feat: preserve governed OpenAI tool messages in 1.4.0

Add faithful single-choice chat results, tool stream activity, error-frame retries, and isolated protocol caching.

Reject empty cached completions with the same value predicate as transport, add regression coverage, and adopt Apache-2.0.

tests/integration/test_redis_cache.py

@@ -172,3 +172,22 @@ class TestClientWithRealRedis:
         second = await client.chat(_MSGS)
         assert first.cache_hit is False and second.cache_hit is True
         assert second.content == "hello"
+
+
+async def test_real_completion_roundtrip(redis_cache, namespace):
+    """真实 Redis 保存完整嵌套协议对象，命中仍为值类型。"""
+    from dataclasses import replace
+
+    from tests.unit.test_openai_completion import accumulate
+
+    completion = accumulate().finish()
+    mw = CacheMW(backend=redis_cache, model_fingerprint="m", default_namespace=namespace, ttl_s=60)
+
+    async def terminal(_):
+        return replace(_resp(), completion=completion)
+
+    request = ChatRequest(messages=_MSGS, response_mode="openai")
+    first = await mw(request, terminal)
+    second = await mw(request, terminal)
+    assert second.cache_hit and second.completion == first.completion
+    assert second.completion is not first.completion

tests/integration/test_redis_cross_connection.py

@@ -86,6 +86,7 @@ class ScriptedTransport:
         reasoning_effort,
         first_token_event,
         content_event,
+        response_mode,
     ):
         self.calls.append(source.name)
         if self.hang: