app.py

import pandas as pd
import gradio as gr

# Load Parquet data
df = pd.read_parquet('audio_and_text.parquet')

# Define function to fetch audio and text by number or index
def fetch_record(index):
    index = int(index)
    if 0 <= index < len(df):
        record = df.iloc[index]
        number = record['number']
        diacritics_audio = (record['with_diacritics_audio'], "audio.wav")
        no_diacritics_audio = (record['no_diacritics_audio'], "audio.wav")
        diacritics_text = record['diacritics_text']
        no_diacritics_text = record['no_diacritics_text']

        return (f"Record Number: {number}",
                diacritics_audio, diacritics_text,
                no_diacritics_audio, no_diacritics_text)
    else:
        return ("Index out of range.", None, "", None, "")

# Gradio UI setup
with gr.Blocks() as demo:
    gr.Markdown("## 📚 Arabic Audio & Text Browser")
    
    index_input = gr.Number(label="Enter Record Index", value=0, precision=0)

    btn_fetch = gr.Button("Fetch Record")

    record_label = gr.Label(label="Record Info")
    
    with gr.Row():
        with gr.Column():
            diacritics_audio_player = gr.Audio(label="Audio (With Diacritics)")
            diacritics_text_output = gr.Textbox(label="Text (With Diacritics)", lines=5)

        with gr.Column():
            no_diacritics_audio_player = gr.Audio(label="Audio (No Diacritics)")
            no_diacritics_text_output = gr.Textbox(label="Text (No Diacritics)", lines=5)

    btn_fetch.click(fn=fetch_record, 
                    inputs=index_input,
                    outputs=[record_label, 
                             diacritics_audio_player, diacritics_text_output,
                             no_diacritics_audio_player, no_diacritics_text_output])

# Launch the app
demo.launch()