vectorvein 0.1.9__tar.gz → 0.1.11__tar.gz

{vectorvein-0.1.9 → vectorvein-0.1.11}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.1
 Name: vectorvein
-Version: 0.1.9
+Version: 0.1.11
 Summary: Default template for PDM package
 Author-Email: Anderson <andersonby@163.com>
 License: MIT

{vectorvein-0.1.9 → vectorvein-0.1.11}/pyproject.toml

@@ -16,7 +16,7 @@ description = "Default template for PDM package"
 name = "vectorvein"
 readme = "README.md"
 requires-python = ">=3.10"
-version = "0.1.9"
+version = "0.1.11"
 
 [project.license]
 text = "MIT"

{vectorvein-0.1.9 → vectorvein-0.1.11}/src/vectorvein/chat_clients/anthropic_client.py

@@ -118,6 +118,7 @@ class AnthropicChatClient(BaseChatClient):
         max_tokens: int | None = None,
         tools: list | NotGiven = NOT_GIVEN,
         tool_choice: str | NotGiven = NOT_GIVEN,
+        **kwargs,
     ):
         if model is not None:
             self.model = model
@@ -210,6 +211,7 @@ class AnthropicChatClient(BaseChatClient):
             max_tokens=max_tokens,
             tools=tools_params,
             tool_choice=tool_choice,
+            **kwargs,
         )
 
         if self.stream:
@@ -318,6 +320,7 @@ class AsyncAnthropicChatClient(BaseAsyncChatClient):
         max_tokens: int | None = None,
         tools: list | NotGiven = NOT_GIVEN,
         tool_choice: str | NotGiven = NOT_GIVEN,
+        **kwargs,
     ):
         if model is not None:
             self.model = model
@@ -410,6 +413,7 @@ class AsyncAnthropicChatClient(BaseAsyncChatClient):
             max_tokens=max_tokens,
             tools=tools_params,
             tool_choice=tool_choice,
+            **kwargs,
         )
 
         if self.stream:

{vectorvein-0.1.9 → vectorvein-0.1.11}/src/vectorvein/chat_clients/gemini_client.py

@@ -146,9 +146,9 @@ class GeminiChatClient(BaseChatClient):
             result = {
                 "content": "",
                 "usage": {
-                    "prompt_tokens": response["usageMetadata"]["promptTokenCount"],
-                    "completion_tokens": response["usageMetadata"]["candidatesTokenCount"],
-                    "total_tokens": response["usageMetadata"]["totalTokenCount"],
+                    "prompt_tokens": response.get("usageMetadata", {}).get("promptTokenCount", 0),
+                    "completion_tokens": response.get("usageMetadata", {}).get("candidatesTokenCount", 0),
+                    "total_tokens": response.get("usageMetadata", {}).get("totalTokenCount", 0),
                 },
             }
             tool_calls = []
@@ -309,9 +309,9 @@ class AsyncGeminiChatClient(BaseAsyncChatClient):
                 result = {
                     "content": "",
                     "usage": {
-                        "prompt_tokens": response["usageMetadata"]["promptTokenCount"],
-                        "completion_tokens": response["usageMetadata"]["candidatesTokenCount"],
-                        "total_tokens": response["usageMetadata"]["totalTokenCount"],
+                        "prompt_tokens": response.get("usageMetadata", {}).get("promptTokenCount", 0),
+                        "completion_tokens": response.get("usageMetadata", {}).get("candidatesTokenCount", 0),
+                        "total_tokens": response.get("usageMetadata", {}).get("totalTokenCount", 0),
                     },
                 }
                 tool_calls = []

{vectorvein-0.1.9 → vectorvein-0.1.11}/src/vectorvein/chat_clients/minimax_client.py

@@ -25,7 +25,7 @@ def extract_tool_calls(response):
                         "index": index,
                         "id": tool_call["id"],
                         "function": tool_call["function"],
-                        "type": tool_call["type"],
+                        "type": "function",
                     }
                     for index, tool_call in enumerate(tool_calls)
                 ]
@@ -59,6 +59,7 @@ class MiniMaxChatClient(BaseChatClient):
             endpoint_id,
             **kwargs,
         )
+        self.http_client = httpx.Client()
 
     def create_completion(
         self,
@@ -69,6 +70,7 @@ class MiniMaxChatClient(BaseChatClient):
         max_tokens: int | None = None,
         tools: list | None = None,
         tool_choice: str = "auto",
+        **kwargs,
     ):
         if model is not None:
             self.model = model
@@ -134,43 +136,50 @@ class MiniMaxChatClient(BaseChatClient):
             "stream": self.stream,
             "mask_sensitive_info": False,
             **tools_params,
+            **kwargs,
         }
 
