Tag Archives: Community

Sandy Chu: My Internship at the Internet Archive

Posted on by

This summer, continuing a years-long tradition, Open Library and the Internet Archive took part in Google Summer of Code (GSoC), a Google initiative focused on bringing new contributors into open source software development. This year, I was lucky enough to mentor Sandy, a long-time Open Library volunteer, on an exciting project to increase the accessibility of our books with real-time translations. We have invited Sandy to speak about her experience here as we reach the culmination of the GSoC period. It was a pleasure getting to work on this exciting project with you Sandy! – Drini

My name is Sandy Chu and I am a 2025 Google Summer of Code (GSoC) candidate who had the opportunity to work with the amazing Internet Archive engineering team. Prior to participating in the GSoC program, I had contributed as a volunteer software engineer for the Open Library open source repo. As someone who grew up using local libraries as a place to supplement my education and read books that my school could not afford, I was drawn to the Open Library’s mission to empower book lovers and provide a free, valuable resource to all. You can view my initial proposal here.

Coming soon in September, the Open Library will be able to better serve its global audience to access books that were previously not available due to a lack of localization. With the help of open source projects such as the Mozilla Firefox Translation Models and Bergamot Translator library, a new BookReader plugin will have the ability to leverage a user’s browser and hardware resources to toggle translations from a book’s original language to a translation in their language. Additionally, the translated text will also work with the ReadAloud feature to read books in the translated language.

The “Real-Time In-Browser Book Translation w/ Read Aloud (TTS)” project closely aligns with the Open Library’s 2025 goal of providing more with less. Although the Internet Archive hosts and provides its patrons with hundreds of thousands of publicly available works, patrons are limited to a subset of works that were published in their native language. Due to the unique image based implementation of the BookReader application, default browser translator options are not viable for many readers, so this project presents an opportunity to make a big impact for international audiences.

Currently in internal beta, the translation plugin allows patrons to quickly initiate a local translator on-their device and translate the book’s text in just a few seconds per page. With nine distinct languages available for translation from English (and potentially over 40 as we update to Mozilla’s latest models), this project will make countless works more accessible for patrons.

The primary goals of this project were:

  • Translating a book’s original text content to the patron’s desired language with minimal delay or disruption
  • Creating a visually seamless experience to maintain the immersive experience of reading a book without having to go back and forth between a translator and the book
  • Redirecting the existing TTS plugin to use the translated text when the BookReader is in translation mode

Language

Total Readable Books on OL

% of All Readable / Borrowable Books (out of 4,526,060)

Native Speakers Globally

(in millions)

English

3,034,445

67.04%

390

French

332,052

7.33%

74

German

180,341

3.98%

76

Spanish

120,516

2.66%

484

Chinese

90,531

2.00%

1,158

Korean

5,384

0.11%

81

Arabic

2,415

.000533%

142

Retrieved from Wikipedia, which references Ethnologue as its source. Chinese and Arabic dialects are grouped together since they both have a unified written system.

Translations

At the center of the translation plugin are the Neural Machine Translation (NMTs) models provided by Mozilla Foundation’s Firefox Translation Models project. These files contain the lexical and vocabulary conversions from the original language to the target language; these compact models are essential to the real-time, browser-side aspect of this project. Since we are currently using an older subset of models, the translation feature is still considered in the “alpha” stage of maturity and accuracy.

When the translation plugin is enabled by the user, the language registry and model files are fetched from a server within the Internet Archive. After the models have successfully loaded into the user’s browser, we are able to use the Bergamot Translator project scripts to create a dedicated Web Worker, which is initialized to handle the translation tasks in a separate background thread. The Web Worker immediately retrieves the text content within the text selection layer for the currently visible page(s) for translation. Pages that have been rendered but not visible in the BookReader are given a lower priority and translated after the queue of visible content is completed. 

An unmodified page in the BookReader.

The translation plugin script feeds the text within the text selection layer into the model for processing and prepares the stage for the translated output by covering the original image and text selection layer with a beige background. [Pull Request #1410]

The translation plugin has initialized and is providing the original text to the language model.

Once the translation is completed by the dedicated Web Worker, the output is then used as the text content for its respective paragraph chunk and appears as if the work is actually written in the target translation language.

The translation has completed and is now on the page!

Images with captions are also carefully handled so that the translated text box occupies nearly the same space as the text selection layer itself. 

Each translated paragraph is stored in a cache with a unique key to prevent the browser from re-translating recently viewed content [Pull Request #1410/commit]. To prevent readers from having to wait for the translation when “flipping” to the previous/next page, the translation plugin targets the visible pages on the screen then works to complete the translations for the non-visible but loaded pages. If a user decides to flip far from their current page in the work, the translation plugin will detect the newly rendered page and translate / populate the translated text layer while adjusting to a new page on the fly.

The text selection layer has been adjusted to appear in red.
The translation layer occupies roughly the same height and width by copying the text selection layer’s properties.

