app.py

import gradio as gr
from datasets import load_dataset

def peek_schema_text():
    try:
        ds = load_dataset("MultimodalUniverse/jwst", split="train", streaming=True)
        first = next(iter(ds))
        keys = list(first.keys())
        lines = [f"Keys: {keys}", "Types:"]
        lines += [f"  {k}: {type(first[k]).__name__}" for k in keys]
        # show short previews if any text fields exist
        previews = [f"  {k}: {first[k][:200]}" for k in keys if isinstance(first[k], str) and first[k]]
        if previews:
            lines.append("Text previews:")
            lines += previews
        return "\n".join(lines)
    except Exception as e:
        return f"Error: {e}"

demo = gr.Interface(
    fn=peek_schema_text,
    inputs=None,
    outputs="text",
    title="JWST dataset schema peek",
    description="Shows keys and types for the first streamed record."
)

demo.launch()