Recognizers¶

Recognizers are simple interfaces to speech recognition softwares

Base¶

class dobby.recognizers.Recognizer[source]¶

Bases: threading.Thread

Threaded Recognizer base class. A queue can be subscribed to the Recognizer and hence, receive recognized pyjulius.Sentence objects

publish(sentence)[source]¶

Publish a recognized sentence to all subscribers

Parameters:	sentence (pyjulius.Sentence) – the recognized sentence

Note

The type of the sentence parameter may change in the near future when a new Recognizer will be added

subscribe(subscriber)[source]¶

Add a queue to the Recognizer’s subscribers list. A subscriber will receive published pyjulius.Sentence objects

Parameters:	subscriber (Queue.Queue) – subscriber to append

unsubscribe(subscriber)[source]¶

Remove a queue from the subscribers list

Parameters:	subscriber (Queue.Queue) – subscriber to remove

class dobby.recognizers.julius.Julius(host, port, encoding, min_score)[source]¶

Julius Recognizer is based on Julius speech recognition engine. It uses pyjulius to connect to julius instance running in module mode

Parameters:	host (string) – host of the server port (integer) – port of the server encoding (string) – encoding used to decode server’s output min_score (float) – minimum score under which the recognition result will be ignored

run()[source]¶: Run the recognition and publish() the recognized pyjulius.Sentence objects