Share via

IntentRecognizer Class

  • Reference

In addition to performing speech-to-text recognition, the IntentRecognizer extracts structured information about the intent of the speaker.

Inheritance
Recognizer
IntentRecognizer

Constructor

IntentRecognizer(speech_config: SpeechConfig, audio_config: AudioConfig | None = None, intents: Iterable[Tuple[str | LanguageUnderstandingModel, str]] | None = None)

Parameters

Name Description
speech_config
Required

The config for the speech recognizer.
audio_config

The config for the audio input.

Default value: None
intents

Intents from an iterable over pairs of (model, intent_id) or (simple_phrase, intent_id) to be recognized.

Default value: None

Methods

add_all_intents

Adds all intents from the specified Language Understanding Model.
add_intent

Add an intent to the recognizer. There are different ways to do this:

  • add_intent(simple_phrase): Adds a simple phrase that may be spoken by the user, indicating a specific user intent.

  • add_intent(simple_phrase, intent_id): Adds a simple phrase that may be spoken by the user, indicating a specific user intent. Once recognized, the result's intent id will match the id supplied here.

  • add_intent(model, intent_name): Adds a single intent by name from the specified LanguageUnderstandingModel.

  • add_intent(model, intent_name, intent_id): Adds a single intent by name from the specified LanguageUnderstandingModel.

  • add_intent(trigger, intent_id): Adds the IntentTrigger specified. IntentTrigger.
add_intents

Add intents from an iterable over pairs of (model, intent_id) or (simple_phrase, intent_id).
recognize_once

Performs recognition in a blocking (synchronous) mode. Returns after a single utterance is recognized. The end of a single utterance is determined by listening for silence at the end or until a maximum of about 30 seconds of audio is processed. The task returns the recognition text as result. For long-running multi-utterance recognition, use start_continuous_recognition_async instead.
recognize_once_async

Performs recognition in a non-blocking (asynchronous) mode. This will recognize a single utterance. The end of a single utterance is determined by listening for silence at the end or until a maximum of about 30 seconds of audio is processed. For long-running multi-utterance recognition, use start_continuous_recognition_async instead.
start_continuous_recognition

Synchronously initiates continuous recognition operation. User has to connect to EventSignal to receive recognition results. Call stop_continuous_recognition_async to stop the recognition.
start_continuous_recognition_async

Asynchronously initiates continuous recognition operation. User has to connect to EventSignal to receive recognition results. Call stop_continuous_recognition_async to stop the recognition.
start_keyword_recognition

Synchronously configures the recognizer with the given keyword model. After calling this method, the recognizer is listening for the keyword to start the recognition. Call stop_keyword_recognition() to end the keyword initiated recognition.
start_keyword_recognition_async

Asynchronously configures the recognizer with the given keyword model. After calling this method, the recognizer is listening for the keyword to start the recognition. Call stop_keyword_recognition_async() to end the keyword initiated recognition.
stop_continuous_recognition

Synchronously terminates ongoing continuous recognition operation.
stop_continuous_recognition_async

Asynchronously terminates ongoing continuous recognition operation.
stop_keyword_recognition

Synchronously ends the keyword initiated recognition.
stop_keyword_recognition_async

Asynchronously ends the keyword initiated recognition.

add_all_intents

Adds all intents from the specified Language Understanding Model.

add_all_intents(model: LanguageUnderstandingModel)

Parameters

Name Description
model
Required

add_intent

Add an intent to the recognizer. There are different ways to do this:

  • add_intent(simple_phrase): Adds a simple phrase that may be spoken by the user, indicating a specific user intent.

  • add_intent(simple_phrase, intent_id): Adds a simple phrase that may be spoken by the user, indicating a specific user intent. Once recognized, the result's intent id will match the id supplied here.

  • add_intent(model, intent_name): Adds a single intent by name from the specified LanguageUnderstandingModel.

  • add_intent(model, intent_name, intent_id): Adds a single intent by name from the specified LanguageUnderstandingModel.

