vectorvein 0.1.3__tar.gz → 0.1.5__tar.gz

{vectorvein-0.1.3 → vectorvein-0.1.5}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.1
 Name: vectorvein
-Version: 0.1.3
+Version: 0.1.5
 Summary: Default template for PDM package
 Author-Email: Anderson <andersonby@163.com>
 License: MIT

{vectorvein-0.1.3 → vectorvein-0.1.5}/pyproject.toml

@@ -14,7 +14,7 @@ description = "Default template for PDM package"
 name = "vectorvein"
 readme = "README.md"
 requires-python = ">=3.10"
-version = "0.1.3"
+version = "0.1.5"
 
 [project.license]
 text = "MIT"

{vectorvein-0.1.3 → vectorvein-0.1.5}/src/vectorvein/chat_clients/__init__.py

@@ -55,7 +55,7 @@ BackendMap = {
 def create_chat_client(
     backend: BackendType,
     model: str | None = None,
-    stream: bool = True,
+    stream: bool = False,
     temperature: float = 0.7,
     context_length_control: ContextLengthControlType = defs.CONTEXT_LENGTH_CONTROL,
     **kwargs,
@@ -80,7 +80,7 @@ def create_chat_client(
 def create_async_chat_client(
     backend: BackendType,
     model: str | None = None,
-    stream: bool = True,
+    stream: bool = False,
     temperature: float = 0.7,
     context_length_control: ContextLengthControlType = defs.CONTEXT_LENGTH_CONTROL,
     **kwargs,

{vectorvein-0.1.3 → vectorvein-0.1.5}/src/vectorvein/chat_clients/anthropic_client.py

@@ -3,6 +3,7 @@
 import json
 import random
 
+from openai._types import NotGiven as OpenAINotGiven
 from anthropic import Anthropic, AnthropicVertex, AsyncAnthropic, AsyncAnthropicVertex
 from anthropic._types import NotGiven, NOT_GIVEN
 from anthropic.types import (
@@ -22,6 +23,7 @@ from .utils import cutoff_messages
 from ..types import defaults as defs
 from .base_client import BaseChatClient, BaseAsyncChatClient
 from ..types.enums import ContextLengthControlType, BackendType
+from ..types.llm_parameters import ChatCompletionMessage, ChatCompletionDeltaMessage
 
 
 def refactor_tool_use_params(tools: list):
@@ -123,6 +125,8 @@ class AnthropicChatClient(BaseChatClient):
             self.stream = stream
         if temperature is not None:
             self.temperature = temperature
+        if isinstance(tool_choice, OpenAINotGiven):
+            tool_choice = NOT_GIVEN
 
         self.model_setting = self.backend_settings.models[self.model]
 
@@ -218,7 +222,7 @@ class AnthropicChatClient(BaseChatClient):
                             ]
                         elif chunk.content_block.type == "text":
                             message["content"] = chunk.content_block.text
-                        yield message
+                        yield ChatCompletionDeltaMessage(**message)
                     elif isinstance(chunk, RawContentBlockDeltaEvent):
                         if chunk.delta.type == "text_delta":
                             message["content"] = chunk.delta.text
@@ -236,13 +240,13 @@ class AnthropicChatClient(BaseChatClient):
                                     "type": "function",
                                 }
                             ]
-                        yield message
+                        yield ChatCompletionDeltaMessage(**message)
                     elif isinstance(chunk, RawMessageDeltaEvent):
                         result["usage"]["completion_tokens"] = chunk.usage.output_tokens
                         result["usage"]["total_tokens"] = (
                             result["usage"]["prompt_tokens"] + result["usage"]["completion_tokens"]
                         )
-                        yield {"usage": result["usage"]}
+                        yield ChatCompletionDeltaMessage(usage=result["usage"])
 
             return generator()
         else:
@@ -264,7 +268,7 @@ class AnthropicChatClient(BaseChatClient):
             if tool_calls:
                 result["tool_calls"] = refactor_tool_calls(tool_calls)
 
