vectorvein 0.1.56__tar.gz → 0.1.57__tar.gz

{vectorvein-0.1.56 → vectorvein-0.1.57}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.1
 Name: vectorvein
-Version: 0.1.56
+Version: 0.1.57
 Summary: VectorVein python SDK
 Author-Email: Anderson <andersonby@163.com>
 License: MIT

{vectorvein-0.1.56 → vectorvein-0.1.57}/pyproject.toml

@@ -17,7 +17,7 @@ description = "VectorVein python SDK"
 name = "vectorvein"
 readme = "README.md"
 requires-python = ">=3.10"
-version = "0.1.56"
+version = "0.1.57"
 
 [project.license]
 text = "MIT"

{vectorvein-0.1.56 → vectorvein-0.1.57}/src/vectorvein/chat_clients/anthropic_client.py

@@ -249,6 +249,7 @@ class AnthropicChatClient(BaseChatClient):
         tool_choice: ToolChoice | NotGiven = NOT_GIVEN,
         response_format: dict | None = None,
         stream_options: ChatCompletionStreamOptionsParam | None = None,
+        top_p: float | NotGiven | None = NOT_GIVEN,
         **kwargs,
     ) -> ChatCompletionMessage:
         pass
@@ -265,6 +266,7 @@ class AnthropicChatClient(BaseChatClient):
         tool_choice: ToolChoice | NotGiven = NOT_GIVEN,
         response_format: dict | None = None,
         stream_options: ChatCompletionStreamOptionsParam | None = None,
+        top_p: float | NotGiven | None = NOT_GIVEN,
         **kwargs,
     ) -> Generator[ChatCompletionDeltaMessage, None, None]:
         pass
@@ -281,6 +283,7 @@ class AnthropicChatClient(BaseChatClient):
         tool_choice: ToolChoice | NotGiven = NOT_GIVEN,
         response_format: dict | None = None,
         stream_options: ChatCompletionStreamOptionsParam | None = None,
+        top_p: float | NotGiven | None = NOT_GIVEN,
         **kwargs,
     ) -> ChatCompletionMessage | Generator[ChatCompletionDeltaMessage, Any, None]:
         pass
@@ -296,6 +299,7 @@ class AnthropicChatClient(BaseChatClient):
         tool_choice: ToolChoice | NotGiven = NOT_GIVEN,
         response_format: dict | None = None,
         stream_options: ChatCompletionStreamOptionsParam | None = None,
+        top_p: float | NotGiven | None = NOT_GIVEN,
         **kwargs,
     ):
         if model is not None:
@@ -304,10 +308,6 @@ class AnthropicChatClient(BaseChatClient):
             self.stream = stream
         if temperature is not None:
             self.temperature = temperature
-        if isinstance(tools, OpenAINotGiven):
-            tools = NOT_GIVEN
-        if isinstance(tool_choice, OpenAINotGiven):
-            tool_choice = NOT_GIVEN
 
         if self.random_endpoint:
             self.random_endpoint = True
@@ -348,6 +348,7 @@ class AnthropicChatClient(BaseChatClient):
                         tool_choice=_tool_choice,
                         response_format=response_format,
                         stream_options=stream_options,
+                        top_p=top_p,
                         **kwargs,
                     )
                     for chunk in response:
@@ -374,11 +375,19 @@ class AnthropicChatClient(BaseChatClient):
                     tool_choice=_tool_choice,
                     response_format=response_format,
                     stream_options=stream_options,
+                    top_p=top_p,
                     **kwargs,
                 )
 
         assert isinstance(self.raw_client, Anthropic | AnthropicVertex)
 
+        if isinstance(tools, OpenAINotGiven):
+            tools = NOT_GIVEN
+        if isinstance(tool_choice, OpenAINotGiven):
+            tool_choice = NOT_GIVEN
+        if isinstance(top_p, OpenAINotGiven) or top_p is None:
+            top_p = 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]
         if self.model_id is None:
@@ -424,6 +433,7 @@ class AnthropicChatClient(BaseChatClient):
                 max_tokens=max_tokens,
                 tools=tools_params,
                 tool_choice=tool_choice_param,
+                top_p=top_p,
                 **kwargs,
             )
 
