Better handling of empty dataframe in gr.DataFrame

[abidlabs]

We were handling empty dataframes incorrectly in the gr.Dataframe --> we were turning them into single-row arrays with each cell being "".

This fixes that -- would appreciate if folks could test this out to make sure I didn't miss anything as the dataframe frontend code is quite complex.

Fixes: #5162

Test code:

import gradio as gr
import pandas as pd

df_orig = pd.DataFrame([], columns=['a', 'b', 'c'])

df_fill = pd.DataFrame([[1, 2, 3]], columns=['a', 'b', 'c'])


with gr.Blocks() as demo:
    df = gr.Dataframe(value=df_orig, interactive=True, row_count=(0, "dynamic"))
    df2 = gr.Dataframe(value=df_fill, interactive=True, row_count=(0, "dynamic"))
    btn = gr.Button()
    text = gr.Textbox()
    btn.click(fn=lambda x: x.shape, inputs=df, outputs=text)
demo.queue().launch()

Also adds some stories for gr.DataFrame to prevent regressions like #5236 and some stories (unrelatedly) for gr.Dropdown()

[gradio-pr-bot]

🪼 branch checks and previews

•	Name	Status	URL
	Spaces	ready!	Spaces preview
	Website	ready!	Website preview
	Storybook	ready!	Storybook preview
	Visual tests	all good!	Build review
🦄	Changes	detected!	Details

---

Install Gradio from this PR

pip install https://gradio-builds.s3.amazonaws.com/8bc5dded137d309c1cc3319456ebf1019d0a9236/gradio-3.40.1-py3-none-any.whl

Install Gradio Python Client from this PR

pip install "gradio-client @ git+https://github.com/gradio-app/gradio@8bc5dded137d309c1cc3319456ebf1019d0a9236#subdirectory=client/python"

[gradio-pr-bot]

🦄 change detected

This Pull Request includes changes to the following packages.

Package	Version
@gradio/dataframe	patch
gradio	patch

• Maintainers can select this checkbox to manually select packages to update.

With the following changelog entry.

Better handling of empty dataframe in gr.DataFrame

Maintainers or the PR author can modify the PR title to modify this entry.

Something isn't right?

• Maintainers can change the version label to modify the version bump.
• If the bot has failed to detect any changes, or if this pull request needs to update multiple packages to different versions or requires a more comprehensive changelog entry, maintainers can update the changelog file directly.

[pngwn]

This is interesting. So I'm not convinced this issue is actually a bug, it was by design. The reason some empty rows are rendered is so a user can click into the cells and start typing, tabbing, etc. With no rows present this isn't possible.

[abidlabs]

The reason some empty rows are rendered is so a user can click into the cells and start typing, tabbing, etc. With no rows present this isn't possible.

We shouldn't be mutating the data to achieve this. As it stands, if you pass in a dataframe with 0 rows into a gr.DataFrame and try to get it back, it will have a single row, which can cause problems for downstream processing.

I don't think the UX issue is really there either, since we include a prominent "Add Row" button underneath the dataframe (as long as interactive=True).

[pngwn]

Yeah, that makes sense.

[pngwn]

Looks good. Thanks for fixing this @abidlabs!

[abidlabs]

Thanks @pngwn!