-            return result
+            return ChatCompletionMessage(**result)
 
 
 class AsyncAnthropicChatClient(BaseAsyncChatClient):
@@ -307,6 +311,8 @@ class AsyncAnthropicChatClient(BaseAsyncChatClient):
             self.stream = stream
         if temperature is not None:
             self.temperature = temperature
+        if isinstance(tool_choice, OpenAINotGiven):
+            tool_choice = NOT_GIVEN
 
         self.model_setting = self.backend_settings.models[self.model]
 
@@ -401,7 +407,7 @@ class AsyncAnthropicChatClient(BaseAsyncChatClient):
                             ]
                         elif chunk.content_block.type == "text":
                             message["content"] = chunk.content_block.text
-                        yield message
+                        yield ChatCompletionDeltaMessage(**message)
                     elif isinstance(chunk, RawContentBlockDeltaEvent):
                         if chunk.delta.type == "text_delta":
                             message["content"] = chunk.delta.text
@@ -419,13 +425,13 @@ class AsyncAnthropicChatClient(BaseAsyncChatClient):
                                     "type": "function",
                                 }
                             ]
-                        yield message
+                        yield ChatCompletionDeltaMessage(**message)
                     elif isinstance(chunk, RawMessageDeltaEvent):
                         result["usage"]["completion_tokens"] = chunk.usage.output_tokens
                         result["usage"]["total_tokens"] = (
                             result["usage"]["prompt_tokens"] + result["usage"]["completion_tokens"]
                         )
-                        yield {"usage": result["usage"]}
+                        yield ChatCompletionDeltaMessage(usage=result["usage"])
 
             return generator()
         else:
@@ -447,4 +453,4 @@ class AsyncAnthropicChatClient(BaseAsyncChatClient):
             if tool_calls:
                 result["tool_calls"] = refactor_tool_calls(tool_calls)
 
-            return result
+            return ChatCompletionMessage(**result)

{vectorvein-0.1.3 → vectorvein-0.1.5}/src/vectorvein/chat_clients/base_client.py

@@ -1,10 +1,14 @@
 # @Author: Bi Ying
 # @Date:   2024-07-26 14:48:55
 from abc import ABC, abstractmethod
+from typing import Generator, AsyncGenerator, Any
+
+from openai._types import NotGiven, NOT_GIVEN
 
 from ..settings import settings
 from ..types import defaults as defs
 from ..types.enums import ContextLengthControlType, BackendType
+from ..types.llm_parameters import ChatCompletionMessage, ChatCompletionDeltaMessage
 
 
 class BaseChatClient(ABC):
@@ -14,7 +18,7 @@ class BaseChatClient(ABC):
     def __init__(
         self,
         model: str = "",
-        stream: bool = True,
+        stream: bool = False,
         temperature: float = 0.7,
         context_length_control: ContextLengthControlType = defs.CONTEXT_LENGTH_CONTROL,
         random_endpoint: bool = True,
@@ -40,14 +44,33 @@ class BaseChatClient(ABC):
         self,
         messages: list,
         model: str | None = None,
-        stream: bool = True,
+        stream: bool = False,
         temperature: float = 0.7,
         max_tokens: int = 2000,
-        tools: list | None = None,
-        tool_choice: str | None = None,
-    ):
+        tools: list | NotGiven = NOT_GIVEN,
+        tool_choice: str | NotGiven = NOT_GIVEN,
+    ) -> ChatCompletionMessage | Generator[ChatCompletionDeltaMessage, Any, None]:
         pass
 
+    def create_stream(
+        self,
+        messages: list,
+        model: str | None = None,
+        temperature: float = 0.7,
+        max_tokens: int = 2000,
+        tools: list | NotGiven = NOT_GIVEN,
+        tool_choice: str | NotGiven = NOT_GIVEN,
+    ) -> Generator[ChatCompletionDeltaMessage, Any, None]:
+        return self.create_completion(
+            messages=messages,
+            model=model,
+            stream=True,
+            temperature=temperature,
+            max_tokens=max_tokens,
+            tools=tools,
+            tool_choice=tool_choice,
+        )
+
 
