Currently there's no implementation of tool calling in OpenAIHandler. We need to rework that class so as to reflect the idea of an LLM's output being fit to a child's coroutine. Note:
The fitting depends un forwarding of results being enabled in the current node.
Since a node can have multiple children, a single node can result in multiple tool call fits.
Currently there's no implementation of tool calling in
OpenAIHandler
. We need to rework that class so as to reflect the idea of an LLM's output being fit to achild's
coroutine. Note: