nlu.user_intents_checker¶
This file contains the main functions for checking the user intents.
Module Contents¶
Classes¶
Attributes¶
- nlu.user_intents_checker.PATTERN_BASIC¶
- nlu.user_intents_checker.PATTERN_DONT_CARE = ['anything', 'any', 'dont know', 'i do not care', 'i dont care', 'dont care', 'dontcare', 'it...¶
- nlu.user_intents_checker.PATTERN_QUESTION = ['who', 'what', 'when', 'which', 'can', 'could', 'is', 'are']¶
- nlu.user_intents_checker.PATTERN_DONT_LIKE = ['something else', 'anything else', 'dont like it', 'not this', 'another']¶
- nlu.user_intents_checker.PATTERN_DONT_WANT = ['dont', 'not', 'nothing', 'wont', 'shouldnt', 'dont need', 'dont want', 'no', 'not']¶
- nlu.user_intents_checker.PATTERN_WATCHED = ['watched', 'seen', 'yes', 'I have', 'yup', 'yea']¶
- class nlu.user_intents_checker.UserIntentsChecker(config: Dict[str, Any])¶
- is_dontcare(user_utterance: moviebot.core.utterance.utterance.UserUtterance) bool¶
Returns true if any keyword from dont care pattern is present.
TODO (Ivica Kostric): This can be partially merged with check_basic_intent. Regex can be dropped since we are only looking for complete tokens anyway.
- Parameters:
utterance – A processed user utterance.
- Returns:
True if the dont care pattern detected.
- check_basic_intent(user_utterance: moviebot.core.utterance.utterance.UserUtterance, intent: moviebot.core.intents.user_intents.UserIntents) List[moviebot.dialogue_manager.dialogue_act.DialogueAct]¶
Given intent and a list of intent patterns checks if any token in user utterance match the pattern.
- Parameters:
user_utterance – Class containing raw utterance and processed tokens.
intent – Intent for which to compare patterns.
- Returns:
If pattern exists returns that intents dialogue act.
- check_reveal_voluntary_intent(user_utterance: moviebot.core.utterance.utterance.UserUtterance) List[moviebot.dialogue_manager.dialogue_act.DialogueAct]¶
Finds any voluntary revealed information in utterance.
- Parameters:
utterance – Processed user utterance.
- Returns:
A list of dialogue acts with revealed information.
- check_reveal_intent(user_utterance, last_agent_dact) List[moviebot.dialogue_manager.dialogue_act.DialogueAct]¶
Checks utterance for intent “reveal”.
This function is only called if the intent of agent is ELicit to see if user has answered the query.
- Parameters:
utterance – Processed user utterance.
last_agent_dact – Last agent dialogue act.
- Returns:
A list of dialogue acts.
- check_reject_intent(user_utterance: moviebot.core.utterance.utterance.UserUtterance) List[moviebot.dialogue_manager.dialogue_act.DialogueAct]¶
Checks utterance for intent “reject”.
- Parameters:
utterance – Processed user utterance.
- Returns:
A list of dialogue acts.
- check_inquire_intent(user_utterance: moviebot.core.utterance.utterance.UserUtterance) List[moviebot.dialogue_manager.dialogue_act.DialogueAct]¶
Checks utterance for intent “inquire”.
- Parameters:
utterance – Processed user utterance.
- Returns:
A list of dialogue acts.