 class BaseAsyncChatClient(ABC):
     DEFAULT_MODEL: str | None = None
@@ -56,7 +79,7 @@ class BaseAsyncChatClient(ABC):
     def __init__(
         self,
         model: str = "",
-        stream: bool = True,
+        stream: bool = False,
         temperature: float = 0.7,
         context_length_control: ContextLengthControlType = defs.CONTEXT_LENGTH_CONTROL,
         random_endpoint: bool = True,
@@ -82,10 +105,29 @@ class BaseAsyncChatClient(ABC):
         self,
         messages: list,
         model: str | None = None,
-        stream: bool = True,
+        stream: bool = False,
         temperature: float = 0.7,
         max_tokens: int = 2000,
-        tools: list | None = None,
-        tool_choice: str | None = None,
-    ):
+        tools: list | NotGiven = NOT_GIVEN,
+        tool_choice: str | NotGiven = NOT_GIVEN,
+    ) -> ChatCompletionMessage | AsyncGenerator[ChatCompletionDeltaMessage, None]:
         pass
+
+    async def create_stream(
+        self,
+        messages: list,
+        model: str | None = None,
+        temperature: float = 0.7,
+        max_tokens: int = 2000,
+        tools: list | NotGiven = NOT_GIVEN,
+        tool_choice: str | NotGiven = NOT_GIVEN,
+    ) -> AsyncGenerator[ChatCompletionDeltaMessage, None]:
+        return await self.create_completion(
+            messages=messages,
+            model=model,
+            stream=True,
+            temperature=temperature,
+            max_tokens=max_tokens,
+            tools=tools,
+            tool_choice=tool_choice,
+        )

{vectorvein-0.1.3 → vectorvein-0.1.5}/src/vectorvein/chat_clients/gemini_client.py

@@ -10,6 +10,7 @@ from .utils import cutoff_messages
 from ..types import defaults as defs
 from .base_client import BaseChatClient, BaseAsyncChatClient
 from ..types.enums import ContextLengthControlType, BackendType
+from ..types.llm_parameters import ChatCompletionMessage, ChatCompletionDeltaMessage
 
 
 class GeminiChatClient(BaseChatClient):
@@ -120,7 +121,7 @@ class GeminiChatClient(BaseChatClient):
                             message["tool_calls"] = [
                                 {
                                     "index": 0,
-                                    "id": 0,
+                                    "id": "call_0",
                                     "function": {
                                         "arguments": json.dumps(
                                             chunk_content["functionCall"]["args"], ensure_ascii=False
@@ -136,7 +137,7 @@ class GeminiChatClient(BaseChatClient):
                             "completion_tokens": data["usageMetadata"]["candidatesTokenCount"],
                             "total_tokens": data["usageMetadata"]["totalTokenCount"],
                         }
-                        yield message
+                        yield ChatCompletionDeltaMessage(**message)
 
             return generator()
         else:
@@ -160,7 +161,7 @@ class GeminiChatClient(BaseChatClient):
             if tool_calls:
                 result["tool_calls"] = tool_calls
 
-            return result
+            return ChatCompletionMessage(**result)
 
 
 class AsyncGeminiChatClient(BaseAsyncChatClient):
@@ -272,7 +273,7 @@ class AsyncGeminiChatClient(BaseAsyncChatClient):
                             message["tool_calls"] = [
                                 {
                                     "index": 0,
-                                    "id": 0,
+                                    "id": "call_0",
                                     "function": {
                                         "arguments": json.dumps(
                                             chunk_content["functionCall"]["args"], ensure_ascii=False
@@ -288,7 +289,7 @@ class AsyncGeminiChatClient(BaseAsyncChatClient):
                             "completion_tokens": data["usageMetadata"]["candidatesTokenCount"],
                             "total_tokens": data["usageMetadata"]["totalTokenCount"],
                         }
-                        yield message
+                        yield ChatCompletionDeltaMessage(**message)
 
             return generator()
         else:
@@ -314,4 +315,4 @@ class AsyncGeminiChatClient(BaseAsyncChatClient):
                 if tool_calls:
                     result["tool_calls"] = tool_calls
 
-                return result
+                return ChatCompletionMessage(**result)

{vectorvein-0.1.3 → vectorvein-0.1.5}/src/vectorvein/chat_clients/minimax_client.py

