Page MenuHomePhabricator

TranscribusOCR selection in Wikisource fails
Closed, ResolvedPublicBUG REPORT

Description

Steps to replicate the issue (include links if applicable):

What happens?:
Scan records error 400 from the linked webpage indicates failure.

What should have happened instead?:
correct the text layer of the page.

Software version (skip for WMF-hosted wikis like Wikipedia):
Wikisource - MediaWiki 1.41.0-wmf.9 (bacc43a) 2023-05-15T21:19:43

Other information (browser name/version, screenshots, etc.):
Linux Mint Cinnamon, 21.1
Firefox 113.0.1

230520_OCR error.jpg (888×1 px, 279 KB)

Event Timeline

Ineuw renamed this task from TranscribusOCR selection in nWikisource fails to TranscribusOCR selection in Wikisource fails.May 21 2023, 12:46 AM
Ineuw updated the task description. (Show Details)

Hi @Ineuw, could you please associate one or more active project tags with this task (via the Add Action...Change Project Tags dropdown)? That will allow to see a task when looking at project workboards or searching for tasks in certain projects, and get notified about a task when watching a related project tag. Thanks! I'm adding Wikimedia OCR here.

I've reset the tool's Transkribus credentials, and it looks like the requests are working again.

We'll work on making this process more resilient. I've opened T337186 for this.

I've reset the tool's Transkribus credentials, and it looks like the requests are working again.

@Samwilson: In that case, is there more to do in this very task or should it be resolved?

Ineuw claimed this task.

Thanks, it functions now. Consider the issue closed.

Just as a mention, both Tesseract and Transkribus are very slow in general, but much better than Google quality. This comment applies when the Wikimedia servers function normally.