@@ -486,6 +496,7 @@ class AnthropicChatClient(BaseChatClient):
                 max_tokens=max_tokens,
                 tools=tools_params,
                 tool_choice=tool_choice_param,
+                top_p=top_p,
                 **kwargs,
             )
 
@@ -614,6 +625,7 @@ class AsyncAnthropicChatClient(BaseAsyncChatClient):
         tool_choice: ToolChoice | NotGiven = NOT_GIVEN,
         response_format: dict | None = None,
         stream_options: ChatCompletionStreamOptionsParam | None = None,
+        top_p: float | NotGiven | None = NOT_GIVEN,
         **kwargs,
     ) -> ChatCompletionMessage:
         pass
@@ -630,6 +642,7 @@ class AsyncAnthropicChatClient(BaseAsyncChatClient):
         tool_choice: ToolChoice | NotGiven = NOT_GIVEN,
         response_format: dict | None = None,
         stream_options: ChatCompletionStreamOptionsParam | None = None,
+        top_p: float | NotGiven | None = NOT_GIVEN,
         **kwargs,
     ) -> AsyncGenerator[ChatCompletionDeltaMessage, Any]:
         pass
@@ -646,6 +659,7 @@ class AsyncAnthropicChatClient(BaseAsyncChatClient):
         tool_choice: ToolChoice | NotGiven = NOT_GIVEN,
         response_format: dict | None = None,
         stream_options: ChatCompletionStreamOptionsParam | None = None,
+        top_p: float | NotGiven | None = NOT_GIVEN,
         **kwargs,
     ) -> ChatCompletionMessage | AsyncGenerator[ChatCompletionDeltaMessage, Any]:
         pass
@@ -661,6 +675,7 @@ class AsyncAnthropicChatClient(BaseAsyncChatClient):
         tool_choice: ToolChoice | NotGiven = NOT_GIVEN,
         response_format: dict | None = None,
         stream_options: ChatCompletionStreamOptionsParam | None = None,
+        top_p: float | NotGiven | None = NOT_GIVEN,
         **kwargs,
     ):
         if model is not None:
@@ -669,10 +684,6 @@ class AsyncAnthropicChatClient(BaseAsyncChatClient):
             self.stream = stream
         if temperature is not None:
             self.temperature = temperature
-        if isinstance(tools, OpenAINotGiven):
-            tools = NOT_GIVEN
-        if isinstance(tool_choice, OpenAINotGiven):
-            tool_choice = NOT_GIVEN
 
         if self.random_endpoint:
             self.random_endpoint = True
@@ -745,6 +756,13 @@ class AsyncAnthropicChatClient(BaseAsyncChatClient):
 
         assert isinstance(self.raw_client, AsyncAnthropic | AsyncAnthropicVertex)
 
+        if isinstance(tools, OpenAINotGiven):
+            tools = NOT_GIVEN
+        if isinstance(tool_choice, OpenAINotGiven):
+            tool_choice = NOT_GIVEN
+        if isinstance(top_p, OpenAINotGiven) or top_p is None:
+            top_p = 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]
         if self.model_id is None:
@@ -790,6 +808,7 @@ class AsyncAnthropicChatClient(BaseAsyncChatClient):
                 max_tokens=max_tokens,
                 tools=tools_params,
                 tool_choice=tool_choice_param,
+                top_p=top_p,
                 **kwargs,
             )
 
@@ -852,6 +871,7 @@ class AsyncAnthropicChatClient(BaseAsyncChatClient):
                 max_tokens=max_tokens,
                 tools=tools_params,
                 tool_choice=tool_choice_param,
+                top_p=top_p,
                 **kwargs,
             )
 

{vectorvein-0.1.56 → vectorvein-0.1.57}/src/vectorvein/chat_clients/base_client.py

