VoiceReply.py

import requests
import os
import time
import urllib.parse
import random
from utils import read_config

_config = read_config()["llm"]
_SYSTEM_TEMPLATE   = _config.get("system_prompt", "")
_CHAR              = _config.get("char", "Eve")

def _build_system_prompt() -> str:
    """
    Substitute {char} into the system prompt template.
    """
    return _SYSTEM_TEMPLATE.replace("{char}", _CHAR)

def generate_voice_reply(prompt, model="openai-audio", voice="coral", audio_dir="."):
    """
    Generate an audio file (MP3) using the Pollinations text-to-speech API.
    
    The prompt is appended with a system message explaining that this is Eve,
    a WhatsApp bot that can generate images from Flux and help users generate images.
    To generate an image, the user should send a message like: /imagine {prompt}

    Parameters:
        prompt (str): The text prompt for audio generation.
        model (str): Model identifier (default: openai-audio).
        voice (str): Voice parameter (default: coral).
        audio_dir (str): Directory where the audio file will be saved.

    Returns:
        tuple: (audio_file_path (str), audio_data (bytes)) or None if failed.
    """
    print("DEBUG: Starting generate_voice_reply")
    print(f"DEBUG: Received prompt: {prompt}")
    os.makedirs(audio_dir, exist_ok=True)
    randomSeed = random.randint(0, 9999999)
    encoded_prompt = urllib.parse.quote(prompt)
    url = f"http://text.pollinations.ai/{encoded_prompt}?model={model}&voice={voice}&seed={randomSeed}"
    print(f"DEBUG: Fetching audio with URL: {url}")
    
    try:
        response = requests.get(url, timeout=30)
        print("DEBUG: Received response")
    except Exception as e:
        print(f"DEBUG: Error fetching audio: {e}")
        return None

    print(f"DEBUG: Response status code: {response.status_code}")
    if response.status_code != 200:
        print(f"DEBUG: Failed to fetch audio. Status code: {response.status_code} | Response text: {response.text}")
        return None

    audio_data = response.content
    timestamp = int(time.time())
    file_name = f"voice_reply_{timestamp}.mp3"
    audio_file_path = os.path.join(audio_dir, file_name)
    print(f"DEBUG: Saving audio to {audio_file_path}")

    try:
        with open(audio_file_path, "wb") as f:
            f.write(audio_data)
        print(f"DEBUG: Audio saved to {audio_file_path}")
    except Exception as e:
        print(f"DEBUG: Error saving audio file: {e}")
        return None

    return audio_file_path, audio_data

# if __name__ == "__main__":
#     # Example usage
#     prompt = "Hi. how are you."
#     audio_dir = "./audio_replies"
    
    
#     result = generate_voice_reply(prompt, audio_dir=audio_dir)
#     if result:
#         audio_file_path, audio_data = result
#         print(f"DEBUG: Generated audio file: {audio_file_path}")
#     else:
#         print("DEBUG: Failed to generate audio file.")