Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Add roadmap for 2024 #1253

Merged
merged 4 commits into from
Jan 19, 2024
Merged

Add roadmap for 2024 #1253

merged 4 commits into from
Jan 19, 2024

Conversation

maximlt
Copy link
Member

@maximlt maximlt commented Jan 10, 2024

I feel that without some sort of roadmap, hvPlot's development is only driven by whatever other development happens in the HoloViz ecosystem and outside, while its development should be more deliberate to gain more momentum.

So I'm suggesting this tentative, ambitious, and subjective (of course) roadmap for hvPlot for 2024. I'm not sure where the roadmap should eventually live If we manage to agree on its content. I'm opening this PR for others to be able to comment on it.

Copy link
Collaborator

@MarcSkovMadsen MarcSkovMadsen left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

I really like the proposed roadmap. Do we know if documentation or bug fixes are most important to our users?

Copy link
Member

@jbednar jbednar left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Looks great to me!

Comment on lines +20 to +23
- The API is not being documented.
- The API is not being enough demonstrated in a single place; users have to assemble examples from different pages.
- An implicit reliance on HoloViews' documentation (e.g. for the options) that itself isn't in great shape.
- The difficulty to run locally all or parts of the documentation (that relies on somewhat complex datasets and many libraries).
Copy link
Collaborator

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Maybe the discovery of available kwargs/options could be improved by using an LLM for hvplot like:
https://huggingface.co/spaces/ahuang11/tweak-mpl-chat

Copy link
Member Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Yes sure LLMs could help users in some ways! How exactly I'm not sure.

I've tried to keep the roadmap simple, it focuses on the boring but needed stuff I believe we have to address.

Copy link
Collaborator

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

I made a basic LLM chat bot for hvplot here: https://huggingface.co/spaces/ahuang11/tweak-hvplot-chat

it's missing a lot of docs because it was a pain to convert ipynb to markdown, but it has the basics of customization https://huggingface.co/spaces/ahuang11/tweak-hvplot-chat/tree/main/hvplot_docs

Copy link
Collaborator

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Hi @ahuang11

Did you consider using and improving the existing hvplot chat app? https://holoviz-topics.github.io/panel-chat-examples/openai/#chat-with-hvplot

Copy link
Collaborator

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

If I'm not mistaken I believe that one is limited by formatting the kwargs into JSON so openai can do function calling. The one I just implemented is to dump hvPlot/HoloViews docs into a directory, so it expands a bit into HoloViews and can use .opts (it could probably write hooks as well if I gave it the FAQ).

Copy link
Collaborator

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Ok. .opts would just be a supplemental function to the existing chat-with-hvplot app.

@maximlt
Copy link
Member Author

maximlt commented Jan 11, 2024

I really like the proposed roadmap. Do we know if documentation or bug fixes are most important to our users?

My gut feeling says the documentation.

@philippjfr
Copy link
Member

This looks great, thanks @maximlt. I agree with docs as the highest priority.

@maximlt
Copy link
Member Author

maximlt commented Jan 16, 2024

I'd love to see more/better interactivity options for network graph visualizations. More specifically, I want to define callback functions when a node is clicked. This way, I can return details about the node and additional visualizations/panel components. There is either a bug or the functionality isn't implemented properly as there are several unanswered questions regarding this in the discourse.

Hi @Cyb3r-Monk, thanks for your feedback. I think what you're asking for is too specific for the roadmap. What I can recommend you is to look for issues (bugs or feature requests) on https://github.com/holoviz/holoviews directly, since it's very likely what you're asking for would be best achieved/supported using HoloViews. If there are existing issues, please upvote and chime in. If not, please open a new issue with as much information as possible about your use case. Finally, the HoloViz projects are open-source projects, feel free to start working on fixing/improving them, I'm sure contributors will be happy to help you in the process.

@droumis droumis self-requested a review January 18, 2024 18:07
Copy link
Member

@droumis droumis left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

The roadmap looks great and I completely agree about docs, bug fixes, explorer, deprecation of .interactive, and improving the contribution process.

@maximlt maximlt merged commit ad29d30 into main Jan 19, 2024
1 check passed
@maximlt maximlt deleted the roadmap2024 branch January 19, 2024 16:05
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

Successfully merging this pull request may close these issues.

6 participants