These docs are for version 1.x of Rasa Open Source.

Tracker

Trackers maintain the state of a dialogue between the assistant and the user in the form of conversation sessions. To learn more about how to configure the session behavior, check out the docs on Session configuration.

class rasa.core.trackers.DialogueStateTracker(sender_id, slots, max_event_history=None, sender_source=None)

Maintains the state of a conversation.

Methods

• applied_events()
  Returns all actions that should be applied - w/o reverted events.

  Return type: List[Event]

• as_dialogue()
  Return a Dialogue object containing all of the turns. This can be serialised and later used to recover the state of this tracker exactly.

Return type: Dialogue

• change_form_to(form_name)
  Activate or deactivate a form.

Return type: None

• clear_followup_action()
  Clears follow up action when it was executed.

Return type: None

• copy()
  Creates a duplicate of this tracker.

Return type: DialogueStateTracker

• current_slot_values()
  Return the currently set values of the slots.

Return type: Dict[str, Any]

• current_state(event_verbosity=<EventVerbosity.NONE: 1>)
  Return the current tracker state as an object.

Return type: Dict[str, Any]

• events_after_latest_restart()
  Return a list of events after the most recent restart.

Return type: List[Event]

• export_stories(e2e=False, include_source=False)
  Dump the tracker as a story in the Rasa Core story format. Returns the dumped tracker as a string.

Return type: str

• export_stories_to_file(export_path='debug.md')
  Dump the tracker as a story to a file.

Return type: None

• from_dict(sender_id, events_as_dict, slots=None, max_event_history=None)
  Create a tracker from dump. The dump should be an array of dumped events. When restoring the tracker, these events will be replayed to recreate the state.

Return type: DialogueStateTracker

• generate_all_prior_trackers()
  Returns a generator of the previous trackers of this tracker. The resulting array represents the trackers before each action.

Return type: Generator[DialogueStateTracker, None, None]

• get_last_event_for(event_type, action_names_to_exclude=None, skip=0)
  Gets the last event of a given type which was actually applied.

Return type: Optional[Event]

• get_latest_entity_values(entity_type, entity_role=None, entity_group=None)
  Get entity values found for the passed entity type and optional role and group in the latest message.

Return type: Iterator[str]

• get_latest_input_channel()
  Get the name of the input_channel of the latest UserUttered event.

Return type: Optional[str]

• get_slot(key)
  Retrieves the value of a slot.

Return type: Optional[Any]

• idx_after_latest_restart()
  Return the idx of the most recent restart in the list of events. If the conversation has not been restarted, 0 is returned.

Return type: int

• init_copy()
  Creates a new state tracker with the same initial values.

Return type: DialogueStateTracker

• is_paused()
  State whether the tracker is currently paused.

Return type: bool

• last_executed_action_has(name, skip=0)
  Returns whether last ActionExecuted event had a specific name.

Return type: bool

• past_states(domain)
  Generate the past states of this tracker based on the history.

Return type: deque

• recreate_from_dialogue(dialogue)
  Use a serialised Dialogue to update the tracker’s state. This uses the state as is persisted in a TrackerStore. If the tracker is blank before calling this method, the final state will be identical to the tracker from which the dialogue was created.

  Return type: None

• reject_action(action_name)
  Notify active form that it was rejected.

Return type: None

• replay_events()
  Update the tracker based on a list of events.

Return type: None

• set_form_validation(validate)
  Toggle form validation.

Return type: None

• set_latest_action_name(action_name)
  Set latest action name and reset form validation and rejection parameters.

Return type: None

• travel_back_in_time(target_time)
  Creates a new tracker with a state at a specific timestamp. A new tracker will be created and all events previous to the passed time stamp will be replayed. Events that occur exactly at the target time will be included.

Return type: DialogueStateTracker

• trigger_followup_action(action)
  Triggers another action following the execution of the current.

Return type: None

• update(event, domain=None)
  Modify the state of the tracker according to an Event.

Return type: None

👋 I can help you get started with Rasa and answer your technical questions.