@@ -73,6 +73,7 @@ class BaseChatClient(ABC):
         tool_choice: ToolChoice | NotGiven = NOT_GIVEN,
         response_format: dict | None = None,
         stream_options: ChatCompletionStreamOptionsParam | None = None,
+        top_p: float | NotGiven | None = NOT_GIVEN,
         **kwargs,
     ) -> ChatCompletionMessage:
         pass
@@ -90,6 +91,7 @@ class BaseChatClient(ABC):
         tool_choice: ToolChoice | NotGiven = NOT_GIVEN,
         response_format: dict | None = None,
         stream_options: ChatCompletionStreamOptionsParam | None = None,
+        top_p: float | NotGiven | None = NOT_GIVEN,
         **kwargs,
     ) -> Generator[ChatCompletionDeltaMessage, Any, None]:
         pass
@@ -107,6 +109,7 @@ class BaseChatClient(ABC):
         tool_choice: ToolChoice | NotGiven = NOT_GIVEN,
         response_format: dict | None = None,
         stream_options: ChatCompletionStreamOptionsParam | None = None,
+        top_p: float | NotGiven | None = NOT_GIVEN,
         **kwargs,
     ) -> ChatCompletionMessage | Generator[ChatCompletionDeltaMessage, Any, None]:
         pass
@@ -123,6 +126,7 @@ class BaseChatClient(ABC):
         tool_choice: ToolChoice | NotGiven = NOT_GIVEN,
         response_format: dict | None = None,
         stream_options: ChatCompletionStreamOptionsParam | None = None,
+        top_p: float | NotGiven | None = NOT_GIVEN,
         **kwargs,
     ) -> ChatCompletionMessage | Generator[ChatCompletionDeltaMessage, Any, None]:
         pass
@@ -137,6 +141,7 @@ class BaseChatClient(ABC):
         tool_choice: ToolChoice | NotGiven = NOT_GIVEN,
         response_format: dict | None = None,
         stream_options: ChatCompletionStreamOptionsParam | None = None,
+        top_p: float | NotGiven | None = NOT_GIVEN,
         **kwargs,
     ) -> Generator[ChatCompletionDeltaMessage, Any, None]:
         return self.create_completion(
@@ -206,6 +211,7 @@ class BaseAsyncChatClient(ABC):
         tool_choice: ToolChoice | NotGiven = NOT_GIVEN,
         response_format: dict | None = None,
         stream_options: ChatCompletionStreamOptionsParam | None = None,
+        top_p: float | NotGiven | None = NOT_GIVEN,
         **kwargs,
     ) -> ChatCompletionMessage:
         pass
@@ -223,6 +229,7 @@ class BaseAsyncChatClient(ABC):
         tool_choice: ToolChoice | NotGiven = NOT_GIVEN,
         response_format: dict | None = None,
         stream_options: ChatCompletionStreamOptionsParam | None = None,
+        top_p: float | NotGiven | None = NOT_GIVEN,
         **kwargs,
     ) -> AsyncGenerator[ChatCompletionDeltaMessage, None]:
         pass
@@ -240,6 +247,7 @@ class BaseAsyncChatClient(ABC):
         tool_choice: ToolChoice | NotGiven = NOT_GIVEN,
         response_format: dict | None = None,
         stream_options: ChatCompletionStreamOptionsParam | None = None,
+        top_p: float | NotGiven | None = NOT_GIVEN,
         **kwargs,
     ) -> ChatCompletionMessage | AsyncGenerator[ChatCompletionDeltaMessage, None]:
         pass
@@ -256,6 +264,7 @@ class BaseAsyncChatClient(ABC):
         tool_choice: ToolChoice | NotGiven = NOT_GIVEN,
         response_format: dict | None = None,
         stream_options: ChatCompletionStreamOptionsParam | None = None,
+        top_p: float | NotGiven | None = NOT_GIVEN,
         **kwargs,
     ) -> ChatCompletionMessage | AsyncGenerator[ChatCompletionDeltaMessage, None]:
         pass
@@ -270,6 +279,7 @@ class BaseAsyncChatClient(ABC):
         tool_choice: ToolChoice | NotGiven = NOT_GIVEN,
         response_format: dict | None = None,
         stream_options: ChatCompletionStreamOptionsParam | None = None,
+        top_p: float | NotGiven | None = NOT_GIVEN,
         **kwargs,
     ) -> AsyncGenerator[ChatCompletionDeltaMessage, None]:
         return await self.create_completion(

{vectorvein-0.1.56 → vectorvein-0.1.57}/src/vectorvein/chat_clients/gemini_client.py

@@ -65,6 +65,7 @@ class GeminiChatClient(BaseChatClient):
         tool_choice: ToolChoice | NotGiven = NOT_GIVEN,
         response_format: dict | None = None,
         stream_options: ChatCompletionStreamOptionsParam | None = None,
+        top_p: float | NotGiven | None = NOT_GIVEN,
         **kwargs,
     ) -> ChatCompletionMessage:
         pass
@@ -81,6 +82,7 @@ class GeminiChatClient(BaseChatClient):
         tool_choice: ToolChoice | NotGiven = NOT_GIVEN,
         response_format: dict | None = None,
         stream_options: ChatCompletionStreamOptionsParam | None = None,
+        top_p: float | NotGiven | None = NOT_GIVEN,
         **kwargs,
     ) -> Generator[ChatCompletionDeltaMessage, None, None]:
         pass
@@ -97,6 +99,7 @@ class GeminiChatClient(BaseChatClient):
         tool_choice: ToolChoice | NotGiven = NOT_GIVEN,
         response_format: dict | None = None,
         stream_options: ChatCompletionStreamOptionsParam | None = None,
+        top_p: float | NotGiven | None = NOT_GIVEN,
         **kwargs,
     ) -> ChatCompletionMessage | Generator[ChatCompletionDeltaMessage, Any, None]:
         pass
@@ -112,6 +115,7 @@ class GeminiChatClient(BaseChatClient):
         tool_choice: ToolChoice | NotGiven = NOT_GIVEN,
         response_format: dict | None = None,
         stream_options: ChatCompletionStreamOptionsParam | None = None,
+        top_p: float | NotGiven | None = NOT_GIVEN,
         **kwargs,
     ):
         if model is not None:
@@ -137,16 +141,19 @@ class GeminiChatClient(BaseChatClient):
                 model=self.model_setting.id,
             )
 
+        tools_params = {}
         if tools:
             tools_params = {"tools": [{"function_declarations": [tool["function"] for tool in tools]}]}
-        else:
-            tools_params = {}
 
         response_format_params = {}
         if response_format is not None:
             if response_format.get("type") == "json_object":
                 response_format_params = {"response_mime_type": "application/json"}
 
+        top_p_params = {}
+        if top_p:
+            top_p_params = {"top_p": top_p}
+
         if self.random_endpoint:
             self.random_endpoint = True
             endpoint_choice = random.choice(self.backend_settings.models[self.model].endpoints)
@@ -168,6 +175,7 @@ class GeminiChatClient(BaseChatClient):
             "generationConfig": {
                 "temperature": self.temperature,
                 "maxOutputTokens": max_tokens,
+                **top_p_params,
                 **response_format_params,
             },
             **tools_params,
@@ -230,6 +238,8 @@ class GeminiChatClient(BaseChatClient):
             else:
                 client = httpx.Client()
             response = client.post(url, json=request_body, headers=headers, params=params, timeout=None).json()
+            if "error" in response:
+                raise Exception(response["error"])
             result = {
                 "content": "",
                 "usage": {
@@ -302,6 +312,7 @@ class AsyncGeminiChatClient(BaseAsyncChatClient):
         tool_choice: ToolChoice | NotGiven = NOT_GIVEN,
         response_format: dict | None = None,
         stream_options: ChatCompletionStreamOptionsParam | None = None,
+        top_p: float | NotGiven | None = NOT_GIVEN,
         **kwargs,
     ) -> ChatCompletionMessage:
         pass
@@ -318,6 +329,7 @@ class AsyncGeminiChatClient(BaseAsyncChatClient):
         tool_choice: ToolChoice | NotGiven = NOT_GIVEN,
         response_format: dict | None = None,
         stream_options: ChatCompletionStreamOptionsParam | None = None,
+        top_p: float | NotGiven | None = NOT_GIVEN,
         **kwargs,
     ) -> AsyncGenerator[ChatCompletionDeltaMessage, Any]:
         pass
@@ -334,6 +346,7 @@ class AsyncGeminiChatClient(BaseAsyncChatClient):
         tool_choice: ToolChoice | NotGiven = NOT_GIVEN,
         response_format: dict | None = None,
         stream_options: ChatCompletionStreamOptionsParam | None = None,
+        top_p: float | NotGiven | None = NOT_GIVEN,
         **kwargs,
     ) -> ChatCompletionMessage | AsyncGenerator[ChatCompletionDeltaMessage, Any]:
         pass
@@ -349,6 +362,7 @@ class AsyncGeminiChatClient(BaseAsyncChatClient):
         tool_choice: ToolChoice | NotGiven = NOT_GIVEN,
         response_format: dict | None = None,
         stream_options: ChatCompletionStreamOptionsParam | None = None,
+        top_p: float | NotGiven | None = NOT_GIVEN,
         **kwargs,
     ):
         if model is not None:
@@ -374,16 +388,19 @@ class AsyncGeminiChatClient(BaseAsyncChatClient):
                 model=self.model_setting.id,
             )
 