-        response = httpx.post(
-            url=self.url,
-            headers=self.headers,
-            json=request_body,
-            timeout=60,
-        )
-
         if self.stream:
 
             def generator():
-                for chunk in response.iter_lines():
-                    if chunk:
-                        chunk_data = json.loads(chunk[6:])
-                        if chunk_data["object"] != "chat.completion.chunk":
-                            continue
-                        tool_calls_params = extract_tool_calls(chunk_data)
-                        has_tool_calls = True if tool_calls_params else False
-                        if has_tool_calls:
-                            yield ChatCompletionDeltaMessage(
-                                **{
-                                    "content": chunk_data["choices"][0]["delta"].get("content"),
-                                    "role": "assistant",
-                                    **tool_calls_params,
-                                }
-                            )
-                        else:
-                            yield ChatCompletionDeltaMessage(
-                                **{
-                                    "content": chunk_data["choices"][0]["delta"]["content"],
-                                    "role": "assistant",
-                                }
-                            )
+                with self.http_client.stream(
+                    "POST",
+                    url=self.url,
+                    headers=self.headers,
+                    json=request_body,
+                    timeout=60,
+                ) as response:
+                    for chunk in response.iter_lines():
+                        if chunk:
+                            chunk_data = json.loads(chunk[6:])
+                            if chunk_data["object"] != "chat.completion.chunk":
+                                continue
+                            tool_calls_params = extract_tool_calls(chunk_data)
+                            has_tool_calls = True if tool_calls_params else False
+                            if has_tool_calls:
+                                yield ChatCompletionDeltaMessage(
+                                    **{
+                                        "content": chunk_data["choices"][0]["delta"].get("content"),
+                                        "role": "assistant",
+                                        **tool_calls_params,
+                                    }
+                                )
+                            else:
+                                yield ChatCompletionDeltaMessage(
+                                    **{
+                                        "content": chunk_data["choices"][0]["delta"]["content"],
+                                        "role": "assistant",
+                                    }
+                                )
 
             return generator()
         else:
+            response = httpx.post(
+                url=self.url,
+                headers=self.headers,
+                json=request_body,
+                timeout=60,
+            )
             result = response.json()
             tool_calls_params = extract_tool_calls(result)
             return ChatCompletionMessage(
@@ -221,6 +230,7 @@ class AsyncMiniMaxChatClient(BaseAsyncChatClient):
         max_tokens: int | None = None,
         tools: list | None = None,
         tool_choice: str = "auto",
+        **kwargs,
     ):
         if model is not None:
             self.model = model
@@ -284,6 +294,7 @@ class AsyncMiniMaxChatClient(BaseAsyncChatClient):
             "stream": self.stream,
             "mask_sensitive_info": False,
             **tools_params,
+            **kwargs,
         }
 
         if self.stream:

{vectorvein-0.1.9 → vectorvein-0.1.11}/src/vectorvein/chat_clients/openai_compatible_client.py

@@ -54,6 +54,7 @@ class OpenAICompatibleChatClient(BaseChatClient):
         max_tokens: int | None = None,
         tools: list | NotGiven = NOT_GIVEN,
         tool_choice: str | NotGiven = NOT_GIVEN,
+        **kwargs,
     ):
         if model is not None:
             self.model = model
@@ -120,6 +121,7 @@ class OpenAICompatibleChatClient(BaseChatClient):
             temperature=self.temperature,
             max_tokens=max_tokens,
             **tools_params,
+            **kwargs,
         )
 
         if self.stream:
@@ -207,6 +209,7 @@ class AsyncOpenAICompatibleChatClient(BaseAsyncChatClient):
         max_tokens: int | None = None,
         tools: list | NotGiven = NOT_GIVEN,
         tool_choice: str | NotGiven = NOT_GIVEN,
+        **kwargs,
     ):
         if model is not None:
             self.model = model
@@ -273,6 +276,7 @@ class AsyncOpenAICompatibleChatClient(BaseAsyncChatClient):
             temperature=self.temperature,
             max_tokens=max_tokens,
             **tools_params,
+            **kwargs,
         )
 
         if self.stream:

{vectorvein-0.1.9 → vectorvein-0.1.11}/src/vectorvein/types/defaults.py

@@ -213,6 +213,13 @@ ZHIPUAI_MODELS = {
         "response_format_available": False,
         "max_output_tokens": 4095,
     },
+    "glm-4-plus": {
+        "id": "glm-4-plus",
+        "context_length": 128000,
+        "function_call_available": True,
+        "response_format_available": False,
+        "max_output_tokens": 4095,
+    },
     "glm-4-0520": {
         "id": "glm-4-0520",
         "context_length": 128000,
@@ -255,6 +262,13 @@ ZHIPUAI_MODELS = {
         "response_format_available": False,
         "max_output_tokens": 1024,
     },
+    "glm-4v-plus": {
+        "id": "glm-4v-plus",
+        "context_length": 2000,
+        "function_call_available": False,
+        "response_format_available": False,
+        "max_output_tokens": 1024,
+    },
 }
 
 # Mistral models

{vectorvein-0.1.9 → vectorvein-0.1.11}/tests/test_create_chat_client.py

@@ -193,8 +193,8 @@ model = "deepseek-chat"
 # backend = BackendType.OpenAI
 # model = "gpt-4o"
 # model = "gpt-35-turbo"
-# backend = BackendType.MiniMax
-# model = "abab6.5s-chat"
+backend = BackendType.MiniMax
+model = "abab6.5s-chat"
 # backend = BackendType.Yi
 # model = "yi-large-fc"
 # model = "yi-large-turbo"
@@ -222,13 +222,13 @@ model = "deepseek-chat"
 # model = "mistral-large"
 
 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=True, use_tool=False)
+# 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=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=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))
 end_time = time.perf_counter()
 print(f"Stream time elapsed: {end_time - start_time} seconds")