141 lines
4.3 KiB
Python
141 lines
4.3 KiB
Python
|
|
||
|
|
import os
|
||
|
|
from groq import Groq
|
||
|
|
import yaml
|
||
|
|
import gradio as gr
|
||
|
|
import json
|
||
|
|
import os
|
||
|
|
from openai import OpenAI
|
||
|
|
from dotenv import load_dotenv
|
||
|
|
from pathlib import Path
|
||
|
|
|
||
|
|
|
||
|
|
|
||
|
|
# 0. IMPORT ALL FILES!
|
||
|
|
import os
|
||
|
|
import sounddevice as sd
|
||
|
|
import numpy as np
|
||
|
|
import soundfile as sf
|
||
|
|
import queue
|
||
|
|
import sys
|
||
|
|
from scipy.io.wavfile import read
|
||
|
|
from faster_whisper import WhisperModel
|
||
|
|
|
||
|
|
|
||
|
|
load_dotenv()
|
||
|
|
openai_api_key = os.getenv("OPENAI_API_KEY")
|
||
|
|
groq_api_key = os.getenv("GROQ_API_KEY")
|
||
|
|
client_groq = Groq(api_key = groq_api_key)
|
||
|
|
|
||
|
|
|
||
|
|
def record_on_speech(output_file="conversation.wav", samplerate=44100, channels=1, silence_threshold=0.01, silence_duration=1, device=None):
|
||
|
|
"""
|
||
|
|
Records audio from the microphone, starting only when the user speaks and stopping after a period of silence.
|
||
|
|
|
||
|
|
Args:
|
||
|
|
output_file (str): Path to save the recorded audio.
|
||
|
|
samplerate (int): Sampling rate in Hz. Default is 44100.
|
||
|
|
channels (int): Number of audio channels. Default is 1 (mono).
|
||
|
|
silence_threshold (float): RMS threshold to detect silence. Default is 0.01 (normalized amplitude).
|
||
|
|
silence_duration (float): Duration in seconds of silence to stop recording. Default is 2.
|
||
|
|
device (int or str): Input device ID or name. Default is None (use system default).
|
||
|
|
|
||
|
|
Returns:
|
||
|
|
None
|
||
|
|
"""
|
||
|
|
|
||
|
|
if os.path.exists(output_file):
|
||
|
|
os.remove(output_file)
|
||
|
|
print(f"Existing file '{output_file}' was deleted.")
|
||
|
|
|
||
|
|
q = queue.Queue()
|
||
|
|
|
||
|
|
def callback(indata, frames, time, status):
|
||
|
|
"""Callback for audio input."""
|
||
|
|
if status:
|
||
|
|
print(status, file=sys.stderr)
|
||
|
|
q.put(indata.copy())
|
||
|
|
|
||
|
|
def rms_level(data):
|
||
|
|
"""Calculate the RMS level of the audio."""
|
||
|
|
return np.sqrt(np.mean(np.square(data)))
|
||
|
|
|
||
|
|
try:
|
||
|
|
# Open the sound file
|
||
|
|
with sf.SoundFile(output_file, mode='x', samplerate=samplerate,
|
||
|
|
channels=channels, subtype='PCM_16') as file:
|
||
|
|
with sd.InputStream(samplerate=samplerate, device=device,
|
||
|
|
channels=channels, callback=callback):
|
||
|
|
print("Listening for speech...")
|
||
|
|
silent_time = 0
|
||
|
|
recording_started = False
|
||
|
|
|
||
|
|
while True:
|
||
|
|
data = q.get()
|
||
|
|
rms = rms_level(data)
|
||
|
|
|
||
|
|
if not recording_started:
|
||
|
|
if rms > silence_threshold:
|
||
|
|
print("Voice detected, starting recording...")
|
||
|
|
recording_started = True
|
||
|
|
|
||
|
|
if recording_started:
|
||
|
|
file.write(data)
|
||
|
|
|
||
|
|
if rms < silence_threshold:
|
||
|
|
silent_time += len(data) / samplerate
|
||
|
|
else:
|
||
|
|
silent_time = 0
|
||
|
|
|
||
|
|
if silent_time >= silence_duration:
|
||
|
|
print("Silence detected, stopping recording...")
|
||
|
|
break
|
||
|
|
|
||
|
|
except KeyboardInterrupt:
|
||
|
|
print("\nRecording interrupted.")
|
||
|
|
except Exception as e:
|
||
|
|
print(f"Error: {type(e).__name__}: {e}", file=sys.stderr)
|
||
|
|
|
||
|
|
return output_file
|
||
|
|
|
||
|
|
|
||
|
|
def transcribe_audio_groq(aud_path = "conversation.wav"):
|
||
|
|
with open(aud_path, "rb") as file:
|
||
|
|
transcription = client_groq.audio.transcriptions.create(
|
||
|
|
file=(aud_path, file.read()),
|
||
|
|
model="whisper-large-v3-turbo",
|
||
|
|
response_format="verbose_json",
|
||
|
|
prompt="The following is a conversation between Riko and Rayen",
|
||
|
|
)
|
||
|
|
print(transcription.text)
|
||
|
|
|
||
|
|
return transcription.text
|
||
|
|
|
||
|
|
|
||
|
|
|
||
|
|
if __name__ == "__main__":
|
||
|
|
print('Running module')
|
||
|
|
|
||
|
|
conversation_recording = "~/riko_project_v1/conversation.wav"
|
||
|
|
conversation_recording = Path("audio") / "conversation.wav"
|
||
|
|
conversation_recording = str(conversation_recording)
|
||
|
|
|
||
|
|
record_on_speech(
|
||
|
|
output_file=conversation_recording,
|
||
|
|
samplerate=44100,
|
||
|
|
channels=1,
|
||
|
|
silence_threshold=0.02, # Adjust based on your microphone sensitivity
|
||
|
|
silence_duration=1, # Stop after 3 seconds of silence
|
||
|
|
device=None # Use default device, or specify by ID or name
|
||
|
|
)
|
||
|
|
|
||
|
|
user_spoken_text = transcribe_audio_groq(aud_path=conversation_recording)
|
||
|
|
print(user_spoken_text)
|
||
|
|
|
||
|
|
|
||
|
|
|
||
|
|
|
||
|
|
|
||
|
|
|
||
|
|
|
||
|
|
|