  • add_intent(trigger, intent_id): Adds the IntentTrigger specified. IntentTrigger.

add_intent(*args)

Parameters

Name Description
model
Required

The language understanding model containing the intent.
intent_name
Required

The name of the single intent to be included from the language understanding model.
simple_phrase
Required

The phrase corresponding to the intent.
intent_id
Required

A custom id string to be returned in the IntentRecognitionResult's intent_id property.
trigger
Required

The IntentTrigger corresponding to the intent.

add_intents

Add intents from an iterable over pairs of (model, intent_id) or (simple_phrase, intent_id).

add_intents(intents_iter: Iterable[Tuple[str | LanguageUnderstandingModel, str]])

Parameters

Name Description
intents
Required

Intents from an iterable over pairs of (model, intent_id) or (simple_phrase, intent_id) to be recognized.
intents_iter
Required

recognize_once

Performs recognition in a blocking (synchronous) mode. Returns after a single utterance is recognized. The end of a single utterance is determined by listening for silence at the end or until a maximum of about 30 seconds of audio is processed. The task returns the recognition text as result. For long-running multi-utterance recognition, use start_continuous_recognition_async instead.

recognize_once() -> IntentRecognitionResult

Returns

Type Description

The result value of the synchronous recognition.

recognize_once_async

Performs recognition in a non-blocking (asynchronous) mode. This will recognize a single utterance. The end of a single utterance is determined by listening for silence at the end or until a maximum of about 30 seconds of audio is processed. For long-running multi-utterance recognition, use start_continuous_recognition_async instead.

recognize_once_async() -> ResultFuture

Returns

Type Description

A future containing the result value of the asynchronous recognition.

start_continuous_recognition

Synchronously initiates continuous recognition operation. User has to connect to EventSignal to receive recognition results. Call stop_continuous_recognition_async to stop the recognition.

start_continuous_recognition()

start_continuous_recognition_async

Asynchronously initiates continuous recognition operation. User has to connect to EventSignal to receive recognition results. Call stop_continuous_recognition_async to stop the recognition.

start_continuous_recognition_async() -> ResultFuture

Returns

Type Description

A future that is fulfilled once recognition has been initialized.

start_keyword_recognition

Synchronously configures the recognizer with the given keyword model. After calling this method, the recognizer is listening for the keyword to start the recognition. Call stop_keyword_recognition() to end the keyword initiated recognition.

start_keyword_recognition(model: KeywordRecognitionModel)

Parameters

Name Description
model
Required

the keyword recognition model that specifies the keyword to be recognized.

start_keyword_recognition_async

Asynchronously configures the recognizer with the given keyword model. After calling this method, the recognizer is listening for the keyword to start the recognition. Call stop_keyword_recognition_async() to end the keyword initiated recognition.

start_keyword_recognition_async(model: KeywordRecognitionModel)

Parameters

Name Description
model
Required

the keyword recognition model that specifies the keyword to be recognized.

Returns

Type Description

A future that is fulfilled once recognition has been initialized.

stop_continuous_recognition

Synchronously terminates ongoing continuous recognition operation.

stop_continuous_recognition()

stop_continuous_recognition_async

Asynchronously terminates ongoing continuous recognition operation.

stop_continuous_recognition_async()

Returns

Type Description

A future that is fulfilled once recognition has been stopped.

stop_keyword_recognition

Synchronously ends the keyword initiated recognition.

stop_keyword_recognition()

stop_keyword_recognition_async

Asynchronously ends the keyword initiated recognition.

stop_keyword_recognition_async()

Returns

Type Description

A future that is fulfilled once recognition has been stopped.

Attributes

authorization_token

The authorization token that will be used for connecting to the service.

Note

The caller needs to ensure that the authorization token is valid. Before the

authorization token expires, the caller needs to refresh it by calling this setter with a

new valid token. Otherwise, the recognizer will encounter errors during recognition.

canceled

Signal for events containing canceled recognition results (indicating a recognition attempt that was canceled as a result or a direct cancellation request or, alternatively, a transport or protocol failure).

Callbacks connected to this signal are called with a IntentRecognitionCanceledEventArgs, instance as the single argument.

endpoint_id

The endpoint ID of a customized speech model that is used for recognition, or a custom voice model for speech synthesis.

properties

A collection of properties and their values defined for this Recognizer.

recognized

Signal for events containing final recognition results (indicating a successful recognition attempt).

Callbacks connected to this signal are called with a IntentRecognitionEventArgs instance as the single argument, dependent on the type of recognizer.

recognizing

Signal for events containing intermediate recognition results.

Callbacks connected to this signal are called with a IntentRecognitionEventArgs instance as the single argument.

session_started

Signal for events indicating the start of a recognition session (operation).

Callbacks connected to this signal are called with a SessionEventArgs instance as the single argument.

session_stopped

Signal for events indicating the end of a recognition session (operation).

Callbacks connected to this signal are called with a SessionEventArgs instance as the single argument.

speech_end_detected

Signal for events indicating the end of speech.

Callbacks connected to this signal are called with a RecognitionEventArgs instance as the single argument.

speech_start_detected

Signal for events indicating the start of speech.

Callbacks connected to this signal are called with a RecognitionEventArgs instance as the single argument.