What is the title of that page? Can you add a link to it? Thanks!

It looks like the downtimes began around April 12, and commit 1ba22eb9083f53c1118175648941a702e80b2a15 was the first major commit recently before that, on March 29.

This can of course always be overridden locally in .env.local.

PR: https://github.com/wikimedia/wikimedia-ocr/pull/31

It's deployed (demo) but there's a CSP error:

I've scheduled this for the European mid-day backport window, ~3 hours from now.

Also: https://www.mediawiki.org/wiki/Extension:WikiEditor/Toolbar_customization#Determining_when_toolbar_load_is_done suggests not to use wikiEditor-toolbar-doneInitialSections.

https://tesseract-ocr.github.io/tessdoc/Data-Files suggests that these are not duplicates, but e.g. ben is Bengali and should only be listed once.

I assume we'll want to map Tesseract's supported language list to return ISO 639-1, which is what we use on-wiki.

The basic UI for this has already been done by @MusikAnimal (as part of T280213). See it here: https://ocr-test.wmcloud.org/?engine=tesseract

No, in the MVP T280848 only step 1 is done, the disabling (it actually also disables the button; is that bad?).

In T280617#7058055, @ifried wrote:

Pinging @Samwilson, since I know Harumi is busy with some Editing team work :)

If we wanted to include support for the experimental & mapped languages, would this be possible? Could we create a ticket for it? Any thoughts/concerns?

The feature flag still needs to be removed.

PR for ebook-convert: https://github.com/wikimedia/ws-export/pull/364

This is pretty much done, and is working at https://ocr.wmcloud.org/api.php

Done.

Sounds sensible. PR: https://github.com/wikimedia/ws-export/pull/363

Given that we're recording the book-generation times but not displaying them anywhere, this ticket seems like it might still be valid. (Not that Community-Tech is going to work on it.)

Here's a task about cleaning up Google's structured output into wikitext: T250185: Make Wikisource-OCR handle paragraphs better

PR: https://github.com/wikimedia/wikimedia-ocr/pull/26

The tool work for this has been done in T280214.

It's definitely possible (and pretty simple) to add a link with prefilled image URL and language, as well as a backlink (probably as an extra, new, URL parameter).

I think we need to also install, before building: https://github.com/wikimedia/ToolforgeBundle/pull/51

Better UI is merged (single input with tag-style language codes). Live on the test site (only, so far). @dom_walden, not sure if you want to have another look at this.

FWIW, my experience with Phe's OCR tool (whose gadget just sets .disable() on #wpTextBox1) is that this is not particularly intuitive for end users.

Ready for review: https://gerrit.wikimedia.org/r/c/mediawiki/extensions/Wikisource/+/682034 (dependent on https://github.com/wikimedia/wikimedia-ocr/pull/24 ).

This sounds similar to the 2020 wishlist proposal Insert attestation using Wikisource as a corpus, which references:

• T139152: Add a feature to allow searching and importing/copying content from sister projects
• T157802: Search relevant sentences from Wikisource to illustrate the Wiktionary

Merged, and released in 1.4.0 (of the bundle).

PR for review: https://github.com/wikimedia/wikimedia-ocr/pull/24

There are actually two sets of host names (one list and one single one) that should be configurable:

Can someone walk me through the flow for this, a bit confused about what hosting it on the extension means. Do we have mocks for this extension UI and if so can folks direct me to them?

I've started looking at this and have a basic functioning system with two buttons in the tool bar (one for each engine).

Closing. Instead of this, we're going to go directly to the new design (in T280848).

PR to disable multithreading: https://github.com/wikimedia/wikimedia-ocr/pull/17

The remaining issues here look like they're to do with font support. I've installed the Culmus fonts (as recommended here) and with, for example, Yehuda CLM the above page looks like this:

This will include the Fraktur OCR button for German Wikisource as well.

Yes, that's all correct @ifried, sorry I forgot to move it before!

The multiple languages part of this will be dealt with in T280214 (because the lang list is common to both engines). We might want to still do some per-engine verification of the language codes though.

PR: https://github.com/wikimedia/wikimedia-ocr/pull/16

2.6.2 is released now and includes the new translations: https://ws-export.wmcloud.org/?uselang=or

Test site is up and running, with auto deployment. The documentation has started at https://wikitech.wikimedia.org/wiki/Nova_Resource:Wikisource/Wikimedia_OCR but there's more to be added.

Great, thanks. The new translations should be live within a week.

@Psubhashish: done.

Thanks for reporting these issues!

Good points! And the patch is nice and simple too (look at me, trying to over-complicate things as usual… :-P )

I wonder if IA Upload should create a Wikidata item as well, where we could store that stuff?

It looks like the wikimedia/simplei18n package doesn't support {{GENDER}}, but even if it did there doesn't seem to be any way for us to get that information about a user. Do other tools handle this correctly?

Fixed in https://github.com/wikisource/ia-upload/pull/53

This is all merged now deployed to the test site: https://ocr-test.toolforge.org

Draft PR: https://github.com/wikimedia/wikimedia-ocr/pull/14

It looks like this is the same Parsoid issue as T274747: see the difference between the wiki page and the Parsoid HTML. The issue is in the Auxiliary Table of Contents template.