Panel 1.0 Release Candidate

We are thrilled to announce the availability of the Panel 1.0 release candidate for general testing! We want to extend our thanks for your patience as we’ve been working tirelessly to adapt to major changes in Bokeh 3.x, improving the API and polishing our user experience. This significant undertaking has paved the way for a new, performant layout engine and greater customizability by exposing CSS stylesheets APIs directly in Python. We believe this release represents a major step forward in customizability, performance, and usability.

We kindly ask for your help in testing the Panel 1.0 release candidate. The bottom up rewrite of the layout engine and component rendering may cause some disruption in the short term so we wanted to get these changes in your hands to gather feedback before the full release. To get the RC release today use:

pip install panel==1.0.0rc1
# OR
conda install -c pyviz/label/dev panel=1.0.0rc1

If you encounter problems please report them on our issue tracker or join the discussion on our Discourse and Discord servers.

To help you with the transition, we have prepared a comprehensive migration guide, which will provide you with a detailed walkthrough of the key updates, the rationale behind them, and the critical considerations to keep in mind when adapting your applications to this new version. Additionally please refer to the development documentation, which includes detailed how-to guides walking you through your learning journey.

To get you excited about the new release a host of new features in the Panel 1.0 release candidate:

  • Overhauled documentation based on the Diataxis framework, featuring a how-to guide for easier navigation
  • Runnable documentation directly in the docs, via JupyterLite, and an improved app gallery
  • Per-component theming and styling using Bootstrap and Material Design systems
  • The ability to bind parameters and reactive functions directly to component parameters
  • The ability to reuse sessions for fast time to initial render performance.
  • Enhanced Markdown rendering with markdown-it-py
  • A Vizzu pane for plots, featuring beautiful animated transitions
  • A FloatPanel layout offering a free-floating, draggable dock panel
  • A Swipe layout for before-and-after comparisons of any component(s)
  • A Switch toggle widget for added interactivity
  • BasicAuth for straightforward password-based authentication
  • Much, much more.

There are also still a few items left items we hope to wrap up before the final release:

  • We are still working on fully supporting ipywidgets rendering
  • Adding a number of missing how-to guides to improve your learning journey
  • Final theming/styling tweaks to ensure your applications look modern and polished
  • General bug fixes as reported by you

Your invaluable feedback will contribute to the most robust and powerful release of Panel yet. Thank you once again for your continued support, and we look forward to seeing the incredible applications you create with Panel 1.0!

Warm regards,

The Panel Team

17 Likes

Congratz on the relase(-candidate), really looking forward to it :slight_smile:

Am i right assuming that it can currently be not installed alongside hvplot because of the bokeh>3 upgrade?

If i try i get

  File "main.py", line 1, in <module>
    import hvplot.pandas
  File "venv\lib\site-packages\hvplot\__init__.py", line 69, in <module>
    from .converter import HoloViewsConverter
  File "venv\lib\site-packages\hvplot\converter.py", line 23, in <module>
    from holoviews.plotting.bokeh import OverlayPlot, colormap_generator
  File "venv\lib\site-packages\holoviews\plotting\bokeh\__init__.py", line 20, in <module>
    from .annotation import (
  File "venv\lib\site-packages\holoviews\plotting\bokeh\annotation.py", line 16, in <module>
    from bokeh.models.arrow_heads import OpenHead, NormalHead
ModuleNotFoundError: No module named 'bokeh.models.arrow_heads'

2 Likes

It should be possible to try it out also with hvPlot. But to get the right combination of packages I recommend you follow the description in the blog post first to install the Panel 1.0rc candidate first and then hvplot.

It might be helpful to check the dependencies in the rc branch or main that Panel is tested with to figure out which versions work with Panel if you are installing something independently. panel/setup.py at main · holoviz/panel (github.com)

2 Likes

I’ve been posting a bit about this on social media. If you like the post please help me share the message by commenting and reposting. Thanks.

3 Likes

@Stubatiger You’ll also need to install holoviews==1.16.0a5 from PyPI or the conda pyviz/label/dev channel.

3 Likes

Here are a few impressions from the RC

FloatPanel

float-panel

Faster Loading

faster-loading

Tooltips

tooltip

Live, interactive docs

Check out the dev docs

altair (1)

Swipe

Compare images, plots, tables and more

swipe

5 Likes

Great job Panel team, I will try and create an environment.yml that people can conda install to make the setup easier. I will add

  1. pandas
  2. xarray
  3. hvplot
  4. pydeck
  5. geoviews
  6. holoviews

As a starter, let me know if anyone already has such and env for Jupyter Notebook.

1 Like

Hi @StuckDuckF, Marc dropped a link earlier in the thread that should be helpful for choosing versions for those (or other) libraries: panel/setup.py at main · holoviz/panel · GitHub

1 Like

Hi Philippfr,

Thanks for this awesome release and the great new documentation. I’m new to Panel but this is all very exciting. It seems there might be a bug with this documentation page though: Make interactive data workflows — Panel v1.0.0rc6

screenshot showing errors below code snippets:

Indeed, thanks @MaxPower. We just kicked off a new build with a fix for that.

1 Like

Thanks @Marc for the tips.
I am very new to holoviz and very excited about Panel. I am working in Jupiterlab and getting same error since upgraded.

Preformatted text`ModuleNotFoundError: No module named 'bokeh.models.arrow_heads'

Which blog post are you referring to for following the description?
I’m trying to build a new virtual environment and start from scratch, because uninstalling and installing the holoviz caused the above error. I’m getting the impression that I need to carefully install each piece separately in right order. I’m trying to learn what the right order is to install all the pieces. If the list also mentions the right versions that would be extremely helpful.

1 Like

Hi @Mana

Welcome to the community.

To me the exception sounds like you have an older version of Bokeh installed.

I would recommend creating a new virtual environment and installkng the packages you need in one go, Then your installer will help you install packages that work together.

If it does not work for you please ask for help again.

1 Like

Suspect you are using holoviews alongside Panel in which case you will have to upgrade to holoviews==1.16.0

1 Like

It worked! :slight_smile:
Thanks for the tip and prompt response. Great to see the community support.

1 Like

I did install holoviews==1.16.0 and everything is working now. Thanks a lot.

1 Like