import gradio as gr
import requests
import json

def extract_words(json_data):
    words = [item['word'] for item in json_data]
    return ''.join(words)

def transcribe_audio(file):
    api_key = "sk-proj-6J4EfDKmIduACsB38LXlT3BlbkFJSVTYDzfbVeAvbKHp1l1z"
    url = "https://api.openai.com/v1/audio/transcriptions"
    headers = {
        "Authorization": f"Bearer {api_key}"
    }
    files = {
        "file": file,
        "model": (None, "whisper-1"),
        "language": (None, "ja"),
        "response_format": (None, "verbose_json"),
        "timestamp_granularities": (None, "word")
    }
    
    response = requests.post(url, headers=headers, files=files)
    response_json = response.json()
    
    if "words" in response_json:
        return extract_words(response_json['words'])
    else:
        return "Error in transcription: " + response_json.get("error", {}).get("message", "Unknown error")

iface = gr.Interface(
    fn=transcribe_audio,
    inputs=gr.inputs.Audio(source="upload", type="file"),
    outputs="text",
    title="音声文字起こし",
    description="音声ファイルをアップロードして、文字起こしを表示します"
)

iface.launch()