from transformers import pipeline
import gradio as gr
import os

# Load the ASR model from Hugging Face Hub
asr = pipeline(task="automatic-speech-recognition", model="openai/whisper-small")

# Define the transcription function
def transcribe_long_form(filepath):
    if filepath is None:
        return "No audio file provided, please upload a file or record one."
    output = asr(filepath)
    return output['text']

# Custom CSS to improve the interface
css = """
body { font-family: Arial, sans-serif; }
button { background-color: #4CAF50; color: white; border: none; padding: 10px 20px; }
"""

# Set up the Gradio interface
with gr.Blocks(css=css) as demo:
    gr.Markdown("### Audio Transcription Service")
    gr.Markdown("Upload an audio file or use your microphone to record one. Then press the 'Transcribe' button to see the transcription.")
    with gr.Tab("Transcribe Audio"):
        with gr.Row():
            audio_input = gr.Audio(sources=["microphone", "upload"], type="filepath", label="Upload or Record Audio")
            submit_button = gr.Button("Transcribe")
        transcription_output = gr.Textbox(label="Transcription", lines=10, placeholder="Your transcription will appear here...")

        submit_button.click(
            transcribe_long_form, 
            inputs=[audio_input],
            outputs=[transcription_output]
        )

# Launch the Gradio app
demo.launch(share=True, server_port=int(os.environ.get('PORT1', 7860)))  # Default port 7860 if PORT1 is not set