vectorvein 0.1.20__tar.gz → 0.1.22__tar.gz

{vectorvein-0.1.20 → vectorvein-0.1.22}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.1
 Name: vectorvein
-Version: 0.1.20
+Version: 0.1.22
 Summary: Default template for PDM package
 Author-Email: Anderson <andersonby@163.com>
 License: MIT

{vectorvein-0.1.20 → vectorvein-0.1.22}/pyproject.toml

@@ -16,7 +16,7 @@ description = "Default template for PDM package"
 name = "vectorvein"
 readme = "README.md"
 requires-python = ">=3.10"
-version = "0.1.20"
+version = "0.1.22"
 
 [project.license]
 text = "MIT"

{vectorvein-0.1.20 → vectorvein-0.1.22}/src/vectorvein/chat_clients/base_client.py

@@ -59,6 +59,7 @@ class BaseChatClient(ABC):
         max_tokens: int | None = None,
         tools: list | NotGiven = NOT_GIVEN,
         tool_choice: str | NotGiven = NOT_GIVEN,
+        response_format: dict | None = None,
         **kwargs,
     ) -> ChatCompletionMessage | Generator[ChatCompletionDeltaMessage, Any, None]:
         pass
@@ -130,6 +131,7 @@ class BaseAsyncChatClient(ABC):
         max_tokens: int | None = None,
         tools: list | NotGiven = NOT_GIVEN,
         tool_choice: str | NotGiven = NOT_GIVEN,
+        response_format: dict | None = None,
         **kwargs,
     ) -> ChatCompletionMessage | AsyncGenerator[ChatCompletionDeltaMessage, None]:
         pass

{vectorvein-0.1.20 → vectorvein-0.1.22}/src/vectorvein/chat_clients/gemini_client.py

@@ -52,6 +52,8 @@ class GeminiChatClient(BaseChatClient):
         max_tokens: int | None = None,
         tools: list | None = None,
         tool_choice: str | None = None,
+        response_format: dict | None = None,
+        **kwargs,
     ):
         if model is not None:
             self.model = model
@@ -81,6 +83,11 @@ class GeminiChatClient(BaseChatClient):
         else:
             tools_params = {}
 
+        if response_format:
+            response_format_params = {"generationConfig": {"response_mime_type": "application/json"}}
+        else:
+            response_format_params = {}
+
         if self.random_endpoint:
             self.random_endpoint = True
             self.endpoint_id = random.choice(self.backend_settings.models[self.model].endpoints)
@@ -99,6 +106,8 @@ class GeminiChatClient(BaseChatClient):
                 "maxOutputTokens": max_tokens,
             },
             **tools_params,
+            **response_format_params,
+            **kwargs,
         }
         if system_prompt:
             request_body["systemInstruction"] = {"parts": [{"text": system_prompt}]}
@@ -226,6 +235,8 @@ class AsyncGeminiChatClient(BaseAsyncChatClient):
         max_tokens: int | None = None,
         tools: list | None = None,
         tool_choice: str | None = None,
+        response_format: dict | None = None,
+        **kwargs,
     ):
         if model is not None:
             self.model = model
@@ -255,6 +266,11 @@ class AsyncGeminiChatClient(BaseAsyncChatClient):
         else:
             tools_params = {}
 
+        if response_format:
+            response_format_params = {"generationConfig": {"response_mime_type": "application/json"}}
+        else:
+            response_format_params = {}
+
         if self.random_endpoint:
             self.random_endpoint = True
             self.endpoint_id = random.choice(self.backend_settings.models[self.model].endpoints)
@@ -273,6 +289,8 @@ class AsyncGeminiChatClient(BaseAsyncChatClient):
                 "maxOutputTokens": max_tokens,
             },
             **tools_params,
+            **response_format_params,
+            **kwargs,
         }
         if system_prompt:
             request_body["systemInstruction"] = {"parts": [{"text": system_prompt}]}