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.

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

Google Docs: Unable to get accurate summarization of page 20 #37996

Closed
stephendonner opened this issue May 1, 2024 · 4 comments · Fixed by brave/brave-core#23489
Closed

Google Docs: Unable to get accurate summarization of page 20 #37996

stephendonner opened this issue May 1, 2024 · 4 comments · Fixed by brave/brave-core#23489
Assignees
@darkdh
Labels
browser-ai bug OS/Desktop priority/P2 A bad problem. We might uplift this to the next planned release. QA Pass-Win64 QA/Test-Plan-Specified QA/Yes regression release-notes/exclude
Milestone
1.68.x - Release

Comments

@stephendonner
Copy link

stephendonner commented May 1, 2024 •
edited
Loading

Description

Google Docs: Unable to get accurate summarization of page 20

Found while testing #36649

Steps to Reproduce

  1. install 1.67.56
  2. launch Brave
  3. load https://docs.google.com/document/d/1p5kTWB3o2Y99rspl1yV3a64bEsXmx39Rp01HhWpR-Ok/edit#heading=h.y1t7tbvgyce or a similar 22-page document (19 blank pages, 1 page with content, 1 more page with totally different content)
  4. click on Leo in the sidebar
  5. advance in the document to page 20
  6. click on Summarize this page
  7. wait

Actual result:

Screenshot 2024-05-01 at 9 44 58 AM

Expected result:

Screen Shot 2024-05-01 at 10 30 54 AM

Reproduces how often:

100%

Brave version (brave://version info)

Brave | 1.67.56 Chromium: 124.0.6367.91 (Official Build) nightly (x86_64)
-- | --
Revision | 5670584eb8a1807ab8ac1584fb099bd21ec1e381
OS | macOS Version 11.7.10 (Build 20G1427)

Version/Channel Information:

  • Can you reproduce this issue with the current release? No
  • Can you reproduce this issue with the beta channel? No
  • Can you reproduce this issue with the nightly channel? Yes

cc @darkdh @bbondy @petemill @mattmcalister @rebron @brave/qa-team
@darkdh darkdh self-assigned this May 1, 2024
@darkdh darkdh removed the regression label May 1, 2024
@darkdh
Copy link
Member

darkdh commented May 1, 2024

Not a regression, we couldn't even summarize more than one page before this PR.

@darkdh
Copy link
Member

darkdh commented May 1, 2024

The problem is print preview only gives us 15 pages of data so the extractor won't get a chance to process beyond 15 pages.
Print dialog doesn't have this issue, needs further investigation.

@darkdh
Copy link
Member

darkdh commented May 1, 2024

Not a regression, we couldn't even summarize more than one page before this PR.

Just saw this in steps
5. advance in the document to page 20

If we do this intentionally to make that page appear in viewport then it is a regression

@mattmcalister mattmcalister moved this to In Progress in Browser AI May 3, 2024
@darkdh darkdh mentioned this issue May 8, 2024
Merged
24 tasks
@mattmcalister mattmcalister added the priority/P2 A bad problem. We might uplift this to the next planned release. label May 8, 2024
@darkdh darkdh moved this from In Progress to In Review in Browser AI May 8, 2024
@github-project-automation github-project-automation bot moved this from In Review to Done in Browser AI May 9, 2024
@brave-builds brave-builds added this to the 1.68.x - Nightly milestone May 9, 2024
@stephendonner stephendonner mentioned this issue May 17, 2024
Closed
@stephendonner
Copy link
Author

Verified PASSED using

Brave | 1.68.85 Chromium: 126.0.6478.40 (Official Build) beta (64-bit)
-- | --
Revision | 8e76a669cf78bdfa933c8d569460092a01166588
OS | Windows 10 Version 22H2 (Build 19045.4474)

Confirmed Claude 3 Haiku correctly summarized the 20th page

Steps:

  1. installed 1.68.85
  2. launched Brave
  3. loaded https://docs.google.com/document/d/1p5kTWB3o2Y99rspl1yV3a64bEsXmx39Rp01HhWpR-Ok/edit
  4. clicked on Leo in the sidebar
  5. advanced to page 20
  6. clicked on Summarize this page with Claude 3 Haiku selected

image

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

@darkdh darkdh

Labels
browser-ai bug OS/Desktop priority/P2 A bad problem. We might uplift this to the next planned release. QA Pass-Win64 QA/Test-Plan-Specified QA/Yes regression release-notes/exclude
Projects
Browser AI
Status: Done
Milestone
1.68.x - Release
Development

Successfully merging a pull request may close this issue.

Make sure PrintPreviewExtraction emit beforeprint and afterprint correctly brave/brave-core
4 participants
@mattmcalister @stephendonner @darkdh @brave-builds