Warning: This document is for an old version of Rasa. The latest version is 1.2.3.

Agent

class rasa.core.agent.Agent(domain=None, policies=None, interpreter=None, generator=None, tracker_store=None, action_endpoint=None, fingerprint=None, model_directory=None, model_server=None, remote_storage=None)

The Agent class provides a convenient interface for the most important Rasa functionality.

This includes training, handling messages, loading a dialogue model, getting the next action, and handling a channel.

create_processor(preprocessor=None)

Instantiates a processor based on the set state of the agent.

Return type:MessageProcessor
execute_action(sender_id, action, output_channel, policy, confidence)

Handle a single message.

Return type:DialogueStateTracker
handle_channels(channels, http_port=5005, route='/webhooks/', cors=None)

Start a webserver attaching the input channels and handling msgs.

Return type:Sanic
handle_message(message, message_preprocessor=None, **kwargs)

Handle a single message.

Return type:Optional[List[Dict[str, Any]]]
handle_text(text_message, message_preprocessor=None, output_channel=None, sender_id='default')

Handle a single message.

If a message preprocessor is passed, the message will be passed to that function first and the return value is then used as the input for the dialogue engine.

The return value of this function depends on the output_channel. If the output channel is not set, set to None, or set to CollectingOutputChannel this function will return the messages the bot wants to respond.

Example:
>>> from rasa.core.agent import Agent
>>> from rasa.core.interpreter import RasaNLUInterpreter
>>> agent = Agent.load("examples/restaurantbot/models/current")
>>> await agent.handle_text("hello")
[u'how can I help you?']
Return type:

Optional[List[Dict[str, Any]]]

is_ready(allow_nlu_only=False)

Check if all necessary components are instantiated to use agent.

Parameters:allow_nlu_only – If True, consider the agent ready event if no policy ensemble is present.
classmethod load(model_path, interpreter=None, generator=None, tracker_store=None, action_endpoint=None, model_server=None, remote_storage=None)

Load a persisted model from the passed path.

Return type:Agent
load_data(training_resource, remove_duplicates=True, unique_last_num_states=None, augmentation_factor=20, tracker_limit=None, use_story_concatenation=True, debug_plots=False, exclusion_percentage=None)

Load training data from a resource.

Return type:List[DialogueStateTracker]
log_message(message, message_preprocessor=None, **kwargs)

Append a message to a dialogue - does not predict actions.

Return type:DialogueStateTracker
parse_message_using_nlu_interpreter(message_data, tracker=None)

Handles message text and intent payload input messages.

The return value of this function is parsed_data.

Parameters:
  • message_data (Text) – Contain the received message in text or intent payload format.
  • tracker (DialogueStateTracker) – Contains the tracker to be used by the interpreter.
Returns:

The parsed message.

Example

{ “text”: ‘/greet{“name”:”Rasa”}’, “intent”: {“name”: “greet”, “confidence”: 1.0}, “intent_ranking”: [{“name”: “greet”, “confidence”: 1.0}], “entities”: [{“entity”: “name”, “start”: 6, “end”: 21, “value”: “Rasa”}], }

Return type:

Dict[str, Any]

persist(model_path, dump_flattened_stories=False)

Persists this agent into a directory for later loading and usage.

Return type:None
predict_next(sender_id, **kwargs)

Handle a single message.

Return type:Optional[Dict[str, Any]]
toggle_memoization(activate)

Toggles the memoization on and off.

If a memoization policy is present in the ensemble, this will toggle the prediction of that policy. When set to False the Memoization policies present in the policy ensemble will not make any predictions. Hence, the prediction result from the ensemble always needs to come from a different policy (e.g. KerasPolicy). Useful to test prediction capabilities of an ensemble when ignoring memorized turns from the training data.

Return type:None
train(training_trackers, **kwargs)

Train the policies / policy ensemble using dialogue data from file.

Parameters:
  • training_trackers – trackers to train on
  • **kwargs – additional arguments passed to the underlying ML trainer (e.g. keras parameters)
Return type:

None