@@ -4,12 +4,14 @@ import json
 import random
 
 import httpx
+from openai._types import NotGiven
 
 from ..settings import settings
 from .utils import cutoff_messages
 from ..types import defaults as defs
 from .base_client import BaseChatClient, BaseAsyncChatClient
 from ..types.enums import ContextLengthControlType, BackendType
+from ..types.llm_parameters import ChatCompletionMessage, ChatCompletionDeltaMessage
 
 
 def extract_tool_calls(response):
@@ -74,6 +76,8 @@ class MiniMaxChatClient(BaseChatClient):
             self.stream = stream
         if temperature is not None:
             self.temperature = temperature
+        if isinstance(tool_choice, NotGiven):
+            tool_choice = "auto"
 
         self.model_setting = self.backend_settings.models[self.model]
         if self.random_endpoint:
@@ -141,33 +145,39 @@ class MiniMaxChatClient(BaseChatClient):
                             if "usage" not in chunk_data:
                                 continue
                             else:
-                                yield {
-                                    "content": chunk_data["choices"][0]["message"].get("content"),
-                                    "role": "assistant",
-                                    **tool_calls_params,
-                                }
+                                yield ChatCompletionDeltaMessage(
+                                    **{
+                                        "content": chunk_data["choices"][0]["message"].get("content"),
+                                        "role": "assistant",
+                                        **tool_calls_params,
+                                    }
+                                )
                         else:
                             if "usage" in chunk_data:
                                 continue
-                            yield {
-                                "content": chunk_data["choices"][0]["delta"]["content"],
-                                "role": "assistant",
-                            }
+                            yield ChatCompletionDeltaMessage(
+                                **{
+                                    "content": chunk_data["choices"][0]["delta"]["content"],
+                                    "role": "assistant",
+                                }
+                            )
 
             return generator()
         else:
             result = response.json()
             tool_calls_params = extract_tool_calls(result)
-            return {
-                "content": result["choices"][0]["message"].get("content"),
-                "usage": {
-                    "prompt_tokens": 0,
-                    "completion_tokens": result["usage"]["total_tokens"],
-                    "total_tokens": result["usage"]["total_tokens"],
-                },
-                "role": "assistant",
-                **tool_calls_params,
-            }
+            return ChatCompletionMessage(
+                **{
+                    "content": result["choices"][0]["message"].get("content"),
+                    "usage": {
+                        "prompt_tokens": 0,
+                        "completion_tokens": result["usage"]["total_tokens"],
+                        "total_tokens": result["usage"]["total_tokens"],
+                    },
+                    "role": "assistant",
+                    **tool_calls_params,
+                }
+            )
 
 
 class AsyncMiniMaxChatClient(BaseAsyncChatClient):
@@ -211,6 +221,8 @@ class AsyncMiniMaxChatClient(BaseAsyncChatClient):
             self.stream = stream
         if temperature is not None:
             self.temperature = temperature
+        if isinstance(tool_choice, NotGiven):
+            tool_choice = "auto"
 
         self.model_setting = self.backend_settings.models[self.model]
         if self.random_endpoint:
@@ -277,18 +289,22 @@ class AsyncMiniMaxChatClient(BaseAsyncChatClient):
                                 if "usage" not in chunk_data:
                                     continue
                                 else:
-                                    yield {
-                                        "content": chunk_data["choices"][0]["message"].get("content"),
-                                        "role": "assistant",
-                                        **tool_calls_params,
-                                    }
+                                    yield ChatCompletionDeltaMessage(
+                                        **{
+                                            "content": chunk_data["choices"][0]["message"].get("content"),
+                                            "role": "assistant",
+                                            **tool_calls_params,
+                                        }
+                                    )
                             else:
                                 if "usage" in chunk_data:
                                     continue
-                                yield {
-                                    "content": chunk_data["choices"][0]["delta"]["content"],
-                                    "role": "assistant",
-                                }
+                                yield ChatCompletionDeltaMessage(
+                                    **{
+                                        "content": chunk_data["choices"][0]["delta"]["content"],
+                                        "role": "assistant",
+                                    }
+                                )
 
