Version: 3.x



align_lists(predictions: List[Text], golds: List[Text]) -> Tuple[List[Text], List[Text]]

Align two lists trying to keep same elements at the same index.

If lists contain different items at some indices, the algorithm will try to find the best alignment and pad with None values where necessary.


actions_since_last_utterance(tracker: DialogueStateTracker) -> List[Text]

Extract all events after the most recent utterance from the user.


async replay_events(tracker: DialogueStateTracker, agent: "Agent") -> None

Take a tracker and replay the logged user utterances against an agent.

During replaying of the user utterances, the executed actions and events created by the agent are compared to the logged ones of the tracker that is getting replayed. If they differ, a warning is logged.

At the end, the tracker stored in the agent's tracker store for the same sender id will have quite the same state as the one that got replayed.


load_tracker_from_json(tracker_dump: Text, domain: Domain) -> DialogueStateTracker

Read the json dump from the file and instantiate a tracker it.