vectorvein 0.2.67__tar.gz → 0.2.68__tar.gz

{vectorvein-0.2.67 → vectorvein-0.2.68}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.1
 Name: vectorvein
-Version: 0.2.67
+Version: 0.2.68
 Summary: VectorVein Python SDK
 Author-Email: Anderson <andersonby@163.com>
 License: MIT

{vectorvein-0.2.67 → vectorvein-0.2.68}/pyproject.toml

@@ -17,7 +17,7 @@ description = "VectorVein Python SDK"
 name = "vectorvein"
 readme = "README.md"
 requires-python = ">=3.10"
-version = "0.2.67"
+version = "0.2.68"
 
 [project.license]
 text = "MIT"

{vectorvein-0.2.67 → vectorvein-0.2.68}/src/vectorvein/chat_clients/openai_compatible_client.py

@@ -117,7 +117,7 @@ class OpenAICompatibleChatClient(BaseChatClient):
         model: str | None = None,
         stream: Literal[False] = False,
         temperature: float | None | NotGiven = NOT_GIVEN,
-        max_tokens: int | None = None,
+        max_tokens: int | None | NotGiven = NOT_GIVEN,
         tools: Iterable[ToolParam] | NotGiven = NOT_GIVEN,
         tool_choice: ToolChoice | NotGiven = NOT_GIVEN,
         response_format: ResponseFormat | NotGiven = NOT_GIVEN,
@@ -158,7 +158,7 @@ class OpenAICompatibleChatClient(BaseChatClient):
         model: str | None = None,
         stream: Literal[True],
         temperature: float | None | NotGiven = NOT_GIVEN,
-        max_tokens: int | None = None,
+        max_tokens: int | None | NotGiven = NOT_GIVEN,
         tools: Iterable[ToolParam] | NotGiven = NOT_GIVEN,
         tool_choice: ToolChoice | NotGiven = NOT_GIVEN,
         response_format: ResponseFormat | NotGiven = NOT_GIVEN,
@@ -199,7 +199,7 @@ class OpenAICompatibleChatClient(BaseChatClient):
         model: str | None = None,
         stream: bool,
         temperature: float | None | NotGiven = NOT_GIVEN,
-        max_tokens: int | None = None,
+        max_tokens: int | None | NotGiven = NOT_GIVEN,
         tools: Iterable[ToolParam] | NotGiven = NOT_GIVEN,
         tool_choice: ToolChoice | NotGiven = NOT_GIVEN,
         response_format: ResponseFormat | NotGiven = NOT_GIVEN,
@@ -239,7 +239,7 @@ class OpenAICompatibleChatClient(BaseChatClient):
         model: str | None = None,
         stream: Literal[False] | Literal[True] = False,
         temperature: float | None | NotGiven = NOT_GIVEN,
-        max_tokens: int | None = None,
+        max_tokens: int | None | NotGiven = NOT_GIVEN,
         tools: Iterable[ToolParam] | NotGiven = NOT_GIVEN,
         tool_choice: ToolChoice | NotGiven = NOT_GIVEN,
         response_format: ResponseFormat | NotGiven = NOT_GIVEN,
@@ -280,6 +280,8 @@ class OpenAICompatibleChatClient(BaseChatClient):
             self.temperature = temperature
         if isinstance(top_p, AnthropicNotGiven):
             top_p = NOT_GIVEN
+        if isinstance(max_tokens, AnthropicNotGiven):
+            max_tokens = NOT_GIVEN
 
         raw_client = self.raw_client  # 调用完 self.raw_client 后，self.model_id 会被赋值
         self.model_setting = self.backend_settings.models[self.model]
@@ -322,7 +324,7 @@ class OpenAICompatibleChatClient(BaseChatClient):
         else:
             tools_params = {}
 
-        if max_tokens is None:
+        if not max_tokens and not max_completion_tokens:
             max_output_tokens = self.model_setting.max_output_tokens
             token_counts = get_message_token_counts(messages=messages, tools=tools, model=self.model)
             if max_output_tokens is not None:
@@ -331,6 +333,10 @@ class OpenAICompatibleChatClient(BaseChatClient):
             else:
                 max_tokens = self.model_setting.context_length - token_counts - 64
 
