diff options
Diffstat (limited to 'agl_service_voiceagent/client.py')
| -rw-r--r-- | agl_service_voiceagent/client.py | 67 |
1 files changed, 63 insertions, 4 deletions
diff --git a/agl_service_voiceagent/client.py b/agl_service_voiceagent/client.py index 88ef785..ee7bc52 100644 --- a/agl_service_voiceagent/client.py +++ b/agl_service_voiceagent/client.py @@ -14,16 +14,19 @@ # See the License for the specific language governing permissions and # limitations under the License. +import sys +sys.path.append("../") import time import grpc from agl_service_voiceagent.generated import voice_agent_pb2 from agl_service_voiceagent.generated import voice_agent_pb2_grpc -def run_client(server_address, server_port, action, mode, nlu_engine, recording_time): +def run_client(server_address, server_port, action, mode, nlu_engine, recording_time,stt_framework,online_mode): SERVER_URL = server_address + ":" + server_port nlu_engine = voice_agent_pb2.RASA if nlu_engine == "rasa" else voice_agent_pb2.SNIPS print("Starting Voice Agent Client...") print(f"Client connecting to URL: {SERVER_URL}") + print("STT Framework:", stt_framework) with grpc.insecure_channel(SERVER_URL) as channel: print("Press Ctrl+C to stop the client.") print("Voice Agent Client started!") @@ -51,18 +54,74 @@ def run_client(server_address, server_port, action, mode, nlu_engine, recording_ elif action == 'ExecuteVoiceCommand': if mode == 'auto': - raise ValueError("[-] Auto mode is not implemented yet.") + # raise ValueError("[-] Auto mode is not implemented yet.") + stub = voice_agent_pb2_grpc.VoiceAgentServiceStub(channel=channel) + stt_framework = voice_agent_pb2.VOSK if stt_framework == "vosk" else voice_agent_pb2.WHISPER + online_mode = voice_agent_pb2.ONLINE if online_mode == True else voice_agent_pb2.OFFLINE + while(True): + wake_request = voice_agent_pb2.Empty() + wake_results = stub.DetectWakeWord(wake_request) + wake_word_detected = False + for wake_result in wake_results: + print("Wake word status: ", wake_word_detected) + if wake_result.status: + print("Wake word status: ", wake_result.status) + wake_word_detected = True + break + print("Wake word detected: ", wake_word_detected) + if wake_word_detected: + print("[+] Wake Word detected! Recording voice command...") + record_start_request = voice_agent_pb2.RecognizeVoiceControl(action=voice_agent_pb2.START, nlu_model=nlu_engine, record_mode=voice_agent_pb2.MANUAL, stt_framework=stt_framework,online_mode=online_mode,) + response = stub.RecognizeVoiceCommand(iter([record_start_request])) + stream_id = response.stream_id + + time.sleep(recording_time) # pause here for the number of seconds passed by user or default 5 seconds + + record_stop_request = voice_agent_pb2.RecognizeVoiceControl(action=voice_agent_pb2.STOP, nlu_model=nlu_engine, record_mode=voice_agent_pb2.MANUAL, stream_id=stream_id,stt_framework=stt_framework,online_mode=online_mode,) + record_result = stub.RecognizeVoiceCommand(iter([record_stop_request])) + print("[+] Voice command recording ended!") + + status = "Uh oh! Status is unknown." + if record_result.status == voice_agent_pb2.REC_SUCCESS: + status = "Yay! Status is success." + elif record_result.status == voice_agent_pb2.VOICE_NOT_RECOGNIZED: + status = "Voice not recognized." + elif record_result.status == voice_agent_pb2.INTENT_NOT_RECOGNIZED: + status = "Intent not recognized." + + # Process the response + print("Status:", status) + print("Command:", record_result.command) + print("Intent:", record_result.intent) + intent_slots = [] + for slot in record_result.intent_slots: + print("Slot Name:", slot.name) + print("Slot Value:", slot.value) + i_slot = voice_agent_pb2.IntentSlot(name=slot.name, value=slot.value) + intent_slots.append(i_slot) + + if record_result.status == voice_agent_pb2.REC_SUCCESS: + print("[+] Executing voice command...") + exec_voice_command_request = voice_agent_pb2.ExecuteInput(intent=record_result.intent, intent_slots=intent_slots) + response = stub.ExecuteCommand(exec_voice_command_request) + print("Response:", response) + wake_word_detected = False + time.sleep(1) + + elif mode == 'manual': stub = voice_agent_pb2_grpc.VoiceAgentServiceStub(channel) + stt_framework = voice_agent_pb2.VOSK if stt_framework == "vosk" else voice_agent_pb2.WHISPER + online_mode = voice_agent_pb2.ONLINE if online_mode == True else voice_agent_pb2.OFFLINE print("[+] Recording voice command in manual mode...") - record_start_request = voice_agent_pb2.RecognizeVoiceControl(action=voice_agent_pb2.START, nlu_model=nlu_engine, record_mode=voice_agent_pb2.MANUAL) + record_start_request = voice_agent_pb2.RecognizeVoiceControl(action=voice_agent_pb2.START, nlu_model=nlu_engine, record_mode=voice_agent_pb2.MANUAL, stt_framework=stt_framework,online_mode=online_mode,) response = stub.RecognizeVoiceCommand(iter([record_start_request])) stream_id = response.stream_id time.sleep(recording_time) # pause here for the number of seconds passed by user or default 5 seconds - record_stop_request = voice_agent_pb2.RecognizeVoiceControl(action=voice_agent_pb2.STOP, nlu_model=nlu_engine, record_mode=voice_agent_pb2.MANUAL, stream_id=stream_id) + record_stop_request = voice_agent_pb2.RecognizeVoiceControl(action=voice_agent_pb2.STOP, nlu_model=nlu_engine, record_mode=voice_agent_pb2.MANUAL, stream_id=stream_id,stt_framework=stt_framework,online_mode=online_mode,) record_result = stub.RecognizeVoiceCommand(iter([record_stop_request])) print("[+] Voice command recording ended!") |