feat(api): DeepSeek V4 Flash 默认、HTTP 错讯与多供应商分层

- 主链路默认 deepseek-v4-flash，DEEPSEEK_THINKING_ENABLED 对齐旧非思考 chat
- 评测台评审装配迁入 adapters/llm（deepseek_eval_judge、zhipu_eval_judge）与 eval_judge_spec
- 拆分 llm_http_openai_chat_errors 与 llm_errors（DeepSeek/智谱品牌与文档链），llm_call 支持 http_error_vendor
- EvalJudgeService 按 spec.provider 传入 allm_json_call；评测台前端文案改为 V4 Flash
- 更新 .env 示例与 staging/production 的 DEEPSEEK_MODEL；补充 openai/供应商错讯测试

Made-with: Cursor

api/app/core/config.py

@@ -56,7 +56,10 @@ class Settings(BaseSettings):
     # ── LLM / DeepSeek ───────────────────────────────────────
     deepseek_api_key: str = ""
     deepseek_base_url: str = "https://api.deepseek.com"
-    deepseek_model: str = "deepseek-chat"
+    # 官方新模型名（V4-Flash）；与弃用名 deepseek-chat 对齐为「非思考」需另设 deepseek_thinking_enabled
+    deepseek_model: str = "deepseek-v4-flash"
+    # V4-Flash 在官方 API 中 thinking 默认为 enabled；主链路为对齐旧版 deepseek-chat 默认关闭
+    deepseek_thinking_enabled: bool = False
     llm_api_key: str = ""
     llm_base_url: str = ""
     llm_model: str = ""
@@ -399,8 +402,10 @@ class Settings(BaseSettings):
     eval_judge_base_url: str = "https://open.bigmodel.cn/api/paas/v4"
     eval_judge_model: str = "glm-5"
     eval_judge_temperature: float = 0.3
-    # 评测评审：DeepSeek（OpenAI 兼容）；默认 deepseek-reasoner 即官网 R1
-    eval_judge_deepseek_model: str = "deepseek-reasoner"
+    # 评测评审：DeepSeek（OpenAI 兼容）；默认 V4-Flash 思考模式（对齐旧版 deepseek-reasoner）
+    eval_judge_deepseek_model: str = "deepseek-v4-flash"
+    # 当仅指定 deepseek-v4-flash、未用弃用名区分时，是否走思考模式（与 eval_judge_deepseek_model 联用）
+    eval_judge_deepseek_thinking_enabled: bool = True
     eval_judge_deepseek_context_window_tokens: int = Field(
         default=64_000,
         ge=4096,