taubench_env
diapason.evals.execution.taubench_env
¶
TauBench task environment — native Diapason agent in tau2 simulation.
Plugs Diapason's inference engine into tau2-bench's orchestrator as a
HalfDuplexAgent, so the multi-turn conversation loop, user simulator,
domain tools, database, and evaluation all come from tau2-bench while the
agent's LLM calls go through Diapason.
JarvisHalfDuplexAgent
¶
A tau2 HalfDuplexAgent backed by Diapason's inference engine.
Replaces tau2's built-in LLMAgent while keeping the same interface so the Orchestrator, UserSimulator, and evaluation work unchanged.
stop(*args, **kwargs)
¶
Cleanup hook called by orchestrator.
TauBenchTaskEnv
¶
Per-task environment for TauBench evaluation.
Creates an Diapason-powered agent, plugs it into tau2's orchestrator, runs the simulation, and stores results in record.metadata for the scorer.