+        tools_params = {}
         if tools:
             tools_params = {"tools": [{"function_declarations": [tool["function"] for tool in tools]}]}
-        else:
-            tools_params = {}
 
         response_format_params = {}
         if response_format is not None:
             if response_format.get("type") == "json_object":
                 response_format_params = {"response_mime_type": "application/json"}
 
+        top_p_params = {}
+        if top_p:
+            top_p_params = {"top_p": top_p}
+
         if self.random_endpoint:
             self.random_endpoint = True
             endpoint_choice = random.choice(self.backend_settings.models[self.model].endpoints)
@@ -405,6 +422,7 @@ class AsyncGeminiChatClient(BaseAsyncChatClient):
             "generationConfig": {
                 "temperature": self.temperature,
                 "maxOutputTokens": max_tokens,
+                **top_p_params,
                 **response_format_params,
             },
             **tools_params,
@@ -469,6 +487,8 @@ class AsyncGeminiChatClient(BaseAsyncChatClient):
             async with client:
                 response = await client.post(url, json=request_body, headers=headers, params=params, timeout=None)
                 response = response.json()
+                if "error" in response:
+                    raise Exception(response["error"])
                 result = {
                     "content": "",
                     "usage": {

{vectorvein-0.1.56 → vectorvein-0.1.57}/src/vectorvein/chat_clients/minimax_client.py

@@ -91,6 +91,7 @@ class MiniMaxChatClient(BaseChatClient):
         tool_choice: ToolChoice | NotGiven = NOT_GIVEN,
         response_format: dict | None = None,
         stream_options: ChatCompletionStreamOptionsParam | None = None,
+        top_p: float | NotGiven | None = NOT_GIVEN,
         **kwargs,
     ) -> ChatCompletionMessage:
         pass
@@ -107,6 +108,7 @@ class MiniMaxChatClient(BaseChatClient):
         tool_choice: ToolChoice | NotGiven = NOT_GIVEN,
         response_format: dict | None = None,
         stream_options: ChatCompletionStreamOptionsParam | None = None,
+        top_p: float | NotGiven | None = NOT_GIVEN,
         **kwargs,
     ) -> Generator[ChatCompletionDeltaMessage, None, None]:
         pass
@@ -123,6 +125,7 @@ class MiniMaxChatClient(BaseChatClient):
         tool_choice: ToolChoice | NotGiven = NOT_GIVEN,
         response_format: dict | None = None,
         stream_options: ChatCompletionStreamOptionsParam | None = None,
+        top_p: float | NotGiven | None = NOT_GIVEN,
         **kwargs,
     ) -> ChatCompletionMessage | Generator[ChatCompletionDeltaMessage, Any, None]:
         pass
@@ -138,6 +141,7 @@ class MiniMaxChatClient(BaseChatClient):
         tool_choice: ToolChoice | NotGiven = NOT_GIVEN,
         response_format: dict | None = None,
         stream_options: ChatCompletionStreamOptionsParam | None = None,
+        top_p: float | NotGiven | None = NOT_GIVEN,
         **kwargs,
     ):
         if model is not None:
@@ -190,6 +194,11 @@ class MiniMaxChatClient(BaseChatClient):
         else:
             tools_params = {}
 
