Skip to content

fix(lightspeed): ux improvements and persist display mode preference #2180

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.

Sign up for GitHub

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

Jump to bottom
Open
debsmita1 wants to merge 2 commits into
base: main
Choose a base branch
from
Open

fix(lightspeed): ux improvements and persist display mode preference #2180

debsmita1 wants to merge 2 commits into from

Conversation

@debsmita1
Copy link
Member

@debsmita1 debsmita1 commented Jan 27, 2026 •
edited
Loading

Hey, I just made a Pull Request!

Fixes:
https://issues.redhat.com/browse/RHDHBUGS-2537

  • Updates the prompt message to be "Enter a prompt for Lightspeed"
  • Loads the input box with the current conversation name, when renaming the conversation
  • Removes the "Learn More" link from the button 'cause it goes nowhere. Also, removes the "Verify accuracy" title
  • Reduced the pinned icon size to be a bit smaller, so the pinned chats are aligned with the recent chats
  • Moved the tool calling to the top, and now it appears after the Show Thinking component
  • Sanitized the tool calling response inside the component to remove the backticks, using the PF's Message component here
  • Persisting the display mode preference in the user settings, like the other preferences

GIF/Screenshots:

Screenshot 2026-01-27 at 8 49 24 PM Screenshot 2026-01-27 at 8 49 40 PM Screenshot 2026-01-27 at 8 49 54 PM
Screen.Recording.2026-01-27.at.8.50.25.PM.mov
Screenshot 2026-01-27 at 8 54 05 PM

Uploading Screen Recording 2026-01-27 at 9.01.09 PM.mov…

✔️ Checklist

  • A changeset describing the change and affected packages. (more info)
  • Added or Updated documentation
  • Tests for new functionality and regression tests for bug fixes
  • Screenshots attached (for UI changes)

@rhdh-gh-app
Copy link

rhdh-gh-app bot commented Jan 27, 2026

Changed Packages

Package Name Package Path Changeset Bump Current Version
@red-hat-developer-hub/backstage-plugin-lightspeed workspaces/lightspeed/plugins/lightspeed patch v1.2.1

@debsmita1 debsmita1 requested review from aprilma419, christoph-jerolimov and its-mitesh-kumar and removed request for yangcao77 January 27, 2026 16:12
HusneShabbir
HusneShabbir reviewed Jan 27, 2026
View reviewed changes
Copy link
Contributor

@HusneShabbir HusneShabbir left a comment

Choose a reason for hiding this comment

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

All the changes look good to me, but could you please take a look at RHDHBUGS-2556?
The model's thinking content still leaks outside of the designated "Show Thinking" box, which was observed even after your changes.

Main Branch :

Screen.Recording.2026-01-28.at.1.11.35.AM.mov

Your PR:

Screen.Recording.2026-01-28.at.1.08.58.AM.mov

@aprilma419
Copy link

aprilma419 commented Jan 28, 2026 •
edited
Loading

+1 to the issue @HusneShabbir has brought up. The rest LGTM!

Just one comment on the chat naming when in the thinking mode - instead of showing "< think >Okay, let's see...", could we show a skeleton-style loading state for it, just like how Gemini handles it here?

Screenshot 2026-01-28 at 12 23 20

@debsmita1
Copy link
Member Author

debsmita1 commented Jan 28, 2026

All the changes look good to me, but could you please take a look at RHDHBUGS-2556? The model's thinking content still leaks outside of the designated "Show Thinking" box, which was observed even after your changes.

Main Branch :

Screen.Recording.2026-01-28.at.1.11.35.AM.mov
Your PR:

Screen.Recording.2026-01-28.at.1.08.58.AM.mov

@benwilcock @karthikjeeyar @rohitkrai03, we have this scenario where the model enters the thinking state again after the tool calling is made. I have checked, and this is very normal for models that support reasoning with tool calling.
How Gemini & Cursor work is that they suppress repeated reasoning

I can think of 2 possible approaches here:

  1. We show the thinking content only the first time the model enters the thinking state and suppress all intermediate reasoning
  2. We append any intermediate reasoning to the existing Show thinking component

@debsmita1
Copy link
Member Author

debsmita1 commented Jan 28, 2026

+1 to the issue @HusneShabbir has brought up. The rest LGTM!

Just one comment on the chat naming when in the thinking mode - instead of showing "< think >Okay, let's see...", could we show a skeleton-style loading state for it, just like how Gemini handles it here?

Screenshot 2026-01-28 at 12 23 20

@aprilma419, we have raised a bug for this. Discussion here: https://redhat-internal.slack.com/archives/C05KGRQLPLG/p1769513912115459?thread_ts=1769505375.146949&cid=C05KGRQLPLG

@sonarqubecloud
Copy link

sonarqubecloud bot commented Jan 28, 2026

Quality Gate Passed Quality Gate passed

Issues
2 New issues
0 Accepted issues

Measures
0 Security Hotspots
0.0% Coverage on New Code
0.0% Duplication on New Code

See analysis details on SonarQube Cloud

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Reviewers

@HusneShabbir HusneShabbir HusneShabbir left review comments

@karthikjeeyar karthikjeeyar Awaiting requested review from karthikjeeyar karthikjeeyar is a code owner

@rohitkrai03 rohitkrai03 Awaiting requested review from rohitkrai03 rohitkrai03 is a code owner

@christoph-jerolimov christoph-jerolimov Awaiting requested review from christoph-jerolimov

@its-mitesh-kumar its-mitesh-kumar Awaiting requested review from its-mitesh-kumar

@aprilma419 aprilma419 Awaiting requested review from aprilma419

At least 1 approving review is required to merge this pull request.

Assignees

No one assigned

Labels

None yet

Projects

None yet

Milestone

No milestone

Development

Successfully merging this pull request may close these issues.

3 participants

@debsmita1 @aprilma419 @HusneShabbir