mirror of
https://github.com/hwchase17/langchain.git
synced 2026-06-09 10:17:00 +00:00
5fd1e67808
Changes:
- ToolCall, InvalidToolCall and ToolCallChunk can all accept a "type"
parameter now
- LLM integration packages add "type" to all the above
- Tool supports ToolCall inputs that have "type" specified
- Tool outputs ToolMessage when a ToolCall is passed as input
- Tools can separately specify ToolMessage.content and
ToolMessage.raw_output
- Tools emit events for validation errors (using on_tool_error and
on_tool_end)
Example:
```python
@tool("structured_api", response_format="content_and_raw_output")
def _mock_structured_tool_with_raw_output(
arg1: int, arg2: bool, arg3: Optional[dict] = None
) -> Tuple[str, dict]:
"""A Structured Tool"""
return f"{arg1} {arg2}", {"arg1": arg1, "arg2": arg2, "arg3": arg3}
def test_tool_call_input_tool_message_with_raw_output() -> None:
tool_call: Dict = {
"name": "structured_api",
"args": {"arg1": 1, "arg2": True, "arg3": {"img": "base64string..."}},
"id": "123",
"type": "tool_call",
}
expected = ToolMessage("1 True", raw_output=tool_call["args"], tool_call_id="123")
tool = _mock_structured_tool_with_raw_output
actual = tool.invoke(tool_call)
assert actual == expected
tool_call.pop("type")
with pytest.raises(ValidationError):
tool.invoke(tool_call)
actual_content = tool.invoke(tool_call["args"])
assert actual_content == expected.content
```
---------
Co-authored-by: Erick Friis <erick@langchain.dev>
langchain-azure-dynamic-sessions
This package contains the LangChain integration for Azure Container Apps dynamic sessions. You can use it to add a secure and scalable code interpreter to your agents.
Installation
pip install -U langchain-azure-dynamic-sessions
Usage
You first need to create an Azure Container Apps session pool and obtain its management endpoint. Then you can use the SessionsPythonREPLTool tool to give your agent the ability to execute Python code.
from langchain_azure_dynamic_sessions import SessionsPythonREPLTool
# get the management endpoint from the session pool in the Azure portal
tool = SessionsPythonREPLTool(pool_management_endpoint=POOL_MANAGEMENT_ENDPOINT)
prompt = hub.pull("hwchase17/react")
tools=[tool]
react_agent = create_react_agent(
llm=llm,
tools=tools,
prompt=prompt,
)
react_agent_executor = AgentExecutor(agent=react_agent, tools=tools, verbose=True, handle_parsing_errors=True)
react_agent_executor.invoke({"input": "What is the current time in Vancouver, Canada?"})
By default, the tool uses DefaultAzureCredential to authenticate with Azure. If you're using a user-assigned managed identity, you must set the AZURE_CLIENT_ID environment variable to the ID of the managed identity.