`aip_sdk.EvalTarget`

aip_sdk.EvalTarget

One evaluation target in a v2 config: where metrics run, plus what runs there.

partition_type is the granularity to score at — "session", "trace" or "span". partition_id names a saved partition selector, when the evaluation scores one. At least one of the two must be set on a target you author; EvalConfig.resolved_targets() also uses this model for the untargeted view of a v1 config, where both are None. span_kind narrows a span target to a single kind ("TOOL", "LLM", "AGENT", …) and is only valid when partition_type is "span".

metrics and metric_configs keep their exact v1 shapes — a flat name list, and a mapping keyed by bare metric name whose values stay free-form. Only the nesting under a target is new.

Instances are frozen: a target is a value object, and allowing assignment would let a caller set span_kind on a non-span target after validation had already passed.