             return generator()
         else:
@@ -300,16 +316,18 @@ class AsyncMiniMaxChatClient(BaseAsyncChatClient):
             )
             result = response.json()
             tool_calls_params = extract_tool_calls(result)
-            return {
-                "content": result["choices"][0]["message"].get("content"),
-                "usage": {
-                    "prompt_tokens": 0,
-                    "completion_tokens": result["usage"]["total_tokens"],
-                    "total_tokens": result["usage"]["total_tokens"],
-                },
-                "role": "assistant",
-                **tool_calls_params,
-            }
+            return ChatCompletionMessage(
+                **{
+                    "content": result["choices"][0]["message"].get("content"),
+                    "usage": {
+                        "prompt_tokens": 0,
+                        "completion_tokens": result["usage"]["total_tokens"],
+                        "total_tokens": result["usage"]["total_tokens"],
+                    },
+                    "role": "assistant",
+                    **tool_calls_params,
+                }
+            )
 
     async def __aexit__(self, exc_type, exc, tb):
         await self.http_client.aclose()

{vectorvein-0.1.3 → vectorvein-0.1.5}/src/vectorvein/chat_clients/openai_compatible_client.py

@@ -2,7 +2,6 @@
 # @Date:   2024-07-26 14:48:55
 import json
 import random
-from typing import Union, AsyncGenerator
 
 from openai._types import NotGiven, NOT_GIVEN
 from openai._streaming import Stream, AsyncStream
@@ -18,6 +17,7 @@ from .utils import (
 from ..settings import settings
 from ..types import defaults as defs
 from ..types.enums import ContextLengthControlType, BackendType
+from ..types.llm_parameters import ChatCompletionMessage, ChatCompletionDeltaMessage
 
 
 class OpenAICompatibleChatClient(BaseChatClient):
@@ -122,21 +122,21 @@ class OpenAICompatibleChatClient(BaseChatClient):
                     if not chunk.choices[0].delta:
                         continue
                     if self.model_setting.function_call_available:
-                        yield chunk.choices[0].delta.model_dump()
+                        yield ChatCompletionDeltaMessage(**chunk.choices[0].delta.model_dump())
                     else:
                         message = chunk.choices[0].delta.model_dump()
                         full_content += message["content"] if message["content"] else ""
                         if tools:
-                            tool_call_data = ToolCallContentProcessor(result["content"]).tool_calls
+                            tool_call_data = ToolCallContentProcessor(full_content).tool_calls
                             if tool_call_data:
                                 message["tool_calls"] = tool_call_data["tool_calls"]
                         if full_content in ("<", "<|", "<|▶", "<|▶|") or full_content.startswith("<|▶|>"):
                             message["content"] = ""
                             result = message
                             continue
-                        yield message
+                        yield ChatCompletionDeltaMessage(**message)
                 if result:
-                    yield result
+                    yield ChatCompletionDeltaMessage(**result)
 
             return generator()
         else:
@@ -155,7 +155,8 @@ class OpenAICompatibleChatClient(BaseChatClient):
                     if tool_call_data:
                         result["tool_calls"] = tool_call_data["tool_calls"]
                         result["content"] = tool_call_content_processor.non_tool_content
-            return result
+
+            return ChatCompletionMessage(**result)
 
 
 class AsyncOpenAICompatibleChatClient(BaseAsyncChatClient):
@@ -191,7 +192,7 @@ class AsyncOpenAICompatibleChatClient(BaseAsyncChatClient):
         max_tokens: int = 2000,
         tools: list | NotGiven = NOT_GIVEN,
         tool_choice: str | NotGiven = NOT_GIVEN,
-    ) -> Union[AsyncGenerator[str, None], str]:
+    ):
         if model is not None:
             self.model = model
         if stream is not None:
@@ -260,21 +261,21 @@ class AsyncOpenAICompatibleChatClient(BaseAsyncChatClient):
                     if not chunk.choices[0].delta:
                         continue
                     if self.model_setting.function_call_available:
-                        yield chunk.choices[0].delta.model_dump()
+                        yield ChatCompletionDeltaMessage(**chunk.choices[0].delta.model_dump())
                     else:
                         message = chunk.choices[0].delta.model_dump()
                         full_content += message["content"] if message["content"] else ""
                         if tools:
-                            tool_call_data = ToolCallContentProcessor(result["content"]).tool_calls
+                            tool_call_data = ToolCallContentProcessor(full_content).tool_calls
                             if tool_call_data:
                                 message["tool_calls"] = tool_call_data["tool_calls"]
                         if full_content in ("<", "<|", "<|▶", "<|▶|") or full_content.startswith("<|▶|>"):
                             message["content"] = ""
                             result = message
                             continue
-                        yield message
+                        yield ChatCompletionDeltaMessage(**message)
                 if result:
-                    yield result
+                    yield ChatCompletionDeltaMessage(**result)
 