+        if top_p:
+            top_p_params = {"top_p": top_p}
+        else:
+            top_p_params = {}
+
         if max_tokens is None:
             max_output_tokens = self.model_setting.max_output_tokens
             if max_output_tokens is not None:
@@ -218,6 +227,7 @@ class MiniMaxChatClient(BaseChatClient):
             "temperature": self.temperature,
             "stream": self.stream,
             "mask_sensitive_info": False,
+            **top_p_params,
             **tools_params,
             **kwargs,
         }
@@ -326,6 +336,7 @@ class AsyncMiniMaxChatClient(BaseAsyncChatClient):
         tool_choice: ToolChoice | NotGiven = NOT_GIVEN,
         response_format: dict | None = None,
         stream_options: ChatCompletionStreamOptionsParam | None = None,
+        top_p: float | NotGiven | None = NOT_GIVEN,
         **kwargs,
     ) -> ChatCompletionMessage:
         pass
@@ -342,6 +353,7 @@ class AsyncMiniMaxChatClient(BaseAsyncChatClient):
         tool_choice: ToolChoice | NotGiven = NOT_GIVEN,
         response_format: dict | None = None,
         stream_options: ChatCompletionStreamOptionsParam | None = None,
+        top_p: float | NotGiven | None = NOT_GIVEN,
         **kwargs,
     ) -> AsyncGenerator[ChatCompletionDeltaMessage, Any]:
         pass
@@ -358,6 +370,7 @@ class AsyncMiniMaxChatClient(BaseAsyncChatClient):
         tool_choice: ToolChoice | NotGiven = NOT_GIVEN,
         response_format: dict | None = None,
         stream_options: ChatCompletionStreamOptionsParam | None = None,
+        top_p: float | NotGiven | None = NOT_GIVEN,
         **kwargs,
     ) -> ChatCompletionMessage | AsyncGenerator[ChatCompletionDeltaMessage, Any]:
         pass
@@ -373,6 +386,7 @@ class AsyncMiniMaxChatClient(BaseAsyncChatClient):
         tool_choice: ToolChoice | NotGiven = NOT_GIVEN,
         response_format: dict | None = None,
         stream_options: ChatCompletionStreamOptionsParam | None = None,
+        top_p: float | NotGiven | None = NOT_GIVEN,
         **kwargs,
     ):
         if model is not None:
@@ -423,6 +437,11 @@ class AsyncMiniMaxChatClient(BaseAsyncChatClient):
         else:
             tools_params = {}
 
+        if top_p:
+            top_p_params = {"top_p": top_p}
+        else:
+            top_p_params = {}
+
         if max_tokens is None:
             max_output_tokens = self.model_setting.max_output_tokens
             if max_output_tokens is not None:
@@ -451,6 +470,7 @@ class AsyncMiniMaxChatClient(BaseAsyncChatClient):
             "temperature": self.temperature,
             "stream": self.stream,
             "mask_sensitive_info": False,
+            **top_p_params,
             **tools_params,
             **kwargs,
         }

{vectorvein-0.1.56 → vectorvein-0.1.57}/src/vectorvein/chat_clients/openai_compatible_client.py

