tokenizer,
@@ -75,8 +75,8 @@ def test_scheduler_concurrent_add_remove_task(mock_model_and_tokenizer):
     """Test concurrent add and remove task operations."""
     mock_model, mock_tokenizer = mock_model_and_tokenizer
 
-    with patch("astrai.inference.scheduler.AutoModel"):
-        with patch("astrai.inference.scheduler.AutoTokenizer"):
+    with patch("astrai.inference.core.scheduler.AutoModel"):
+        with patch("astrai.inference.core.scheduler.AutoTokenizer"):
             scheduler = InferenceScheduler(
                 model=mock_model,
                 tokenizer=mock_tokenizer,
@@ -124,8 +124,8 @@ def test_scheduler_concurrent_get_stats(mock_model_and_tokenizer):
     """Test concurrent get_stats operations."""
     mock_model, mock_tokenizer = mock_model_and_tokenizer
 
-    with patch("astrai.inference.scheduler.AutoModel"):
-        with patch("astrai.inference.scheduler.AutoTokenizer"):
+    with patch("astrai.inference.core.scheduler.AutoModel"):
+        with patch("astrai.inference.core.scheduler.AutoTokenizer"):
             scheduler = InferenceScheduler(
                 model=mock_model,
                 tokenizer=mock_tokenizer,

tests/inference/test_server.py

@@ -2,7 +2,7 @@
 
 import pytest
 
-from astrai.inference.server import app
+from astrai.inference import app
 
 
 def test_health_no_model(client):

tests/inference/test_task.py

@@ -2,7 +2,7 @@
 
 from unittest.mock import MagicMock
 
-from astrai.inference.task import STOP, Task, TaskManager, TaskStatus
+from astrai.inference import STOP, Task, TaskManager, TaskStatus
 
 
 def _make_mock_tokenizer():