fix: 客户端响应结构损坏归入 LLMResponseError；3xx 不再误判成功

tests/test_inference_client.py

@@ -2,7 +2,12 @@ import pytest
 import httpx
 
 from genesis.inference.client import HttpLLMClient, LLMClient
-from genesis.inference.exceptions import LLMNetworkError, LLMNotConfiguredError, LLMTimeoutError
+from genesis.inference.exceptions import (
+    LLMNetworkError,
+    LLMNotConfiguredError,
+    LLMResponseError,
+    LLMTimeoutError,
+)
 from genesis.inference.types import ChatMessage
 
 
@@ -99,4 +104,36 @@ def test_chat_4xx_no_retry():
             model="m", messages=[ChatMessage(role="user", content="x")],
             temperature=0.2, max_tokens=100,
         )
-    assert calls["n"] == 1  # 4xx 不重试
+    assert calls["n"] == 1  # 4xx 不重试
+
+
+def test_chat_3xx_no_retry():
+    # 3xx（重定向，httpx 不自动跟随）不得被当作成功，应立即抛 LLMNetworkError
+    calls = {"n": 0}
+
+    def handler(request):
+        calls["n"] += 1
+        return httpx.Response(302, headers={"Location": "https://elsewhere"})
+
+    with pytest.raises(LLMNetworkError):
+        make_client(handler).chat(
+            model="m", messages=[ChatMessage(role="user", content="x")],
+            temperature=0.2, max_tokens=100,
+        )
+    assert calls["n"] == 1  # 3xx 不重试
+
+
+def test_chat_malformed_response_raises_llm_response_error():
+    # 2xx 但响应体结构损坏（非 JSON / 缺 choices/message/content）→ LLMResponseError
+    def not_json(request):
+        return httpx.Response(200, text="not-json")
+
+    def missing_key(request):
+        return httpx.Response(200, json={"choices": []})
+
+    for handler in (not_json, missing_key):
+        with pytest.raises(LLMResponseError):
+            make_client(handler).chat(
+                model="m", messages=[ChatMessage(role="user", content="x")],
+                temperature=0.2, max_tokens=100,
+            )