+        if "o3-mini" in self.model_id:
+            max_completion_tokens = max_tokens
+            max_tokens = NOT_GIVEN
+
         self._acquire_rate_limit(self.endpoint, self.model, messages)
 
         if self.stream:
@@ -587,7 +593,7 @@ class AsyncOpenAICompatibleChatClient(BaseAsyncChatClient):
         model: str | None = None,
         stream: Literal[False] = False,
         temperature: float | None | NotGiven = NOT_GIVEN,
-        max_tokens: int | None = None,
+        max_tokens: int | None | NotGiven = NOT_GIVEN,
         tools: Iterable[ToolParam] | NotGiven = NOT_GIVEN,
         tool_choice: ToolChoice | NotGiven = NOT_GIVEN,
         response_format: ResponseFormat | NotGiven = NOT_GIVEN,
@@ -628,7 +634,7 @@ class AsyncOpenAICompatibleChatClient(BaseAsyncChatClient):
         model: str | None = None,
         stream: Literal[True],
         temperature: float | None | NotGiven = NOT_GIVEN,
-        max_tokens: int | None = None,
+        max_tokens: int | None | NotGiven = NOT_GIVEN,
         tools: Iterable[ToolParam] | NotGiven = NOT_GIVEN,
         tool_choice: ToolChoice | NotGiven = NOT_GIVEN,
         response_format: ResponseFormat | NotGiven = NOT_GIVEN,
@@ -669,7 +675,7 @@ class AsyncOpenAICompatibleChatClient(BaseAsyncChatClient):
         model: str | None = None,
         stream: bool,
         temperature: float | None | NotGiven = NOT_GIVEN,
-        max_tokens: int | None = None,
+        max_tokens: int | None | NotGiven = NOT_GIVEN,
         tools: Iterable[ToolParam] | NotGiven = NOT_GIVEN,
         tool_choice: ToolChoice | NotGiven = NOT_GIVEN,
         response_format: ResponseFormat | NotGiven = NOT_GIVEN,
@@ -709,7 +715,7 @@ class AsyncOpenAICompatibleChatClient(BaseAsyncChatClient):
         model: str | None = None,
         stream: Literal[False] | Literal[True] = False,
         temperature: float | None | NotGiven = NOT_GIVEN,
-        max_tokens: int | None = None,
+        max_tokens: int | None | NotGiven = NOT_GIVEN,
         tools: Iterable[ToolParam] | NotGiven = NOT_GIVEN,
         tool_choice: ToolChoice | NotGiven = NOT_GIVEN,
         response_format: ResponseFormat | NotGiven = NOT_GIVEN,
@@ -750,6 +756,8 @@ class AsyncOpenAICompatibleChatClient(BaseAsyncChatClient):
             self.temperature = temperature
         if isinstance(top_p, AnthropicNotGiven):
             top_p = NOT_GIVEN
+        if isinstance(max_tokens, AnthropicNotGiven):
+            max_tokens = NOT_GIVEN
 
         raw_client = self.raw_client  # 调用完 self.raw_client 后，self.model_id 会被赋值
         self.model_setting = self.backend_settings.models[self.model]
@@ -792,7 +800,7 @@ class AsyncOpenAICompatibleChatClient(BaseAsyncChatClient):
         else:
             tools_params = {}
 
-        if max_tokens is None:
+        if not max_tokens and not max_completion_tokens:
             max_output_tokens = self.model_setting.max_output_tokens
             token_counts = get_message_token_counts(messages=messages, tools=tools, model=self.model)
             if max_output_tokens is not None:
@@ -801,6 +809,10 @@ class AsyncOpenAICompatibleChatClient(BaseAsyncChatClient):
             else:
                 max_tokens = self.model_setting.context_length - token_counts - 64
 
+        if "o3-mini" in self.model_id:
+            max_completion_tokens = max_tokens
+            max_tokens = NOT_GIVEN
+
         await self._acquire_rate_limit(self.endpoint, self.model, messages)
 
         if self.stream: