feat: 统一 max_input_tokens API,所有引擎必须返回最大输入 token 数
- openai/ollama/manual 三引擎均暴露 max_input_tokens 属性(统一契约) - ollama 额外上报 prompt_eval_count 到 last_prompt_tokens,使本地模型也走 token 感知压缩 - should_compress 改用 max_input_tokens,系统依当前引擎上限的 70%(compress_ratio)自动触发 - 压缩始终调用当前引擎(current_engine)做总结,符合"调用当前引擎总结压缩" - 新增 tests/test_max_input_tokens.py 覆盖契约 + 70% 触发 + 消息数兜底 Co-Authored-By: WorkBuddy <workbuddy@tencent.com>
This commit is contained in:
1 parent
c87a07706e
commit
5fe1a7c868
6 files changed
+104
-7
No files matched your search
@@ -44,6 +44,8 @@ class ManualAI:
|
||||
self.conversation_history = []
|
||||
self.max_history = 10
|
||||
self._subprocess = None
|
||||
# 最近一次请求的 prompt token 数(手动引擎无真实模型,恒为 None)
|
||||
self.last_prompt_tokens = None
|
||||
self._session_id = str(int(time.time()))
|
||||
self._started = False
|
||||
|
||||
@@ -240,6 +242,11 @@ while True:
|
||||
time.sleep(0.3)
|
||||
'''
|
||||
|
||||
@property
|
||||
def max_input_tokens(self):
|
||||
"""当前引擎支持的最大输入 token 数(统一 API)。手动引擎无真实模型,给名义上限。"""
|
||||
return 128000
|
||||
|
||||
def generate_response(self, user_input, tool_results=None, system_prompt=None):
|
||||
"""生成响应 — 转发给另一个终端的人工"""
|
||||
paths = self._get_paths()
|
||||
|
||||
Reference in new issue
Block a user