app.py

import gradio as gr
import pandas as pd
import requests
import spaces

@spaces.GPU
def convert_parquet_to_jsonl(parquet_file_or_url):
    if parquet_file_or_url.startswith("http"):
        response = requests.get(parquet_file_or_url)
        parquet_file = response.content
    else:
        parquet_file = parquet_file_or_url.name
    df = pd.read_parquet(parquet_file)
    jsonl_data = df.to_json(orient='records', lines=True)
    return jsonl_data

theme = "Ytheme/Minecraft"

demo = gr.Interface(theme=theme,
    fn=convert_parquet_to_jsonl,
    inputs=[gr.File(label="Parquet File"), gr.Textbox(label="Parquet File URL")],
    outputs=[gr.Textbox(label="JSONL Output")],
    title="Parquet to JSONL Converter",
    description="Input a Parquet file by a downloadable link or file upload and convert it to JSONL format"
)

if __name__ == "__main__":
    demo.launch()