             return generator()
         else:
@@ -293,4 +294,4 @@ class AsyncOpenAICompatibleChatClient(BaseAsyncChatClient):
                     if tool_call_data:
                         result["tool_calls"] = tool_call_data["tool_calls"]
                         result["content"] = tool_call_content_processor.non_tool_content
-            return result
+            return ChatCompletionMessage(**result)

{vectorvein-0.1.3 → vectorvein-0.1.5}/src/vectorvein/chat_clients/utils.py

@@ -44,7 +44,7 @@ class ToolCallContentProcessor:
                 "tool_calls": [
                     {
                         "index": 0,
-                        "id": "fc1",
+                        "id": "call_0",
                         "function": {
                             "arguments": arguments,
                             "name": tool_call_data["name"],

{vectorvein-0.1.3 → vectorvein-0.1.5}/src/vectorvein/types/llm_parameters.py

@@ -4,6 +4,7 @@ from typing import List, Dict, Optional
 
 from pydantic import BaseModel, Field
 from openai.types.chat.chat_completion_message import ChatCompletionMessageToolCall
+from openai.types.chat.chat_completion_chunk import ChoiceDeltaToolCall
 
 from . import defaults as defs
 
@@ -58,7 +59,7 @@ class Usage(BaseModel):
     total_tokens: int
 
 
-class ModelOutput(BaseModel):
+class ChatCompletionMessage(BaseModel):
     content: Optional[str] = None
 
     tool_calls: Optional[List[ChatCompletionMessageToolCall]] = None
@@ -67,3 +68,13 @@ class ModelOutput(BaseModel):
     function_call_arguments: Optional[dict] = None
 
     usage: Optional[Usage] = None
+
+class ChatCompletionDeltaMessage(BaseModel):
+    content: Optional[str] = None
+
+    tool_calls: Optional[List[ChoiceDeltaToolCall]] = None
+    """The tool calls generated by the model, such as function calls."""
+
+    function_call_arguments: Optional[dict] = None
+
+    usage: Optional[Usage] = None

{vectorvein-0.1.3 → vectorvein-0.1.5}/tests/sample_settings.py

@@ -464,4 +464,38 @@ sample_settings = {
             "endpoints": ["zhipuai-default"],
         },
     },
+    "qwen_models": {
+        "qwen1.5-1.8b-chat": {
+            "id": "Qwen/Qwen1.5-1.8B-Chat",
+            "endpoints": ["together-default"],
+        },
+        "qwen1.5-4b-chat": {
+            "id": "Qwen/Qwen1.5-4B-Chat",
+            "endpoints": ["together-default"],
+        },
+        "qwen1.5-7b-chat": {
+            "id": "Qwen/Qwen1.5-7B-Chat",
+            "endpoints": ["together-default"],
+        },
+        "qwen1.5-14b-chat": {
+            "id": "Qwen/Qwen1.5-14B-Chat",
+            "endpoints": ["together-default"],
+        },
+        "qwen1.5-32b-chat": {
+            "id": "Qwen/Qwen1.5-32B-Chat",
+            "endpoints": ["together-default"],
+        },
+        "qwen1.5-72b-chat": {
+            "id": "Qwen/Qwen1.5-72B-Chat",
+            "endpoints": ["together-default"],
+        },
+        "qwen1.5-110b-chat": {
+            "id": "Qwen/Qwen1.5-110B-Chat",
+            "endpoints": ["together-default"],
+        },
+        "qwen2-72b-instruct": {
+            "id": "Qwen/Qwen2-72B-Instruct",
+            "endpoints": ["together-default"],
+        },
+    },
 }

