import argparse
from openai import OpenAI

def get_args():
    parser = argparse.ArgumentParser()
    parser.add_argument("--ip", type=str, required=True)
    parser.add_argument("--port", type=int, default=8080)
    parser.add_argument("--text", "-t", type=str, required=False)
    parser.add_argument("--phonemes", "-p", type=str, required=False)
    parser.add_argument("--model", "-m", type=str, required=False, default="kokoro", choices=["kokoro",])
    parser.add_argument("--language", "-l", type=str, required=False, default="en", choices=["en", "zh", "ja"])
    parser.add_argument("--voice", "-v", type=str, required=False, default="jm_kumo")
    parser.add_argument("--speed", type=float, default=0.8)
    parser.add_argument("--response_format", "-f", type=str, default="wav")
    parser.add_argument("--output", type=str, default="output")
    return parser.parse_args()

args = get_args()
if not args.text and not args.phonemes:
    raise SystemExit("Either --text or --phonemes must be provided.")

client = OpenAI(
    base_url=f'http://{args.ip}:{args.port}/v1',
    api_key="dummy_key"
)

payload = dict(
    model=args.model,
    voice=args.voice,
    response_format=args.response_format,
    speed=args.speed,
    instructions=args.language
)
if args.phonemes:
    payload["phonemes"] = args.phonemes
else:
    payload["input"] = args.text

speech = client.audio.speech.create(
            **payload
        )

audio_file = args.output + '.' + args.response_format
with open(audio_file, "wb") as f:
    f.write(speech.content)

print(f"Saved result to {audio_file}")