firefox-translations VS PDF.js

Interestingly, I think this is actually related to the offline translation features built into Firefox. Both are products of "Project Bergamot", but the Mozilla-maintained version was later merged into the Firefox application:

https://browser.mt/

https://blog.mozilla.org/en/mozilla/local-translation-add-on...

https://hacks.mozilla.org/2022/06/training-efficient-neural-...

https://github.com/mozilla/firefox-translations

https://firefox-source-docs.mozilla.org/toolkit/components/t...

Extra webpage with screenshot and links, impossible to search for normally:

https://translatelocally.com/downloads/

Does one thing and does it well.

Oh— For downloading models, it's much easier to pipe/`xargs` `translateLocally --available-models` into `translateLocally -d` than go through the GUI.

---

Other self-hostable translation tools:

https://www.apertium.org/index.eng.html

- Traditional rule-based translation. Seems to work pretty well, but no good desktop frontend.

https://www.argosopentech.com/

- Works, but crashy desktop app.

https://libretranslate.com/

- API wrapping Argos Translate.

https://lingva.thedaviddelta.com/

- Google Translate scraper/privacy frontend.

https://euroglot.com/

- Proprietary, subscription trialware.

You can submit feature requests to this Github repository: Add Japanese Translation.

I reported this to https://github.com/mozilla/firefox-translations/issues/688

Mozilla did release DeepSpeech[0] and Firefox Translation[1] (the latter of which they included in Firefox, to offer client-side webpage translations.)

They definitely have fewer resources than OpenAI, and they do not produce SOTA research (their publications have plummeted to 1/year anyway[2]). So the only way for them to make progress is to seek government grants or make challenges like these.

This challenge is unlikely to be profitable for the winning team: the expected value of winnings are likely around $1K when taking into account the probability that another team gets a better rank, but ML research projects are often more expensive (recently, Alpaca spent upwards of $600 on computation alone; and of course pretraining large models is much more expensive). So the main gain will be publicity.

[0]: https://github.com/mozilla/deepspeech

[1]: https://github.com/mozilla/firefox-translations/

[2]: https://research.mozilla.org/

Firefox has an offline translation extension: https://github.com/mozilla/firefox-translations

That will be released in the next version: https://github.com/mozilla/firefox-translations/commit/af29a6ca83c4e98566d510c0f150667a43e506d2

https://github.com/mozilla/firefox-translations/issues/374

Maybe it will get implemented

Did you install Firefox Translations? If so, the feature to turn off translation bar will come very soon in the next version according to this bug.

See https://github.com/mozilla/firefox-translations for in development languages.

readPdf: used for reading the dropped file and displaying it on the screen, it uses PDF.js to load the file, get all fields and display it on the browser.

I first started building the app in the browser, using PDF.js and Download.js to take a PDF and edit it, and then download it to your computer.

pdf2json is a module that transforms PDF files from binary to JSON format, using pdf.js for its core functionality. It also incorporates support for interactive form elements, enhancing its utility in processing and interpreting PDF content.

Why not PDF.js?

I'd think opening a PDF in your browser would be at the same risk-level you associate with going to any random URL. On Firefox at least, I'm pretty sure the built-in PDF viewer is simply JS parsing and rendering the PDF anyway -- nothing with elevated permissions:

https://mozilla.github.io/pdf.js/

The PDF features are actually an extension, just one built in as Firefox's default pdf viewer.

It's called pdf.js https://github.com/mozilla/pdf.js/

You can actually use this pdf viewer in another browser like Chrome if you'd like, there's a demo URL on there.

We use Mozilla's PDF.js via the pdfjs-dist NPM module to load pages from a PDF file. The loadPdfPages function reads the PDF file and extracts its content. It returns an array where each object contains the page number and the text of that page.

The problem is that there simply wasn't a better option at the time.

Ogg Vorbis was a novelty at best, and it was the only decently widely adopted open source competitor for any of the items listed that was available at the time.

HTML5 was only just published when Chrome launched. So Flash was at that point the only option available to show a video in the browser (sure, downloading a RealPlayer file was always an option, but it was clunky, creators didn't like people being able to save stuff locally, and was also not open source). Chrome in fact arguably accelerated the process of getting web video open sourced: Google bought On2 in 2010 to get the rights to VP8 (the only decent H.264 competitor available at that point) so they could immediately open source it. The plan was in fact to remove H.264 from Chrome entirely once VP8/VP9 adoption ramped up[1], but that didn’t end up happening.

Flash was integrated into Chrome because people were going to use it anyway, and having Google distribute it at least let them both sandbox it and roll out automatic updates (a massive vector for malware at the time was ads pretending to be Flash updates, which worked because people were just that used to constant Flash security patches, most of which required a full reboot to apply; Chrome fixed both of those issues). Apple are the ones who ultimately dealt the death blow to Flash, and it was really just because Adobe could not optimize it for phone CPUs no matter what they tried (even the few Android releases of Flash that we got were practically unusable). That also further accelerated the adoption of open source HTML5 technologies.

PDF is an open source format, and has been since 2008. While I don't know if pressure from Google is what did it, that wouldn’t surprise me. Regardless, the Chrome PDF reader, PDFium, is open source[2] and Mozilla's equivalent project from 2011, PDF.js, is also open source.[3] Both of these projects replaced the distinctly closed source Adobe Reader plugin that was formerly mandatory for viewing PDFs in the browser.

Chrome is directly responsible for eliminating a lot of proprietary software from mainstream use and replacing it with high-quality open source tools. While they've caused problems in other areas of browser development that are worthy of criticism, Chrome's track record when it comes to open sourcing their tech has been very good.

[1]: https://blog.chromium.org/2011/01/html-video-codec-support-i...

[2]: https://github.com/chromium/pdfium

[3]: https://github.com/mozilla/pdf.js

Firefox Nightly is an unstable test version. You should report PDF issues to this GitHub repository.