diff options
Diffstat (limited to 'agl_service_voiceagent/servicers/voice_agent_servicer.py')
| -rw-r--r-- | agl_service_voiceagent/servicers/voice_agent_servicer.py | 31 |
1 files changed, 21 insertions, 10 deletions
diff --git a/agl_service_voiceagent/servicers/voice_agent_servicer.py b/agl_service_voiceagent/servicers/voice_agent_servicer.py index 2a4de33..c149b6d 100644 --- a/agl_service_voiceagent/servicers/voice_agent_servicer.py +++ b/agl_service_voiceagent/servicers/voice_agent_servicer.py @@ -199,6 +199,7 @@ class VoiceAgentServicer(voice_agent_pb2_grpc.VoiceAgentServiceServicer): version=self.service_version, status=True, wake_word=self.wake_word, + online_mode = self.online_mode ) # Convert the response object to a JSON string and log it @@ -280,8 +281,12 @@ class VoiceAgentServicer(voice_agent_pb2_grpc.VoiceAgentServiceServicer): "recorder": recorder, "audio_file": audio_file } - - recorder.start_recording() + + def record(): + recorder.start_recording() + + record_thread = threading.Thread(target=record) + record_thread.start() elif request.action == voice_agent_pb2.STOP: stream_uuid = request.stream_id @@ -294,6 +299,7 @@ class VoiceAgentServicer(voice_agent_pb2_grpc.VoiceAgentServiceServicer): recorder = self.rvc_stream_uuids[stream_uuid]["recorder"] audio_file = self.rvc_stream_uuids[stream_uuid]["audio_file"] del self.rvc_stream_uuids[stream_uuid] + print(use_online_mode) recorder.stop_recording() @@ -316,12 +322,19 @@ class VoiceAgentServicer(voice_agent_pb2_grpc.VoiceAgentServiceServicer): recognizer_uuid = self.stt_model.setup_vosk_recognizer() stt = self.stt_model.recognize_from_file(recognizer_uuid, audio_file,stt_framework=stt_framework) used_kaldi = True - print(stt) if stt not in ["FILE_NOT_FOUND", "FILE_FORMAT_INVALID", "VOICE_NOT_RECOGNIZED", ""]: if request.nlu_model == voice_agent_pb2.SNIPS: - extracted_intent = self.snips_interface.extract_intent(stt) - intent, intent_actions = self.snips_interface.process_intent(extracted_intent) + try: + extracted_intent = self.snips_interface.extract_intent(stt) + except Exception as e: + print(e) + extracted_intent = "" + if extracted_intent != "": + intent, intent_actions = self.snips_interface.process_intent(extracted_intent) + else: + intent = "" + intent_actions = {} if not intent or intent == "": status = voice_agent_pb2.INTENT_NOT_RECOGNIZED @@ -346,14 +359,12 @@ class VoiceAgentServicer(voice_agent_pb2_grpc.VoiceAgentServiceServicer): status = voice_agent_pb2.NLU_MODEL_NOT_SUPPORTED else: - stt = "" status = voice_agent_pb2.VOICE_NOT_RECOGNIZED # cleanup the kaldi recognizer if used_kaldi: self.stt_model.cleanup_recognizer(recognizer_uuid) used_kaldi = False - # delete the audio file if not self.store_voice_command: delete_file(audio_file) @@ -516,7 +527,7 @@ class VoiceAgentServicer(voice_agent_pb2_grpc.VoiceAgentServiceServicer): exec_response = "Uh oh, I failed to stop the media." exec_status = voice_agent_pb2.EXEC_ERROR else: - exec_response = "Sorry, I failed to execute command against intent 'MediaControl'. Maybe try again with more specific instructions." + exec_response = "Sorry, I failed to execute command." exec_status = voice_agent_pb2.EXEC_ERROR @@ -572,7 +583,7 @@ class VoiceAgentServicer(voice_agent_pb2_grpc.VoiceAgentServiceServicer): if "value" in execution_item: value = execution_item["value"] if self.set_current_values(signal, value): - exec_response = f"Yay, I successfully updated the intent '{intent}' to value '{value}'." + exec_response = f"Yay, I successfully updated '{intent}' to value '{value}'." exec_status = voice_agent_pb2.EXEC_SUCCESS elif "factor" in execution_item: @@ -593,7 +604,7 @@ class VoiceAgentServicer(voice_agent_pb2_grpc.VoiceAgentServiceServicer): value = current_value - factor value = str(value) if self.set_current_values(signal, value): - exec_response = f"Yay, I successfully updated the intent '{intent}' to value '{value}'." + exec_response = f"Yay, I successfully updated '{intent}' to value '{value}'." exec_status = voice_agent_pb2.EXEC_SUCCESS else: |