{vectorvein-0.1.3 → vectorvein-0.1.5}/tests/test_create_chat_client.py

@@ -145,30 +145,36 @@ def test_sync(backend, model, stream: bool = False, use_tool: bool = False):
     else:
         messages = messages_simple
         tools_params = {}
-    response = client.create_completion(messages=format_messages(messages, backend=backend), **tools_params)
-    if stream:
+
+    if not stream:
+        response = client.create_completion(messages=format_messages(messages, backend=backend), **tools_params)
+        print(response)
+    else:
+        response = client.create_stream(messages=format_messages(messages, backend=backend), **tools_params)
         for chunk in response:
             print(chunk)
             print("=" * 20)
-    else:
-        print(response)
 
 
 async def test_async(backend, model, stream: bool = False, use_tool: bool = False):
-    client = create_async_chat_client(backend, model=model, stream=stream)
+    client = create_async_chat_client(backend, model=model)
     if use_tool:
         messages = messages_for_tools_simple
         tools_params = {"tools": tools_simple}
     else:
         messages = messages_simple
         tools_params = {}
-    response = await client.create_completion(messages=format_messages(messages, backend=backend), **tools_params)
-    if stream:
+
+    if not stream:
+        response = await client.create_completion(
+            messages=format_messages(messages, backend=backend), stream=False, **tools_params
+        )
+        print(response)
+    else:
+        response = await client.create_stream(messages=format_messages(messages, backend=backend), **tools_params)
         async for chunk in response:
             print(chunk)
             print("=" * 20)
-    else:
-        print(response)
 
 
 backend = BackendType.Moonshot
@@ -177,22 +183,29 @@ backend = BackendType.OpenAI
 model = "gpt-4o"
 backend = BackendType.Anthropic
 model = "claude-3-5-sonnet-20240620"
-backend = BackendType.MiniMax
-model = "abab6.5s-chat"
-backend = BackendType.Gemini
-model = "gemini-1.5-flash"
-backend = BackendType.OpenAI
-model = "gpt-35-turbo"
-backend = BackendType.Yi
-model = "yi-large-fc"
+# backend = BackendType.MiniMax
+# model = "abab6.5s-chat"
+# backend = BackendType.Gemini
+# model = "gemini-1.5-flash"
+# backend = BackendType.OpenAI
+# model = "gpt-35-turbo"
+# backend = BackendType.MiniMax
+# model = "abab6.5s-chat"
+# backend = BackendType.Yi
+# model = "yi-large-fc"
+# backend = BackendType.Mistral
+# model = "mixtral-8x7b"
+backend = BackendType.Qwen
+model = "qwen2-72b-instruct"
+
 start_time = time.perf_counter()
 # test_sync(backend=backend, model=model, stream=False, use_tool=False)
-test_sync(backend=backend, model=model, stream=False, use_tool=True)
+# test_sync(backend=backend, model=model, stream=False, use_tool=True)
 # test_sync(backend=backend, model=model, stream=True, use_tool=False)
 # test_sync(backend=backend, model=model, stream=True, use_tool=True)
 # asyncio.run(test_async(backend=backend, model=model, stream=False, use_tool=False))
 # asyncio.run(test_async(backend=backend, model=model, stream=False, use_tool=True))
 # asyncio.run(test_async(backend=backend, model=model, stream=True, use_tool=False))
-# asyncio.run(test_async(backend=backend, model=model, stream=True, use_tool=True))
+asyncio.run(test_async(backend=backend, model=model, stream=True, use_tool=True))
 end_time = time.perf_counter()
 print(f"Stream time elapsed: {end_time - start_time} seconds")

{vectorvein-0.1.3 → vectorvein-0.1.5}/tests/test_tool_use_multi_turns.py

@@ -1,7 +1,6 @@
 # @Author: Bi Ying
 # @Date:   2024-07-27 11:51:28
 import time
-import json
 
 from vectorvein.settings import settings
 from vectorvein.chat_clients import (