update agents streaming guide #1943
Conversation
|
Mintlify preview ID generated: preview-ccstre-1765920086-1fd0a8a |
| 1. Partial JSON as [tool calls](/oss/langchain/models#tool-calling) are generated | ||
| 2. The completed, parsed tool calls that are executed | ||
| To do this, apply both [`"messages"`](#llm-tokens) and [`"updates"`](#agent-progress) streaming modes. The `"messages"` streaming mode will include [message chunks](/oss/langchain/messages#streaming-and-chunks) from all LLM calls in the agent. The `"updates"` mode will include completed messages with tool calls before they are routed to tools for execution. |
There was a problem hiding this comment.
Could we avoid relying on updates for this information and show how to aggregate the tool message? There's no guarantee that the message comes from the same source?
| def _process_completed_message(source: str, message: AnyMessage) -> None: | ||
| if source == "model" and isinstance(message, AIMessage) and message.tool_calls: |
There was a problem hiding this comment.
I don't think readers will know what source="model" is , probably need a bit more context
| To handle human-in-the-loop [interrupts](/oss/langchain/human-in-the-loop), we build on the [above example](#streaming-tool-calls): | ||
| 1. We configure the agent with [human-in-the-loop middleware and a checkpointer](/oss/langchain/human-in-the-loop#configuring-interrupts) | ||
| 2. We collect interrupts generated during the `"updates"` stream mode |
There was a problem hiding this comment.
Do interrupts surface if we do only messages?
| def get_weather(city: str) -> str: | ||
| """Get weather for a given city.""" | ||
| ) | ||
| def _process_message_chunk(token: AIMessage) -> None: |
There was a problem hiding this comment.
Are we able to remove the process* helpers as they act like sinks into print but usually users will not want a sync but to rewrite into a different generator
| ```python | ||
| interrupts = [] | ||
| for stream_mode, data in agent.stream( | ||
| Command(resume={"decisions": decisions}), |
There was a problem hiding this comment.
We need to show how to resume using interrupt IDs, i.e. so user code supports responding to multiple interrupts in parallel (e.g. two agents running in parallel each using HIL)
No description provided.