refactor: remove legacy LLM_DISABLE_THINKING and LLM_REASONING_EFFORT config, unify thinking_level handling

- Eliminate support for LLM_DISABLE_THINKING and LLM_REASONING_EFFORT in config, code, and tests
- Simplify LLM thinking level logic to rely solely on LLM_THINKING_LEVEL
- Refactor LLMHelper and related endpoints to remove legacy parameter handling
- Update system API and test utilities to match new configuration structure
- Minor code cleanup and formatting improvements
This commit is contained in:
jxxghp
2026-04-25 10:42:03 +08:00
parent 4a81417fb7
commit a05ffc07d4
6 changed files with 143 additions and 389 deletions

View File

@@ -505,12 +505,8 @@ class ConfigModel(BaseModel):
LLM_PROVIDER: str = "deepseek"
# LLM模型名称
LLM_MODEL: str = "deepseek-chat"
# 统一思考模式/深度配置off/auto/minimal/low/medium/high/max/xhigh
LLM_THINKING_LEVEL: Optional[str] = None
# 兼容旧配置:是否尽量关闭模型的思考/推理能力(新配置优先)
LLM_DISABLE_THINKING: bool = True
# 兼容旧配置:思考强度(新配置优先)
LLM_REASONING_EFFORT: Optional[str] = None
# 思考模式/深度配置off/auto/minimal/low/medium/high/max/xhigh
LLM_THINKING_LEVEL: Optional[str] = 'off'
# LLM是否支持图片输入开启后消息图片会按多模态输入发送给模型
LLM_SUPPORT_IMAGE_INPUT: bool = True
# LLM API密钥