@@ -24,6 +24,7 @@ from ..types.llm_parameters import (
     ToolParam,
     ToolChoice,
     OpenAINotGiven,
+    AnthropicNotGiven,
     Usage,
     ChatCompletionMessage,
     ChatCompletionDeltaMessage,
@@ -97,6 +98,7 @@ class OpenAICompatibleChatClient(BaseChatClient):
         tool_choice: ToolChoice | NotGiven = NOT_GIVEN,
         response_format: dict | None = None,
         stream_options: ChatCompletionStreamOptionsParam | None | OpenAINotGiven = NOT_GIVEN,
+        top_p: float | NotGiven | None = NOT_GIVEN,
         **kwargs,
     ) -> ChatCompletionMessage:
         pass
@@ -113,6 +115,7 @@ class OpenAICompatibleChatClient(BaseChatClient):
         tool_choice: ToolChoice | NotGiven = NOT_GIVEN,
         response_format: dict | None = None,
         stream_options: ChatCompletionStreamOptionsParam | None | OpenAINotGiven = NOT_GIVEN,
+        top_p: float | NotGiven | None = NOT_GIVEN,
         **kwargs,
     ) -> Generator[ChatCompletionDeltaMessage, None, None]:
         pass
@@ -129,6 +132,7 @@ class OpenAICompatibleChatClient(BaseChatClient):
         tool_choice: ToolChoice | NotGiven = NOT_GIVEN,
         response_format: dict | None = None,
         stream_options: ChatCompletionStreamOptionsParam | None | OpenAINotGiven = NOT_GIVEN,
+        top_p: float | NotGiven | None = NOT_GIVEN,
         **kwargs,
     ) -> ChatCompletionMessage | Generator[ChatCompletionDeltaMessage, Any, None]:
         pass
@@ -144,6 +148,7 @@ class OpenAICompatibleChatClient(BaseChatClient):
         tool_choice: ToolChoice | NotGiven = NOT_GIVEN,
         response_format: dict | None = None,
         stream_options: ChatCompletionStreamOptionsParam | None | OpenAINotGiven = NOT_GIVEN,
+        top_p: float | NotGiven | None = NOT_GIVEN,
         **kwargs,
     ):
         if model is not None:
@@ -152,6 +157,8 @@ class OpenAICompatibleChatClient(BaseChatClient):
             self.stream = stream
         if temperature is not None:
             self.temperature = temperature
+        if isinstance(top_p, AnthropicNotGiven):
+            top_p = 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]
@@ -206,6 +213,7 @@ class OpenAICompatibleChatClient(BaseChatClient):
                 stream=True,
                 temperature=self.temperature,
                 max_tokens=max_tokens,
+                top_p=top_p,
                 **_stream_options_params,
                 **self.response_format,
                 **tools_params,
@@ -260,6 +268,7 @@ class OpenAICompatibleChatClient(BaseChatClient):
                 stream=False,
                 temperature=self.temperature,
                 max_tokens=max_tokens,
+                top_p=top_p,
                 **self.response_format,
                 **tools_params,
                 **kwargs,
@@ -351,6 +360,7 @@ class AsyncOpenAICompatibleChatClient(BaseAsyncChatClient):
         tool_choice: ToolChoice | NotGiven = NOT_GIVEN,
         response_format: dict | None = None,
         stream_options: ChatCompletionStreamOptionsParam | None | OpenAINotGiven = NOT_GIVEN,
+        top_p: float | NotGiven | None = NOT_GIVEN,
         **kwargs,
     ) -> ChatCompletionMessage:
         pass
@@ -367,6 +377,7 @@ class AsyncOpenAICompatibleChatClient(BaseAsyncChatClient):
         tool_choice: ToolChoice | NotGiven = NOT_GIVEN,
         response_format: dict | None = None,
         stream_options: ChatCompletionStreamOptionsParam | None | OpenAINotGiven = NOT_GIVEN,
+        top_p: float | NotGiven | None = NOT_GIVEN,
         **kwargs,
     ) -> AsyncGenerator[ChatCompletionDeltaMessage, Any]:
         pass
@@ -383,6 +394,7 @@ class AsyncOpenAICompatibleChatClient(BaseAsyncChatClient):
         tool_choice: ToolChoice | NotGiven = NOT_GIVEN,
         response_format: dict | None = None,
         stream_options: ChatCompletionStreamOptionsParam | None | OpenAINotGiven = NOT_GIVEN,
+        top_p: float | NotGiven | None = NOT_GIVEN,
         **kwargs,
     ) -> ChatCompletionMessage | AsyncGenerator[ChatCompletionDeltaMessage, Any]:
         pass
@@ -398,6 +410,7 @@ class AsyncOpenAICompatibleChatClient(BaseAsyncChatClient):
         tool_choice: ToolChoice | NotGiven = NOT_GIVEN,
         response_format: dict | None = None,
         stream_options: ChatCompletionStreamOptionsParam | None | OpenAINotGiven = NOT_GIVEN,
+        top_p: float | NotGiven | None = NOT_GIVEN,
         **kwargs,
     ):
         if model is not None:
@@ -406,6 +419,8 @@ class AsyncOpenAICompatibleChatClient(BaseAsyncChatClient):
             self.stream = stream
         if temperature is not None:
             self.temperature = temperature
+        if isinstance(top_p, AnthropicNotGiven):
+            top_p = 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]
@@ -460,6 +475,7 @@ class AsyncOpenAICompatibleChatClient(BaseAsyncChatClient):
                 stream=self.stream,
                 temperature=self.temperature,
                 max_tokens=max_tokens,
+                top_p=top_p,
                 **_stream_options_params,
                 **self.response_format,
                 **tools_params,
@@ -515,6 +531,7 @@ class AsyncOpenAICompatibleChatClient(BaseAsyncChatClient):
                 stream=self.stream,
                 temperature=self.temperature,
                 max_tokens=max_tokens,
+                top_p=top_p,
                 **self.response_format,
                 **tools_params,
                 **kwargs,

