vectorvein 0.1.36__tar.gz → 0.1.37__tar.gz

{vectorvein-0.1.36 → vectorvein-0.1.37}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.1
 Name: vectorvein
-Version: 0.1.36
+Version: 0.1.37
 Summary: Default template for PDM package
 Author-Email: Anderson <andersonby@163.com>
 License: MIT

{vectorvein-0.1.36 → vectorvein-0.1.37}/pyproject.toml

@@ -16,7 +16,7 @@ description = "Default template for PDM package"
 name = "vectorvein"
 readme = "README.md"
 requires-python = ">=3.10"
-version = "0.1.36"
+version = "0.1.37"
 
 [project.license]
 text = "MIT"

{vectorvein-0.1.36 → vectorvein-0.1.37}/src/vectorvein/chat_clients/anthropic_client.py

@@ -270,18 +270,24 @@ class AnthropicChatClient(BaseChatClient):
 
         if self.endpoint.api_schema_type == "openai":
             if self.stream:
-                return OpenAICompatibleChatClient(
-                    model=self.model,
-                    stream=True,
-                    temperature=self.temperature,
-                    context_length_control=self.context_length_control,
-                    random_endpoint=self.random_endpoint,
-                    endpoint_id=self.endpoint_id,
-                    http_client=self.http_client,
-                    backend_name=self.BACKEND_NAME,
-                ).create_completion(
-                    messages, model, True, temperature, max_tokens, tools, tool_choice, response_format, **kwargs
-                )
+
+                def _generator():
+                    response = OpenAICompatibleChatClient(
+                        model=self.model,
+                        stream=True,
+                        temperature=self.temperature,
+                        context_length_control=self.context_length_control,
+                        random_endpoint=self.random_endpoint,
+                        endpoint_id=self.endpoint_id,
+                        http_client=self.http_client,
+                        backend_name=self.BACKEND_NAME,
+                    ).create_completion(
+                        messages, model, True, temperature, max_tokens, tools, tool_choice, response_format, **kwargs
+                    )
+                    for chunk in response:
+                        yield chunk
+
+                return _generator()
             else:
                 return OpenAICompatibleChatClient(
                     model=self.model,
@@ -586,20 +592,27 @@ class AsyncAnthropicChatClient(BaseAsyncChatClient):
 
         if self.endpoint.api_schema_type == "openai":
             if self.stream:
-                return AsyncOpenAICompatibleChatClient(
-                    model=self.model,
-                    stream=True,
-                    temperature=self.temperature,
-                    context_length_control=self.context_length_control,
-                    random_endpoint=self.random_endpoint,
-                    endpoint_id=self.endpoint_id,
-                    http_client=self.http_client,
-                    backend_name=self.BACKEND_NAME,
-                ).create_completion(
-                    messages, model, True, temperature, max_tokens, tools, tool_choice, response_format, **kwargs
-                )
+
+                async def _generator():
+                    client = AsyncOpenAICompatibleChatClient(
+                        model=self.model,
+                        stream=True,
+                        temperature=self.temperature,
+                        context_length_control=self.context_length_control,
+                        random_endpoint=self.random_endpoint,
+                        endpoint_id=self.endpoint_id,
+                        http_client=self.http_client,
+                        backend_name=self.BACKEND_NAME,
+                    )
+                    response = await client.create_completion(
+                        messages, model, True, temperature, max_tokens, tools, tool_choice, response_format, **kwargs
+                    )
+                    async for chunk in response:
+                        yield chunk
+
+                return _generator()
             else:
-                return AsyncOpenAICompatibleChatClient(
+                client = AsyncOpenAICompatibleChatClient(
                     model=self.model,
                     stream=False,
                     temperature=self.temperature,
@@ -608,7 +621,8 @@ class AsyncAnthropicChatClient(BaseAsyncChatClient):
                     endpoint_id=self.endpoint_id,
                     http_client=self.http_client,
                     backend_name=self.BACKEND_NAME,
-                ).create_completion(
+                )
+                return await client.create_completion(
                     messages, model, False, temperature, max_tokens, tools, tool_choice, response_format, **kwargs
                 )
 

{vectorvein-0.1.36 → vectorvein-0.1.37}/src/vectorvein/chat_clients/openai_compatible_client.py

@@ -25,7 +25,6 @@ from ..types.llm_parameters import (
     NOT_GIVEN,
     ToolParam,
     ToolChoice,
-    BackendSettings,
     ChatCompletionMessage,
     ChatCompletionDeltaMessage,
 )