How to fix plots and image color rendering on Notebooks?

[wschoi]

I am currently running dark mode for my Databricks Notebooks, and am using the "new UI" released a few days ago (May 2023) and the "New notebook editor."

Currently all plots (like matplotlib) are showing wrong colors.

For example, denoting:

```

...

plt.figure()

plt.axvline(x=0, color='red')

plt.axhline(y=0, color='red')

...

```

is resulting in teal colored axes. Images also seem to have their colors and hues mixed up.

Is there a way to fix this without giving up any of my current settings for the notebook environment?

[Debayan]

Hi, Could you please provide a screenshot of the same which will help us understand it better?

Please tag @Debayan​ with your next comment so that I will get notified. Thank you!

[wschoi]

@Debayan Mukherjee​ 

Thanks in advance.

Below is a very simple example of the issue.

I denote the colors "red" for the axes and "blue" for the actual plot, but end up with teal colored axes and a yellow plot.

This is the same for any type of non-text output in the notebook environment, including images, which is the main cause of concern as it can make modeling and other tasks a bit confusing.

Not sure if there is a setting I missed, or if this is some kind of an issue.

[Debayan]

@Wonseok Choi​ thanks for the details. Could you please confirm if you have followed any documents on the same? (https://docs.databricks.com/visualizations/index.html)

[leonardoazzi]

Hello all,

Thank you @wschoi for reporting this issue. I've lost a lot of time trying to figure out if my image plotting was wrong.

[aleph_null]

Any updated on this issue? This is a huge drawback to use the dark theme

[andreapeterson]

Still seeing this issue in dark theme June 2025

[boztop]

Two years and counting, no solution to this obvious and very annoying problem so far.

[nikhilj0421]

Hi Team, I just checked this post and tried out the above simple plot graph with dark mode UI. It is working as expected for me. Could you please check now?

[andreapeterson]

Nope, just checked and the HTML on my notebook dashboard is still displaying wrong color in dark mode. works in light

[nikhilj0421]

@andreapeterson could you please share the screenshot along with the code?

[andreapeterson]

Sure, attached is a small snippit of the worksheet along with 2 pics of display of the code

displayHTML("""
  <link href="https://fonts.googleapis.com/css2?family=Mukta:wght@400;600&display=swap" rel="stylesheet">
  <div style="font-family: 'Mukta', sans-serif; background-color: #fafafa; border-left: 6px solid #4435e6; padding: 16px 24px; border-radius: 8px;">
    <h2 style="color: #080b15; margin-top: 0; font-weight: 600;">
      🔒 ML on Masked vs Unmasked Data with ALTR
    </h2>
    <p style="color: #4435e6; font-size: 16px; margin-bottom: 0;">
      We compared model accuracy between a user with masked data access and full unmasked access.
    </p>
  </div>
  <h2 style="
    font-family: 'Mukta', sans-serif;
    font-size: 20px;
    font-weight: 600;
    color: #4435e6;
    margin-top: 24px;
">
    Unmasked Training Run: <code style='color: #080b15;'>ml_privacy_team</code> group
</h2>
""")

[nikhilj0421]

Checking on it. 

[nikhilj0421]

@andreapeterson, this is likely because the CSS invert filter is applied to certain output types. We have an existing internal ticket(PROD-51027) for this issue, and being worked on. 

Either we have a plan to add a setting/toggle for this, which means either invert the colour or keep showing the white background while the site is in dark mode.

[andreapeterson]

Ah okay, thank you!!