{vectorvein-0.1.56 → vectorvein-0.1.57}/src/vectorvein/chat_clients/utils.py

@@ -201,6 +201,10 @@ def get_token_counts(text: str | dict, model: str = "", use_token_server_first:
         result = response.json()
         return result["data"]["total_tokens"]
     elif model.startswith("gemini"):
+        # TODO: gemini-exp-1114 暂时不支持，使用 gemini-1.5-flash 代替
+        if model == "gemini-exp-1114":
+            model = "gemini-1.5-flash"
+
         model_setting = settings.gemini.models[model]
         if len(model_setting.endpoints) == 0:
             return len(get_gpt_35_encoding().encode(text))
@@ -208,13 +212,14 @@ def get_token_counts(text: str | dict, model: str = "", use_token_server_first:
         if isinstance(endpoint_id, dict):
             endpoint_id = endpoint_id["endpoint_id"]
         endpoint = settings.get_endpoint(endpoint_id)
+
         base_url = f"{endpoint.api_base}/models/{model_setting.id}:countTokens"
         params = {"key": endpoint.api_key}
         request_body = {
             "contents": {
                 "role": "USER",
                 "parts": [
-                    {"text": "TEXT"},
+                    {"text": text},
                 ],
             },
         }

{vectorvein-0.1.56 → vectorvein-0.1.57}/src/vectorvein/types/defaults.py

@@ -609,6 +609,13 @@ MINIMAX_MODELS: Final[Dict[str, Dict[str, Any]]] = {
         "function_call_available": True,
         "response_format_available": True,
     },
+    "abab7-preview": {
+        "id": "abab7-preview",
+        "context_length": 245760,
+        "max_output_tokens": 245760,
+        "function_call_available": True,
+        "response_format_available": True,
+    },
 }
 
 # Gemini models
@@ -616,7 +623,8 @@ GEMINI_DEFAULT_MODEL: Final[str] = "gemini-1.5-pro"
 GEMINI_MODELS: Final[Dict[str, Dict[str, Any]]] = {
     "gemini-1.5-pro": {
         "id": "gemini-1.5-pro",
-        "context_length": 1048576,
+        "context_length": 2097152,
+        "max_output_tokens": 8192,
         "function_call_available": True,
         "response_format_available": True,
         "native_multimodal": True,
@@ -624,6 +632,22 @@ GEMINI_MODELS: Final[Dict[str, Dict[str, Any]]] = {
     "gemini-1.5-flash": {
         "id": "gemini-1.5-flash",
         "context_length": 1048576,
+        "max_output_tokens": 8192,
+        "function_call_available": True,
+        "response_format_available": True,
+        "native_multimodal": True,
+    },
+    "gemini-exp-1114": {
+        "id": "gemini-exp-1114",
+        "context_length": 32767,
+        "function_call_available": True,
+        "response_format_available": True,
+        "native_multimodal": True,
+    },
+    "gemini-1.5-flash-8b": {
+        "id": "gemini-1.5-flash-8b",
+        "context_length": 1048576,
+        "max_output_tokens": 8192,
         "function_call_available": True,
         "response_format_available": True,
         "native_multimodal": True,