app.py

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

@spaces.GPU
def convert_parquet_to_jsonl(parquet_file=None, parquet_url=None):
    if parquet_file is not None:
        df = pd.read_parquet(parquet_file.name)
    elif parquet_url is not None:
        response = requests.get(parquet_url)
        df = pd.read_parquet(response.content)
    else:
        raise ValueError("Either parquet_file or parquet_url must be provided")
    jsonl_data = df.to_json(orient='records', lines=True)
    with open("output.jsonl", "w") as f:
        f.write(jsonl_data)
    return "output.jsonl"

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

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