Fine-tuning the visual presentation and behind-the-scenes functionality of the plugin were the main challenges for this portion of the project. Ensuring that the translations for each text chunk were done without depending on a previous chunk was an essential behavior we identified in the early stages of the project. Both asynchronous and synchronous behavior is implemented within the code to ensure that users do not have to wait for longer than needed for paragraphs to complete their translations. The translation plugin utilizes event listeners within the BookReader to detect when a newly rendered text layer is created, which then triggers a translation call to the text content from the upcoming page.

Styling the translated text layer also proved to be difficult. Although it is possible to reuse the style properties on the existing (and invisible) selection text layer, additional adjustments were needed to ensure that the visible translation text would not overlap or go beyond the bounds of the original paragraph. In the early phases of the translation plugin development, there were many instances of text chunks exceeding the boundaries set for the translation layer, which resulted in scrollbars appearing within paragraph elements or not aligning properly with the text on the page.

A screen capture from an earlier version of the translate plugin. A scrollbar can be seen in the 2nd paragraph element of the left page.

Another styling issue that caught us off guard was a pre-existing bug that was only visible in the Chrome browser. Since Drini and I were both using Firefox as our default browser, we later learned during a demo that there was an element scaling issue that was immediately visible when the translation plugin was activated. [Pull Request #1421/commit]

ReadAloud

The next major piece of this project was to connect the translation plugin to the ReadAloud feature and allow users to hear the translated text read aloud. 

The normal flow of the TTS (Text-to-Speech) plugin calls a server-side API to retrieve chunks of text and bounding rectangles based on the page and paragraph index. However, since we have the translated text available within the BookReader locally, the extra network calls to the server were dropped in favor of feeding the translated text lines into the TTS engine directly. Tweaking the pre-existing functionality of the TTS plugin to interact with the content generated by the translate plugin required a substantial amount of investigation to figure out where the adjustments needed to be made for the translation plugin to gracefully take over. 

When the TTS plugin is activated, it checks whether or not the translation plugin is enabled within the BookReader. If the translation plugin is active, the TTS plugin retrieves the translated text on the page to use as its text input for the voice engine. 

Voice overs are also automatically adjusted as soon as the TTS begins to streamline the reading process. By checking the source language from the work’s metadata, the default voice of the TTS reader is automatically adjusted to the target language that was set within the translate plugin. The voice menu is also re-rendered to allow users to more easily switch between the source, target, or other languages for the TTS reader. [Pull Request #1430/commit]

ReadAloud Menu
The ReadAloud voices menu as it is seen without the translation plugin activated.
ReadAloud Menu With Translation
Voices are categorized by the source language, target language, and other languages detected on a user’s system.

Visual parity between the original TTS and translated TTS was maintained as well by highlighting the entire translated paragraph section. Since network info containing the bounding rectangles for a text chunk were no longer available, I was able to use a paragraph element’s offset properties to highlight the text being actively read by the TTS reader [Pull Request #1431/commit].

The BookReader highlights chunks of text that are actively being dictated by the ReadAloud plugin.
ReadAloud highlight with translation active
With a few tweaks, the ReadAloud highlighting feature can also be used to highlight the translated text being dictated by the voice over.

Although this stage of the project did not require as much new code, we encountered a relatively complex issue that would cause the TTS reader to not progress if the translation plugin is activated in a part of a book that contained one or more blank pages. The translation adjusted implementation of the TTS plugin would wait for a new page to be loaded and rendered within the browser but remain stuck on a page due to a synchronization issue. After two weeks of extensive investigation and testing, we were able to resolve the issue by utilizing an existing method that returns all pages that have been loaded but not rendered in the DOM yet [Pull Request #1431/commit] and consolidating the asynchronous translation call with Promise.all().

Next Steps

For now, this feature is currently scheduled to be released for internal testing before being released for full public use. While the majority of goals were completed within the project timespan, there are many additional improvements and expansions that are planned in the future as the BookReader’s translate plugin becomes more mature. The next major steps for this project involve expanding the number of available translation pairs by integrating the latest models from Mozilla’s Translation Model project, receiving and implementing feedback from a round of internal testing, and continually improving the UI of the plugin. Unit tests and offline testing environments are also part of the project’s future goals to help improve the troubleshooting process for developers.

Conclusion

I would like to express my thanks once again to my GSoC mentor Drini for his guidance. The first few weeks of this project felt especially daunting, but the patience and advice that I was given throughout this program helped me realize that this big intimidating project was easier to manage as a number of small tasks were taken step-by-step. I am very glad that I had the chance to be challenged in new ways while being able to leverage my existing JavaScript skills. 

I am extremely grateful that I was able to participate in the GSoC program and to help contribute to a high-impact feature for both the Internet Archive and Open Library. Though my time as a GSoC contributor has officially ended, I intend to continue my work as a contributor with the Open Library team to expand on the functionality of this feature and help increase the availability of published works to a wider global community.

Improving Open Library’s Translation Pipeline

Posted on by

A forward by Drini Cami
Drini Cami here, Open Library staff developer. It’s my pleasure to introduce Rebecca Shoptaw, a 2024 Open Library Engineering Fellow, to the Open Library blog in her first blog post. Rebecca began volunteering with us a few months ago and has already made many great improvements to Open Library. I’ve had the honour of mentoring her during her fellowship, and I’ve been incredibly impressed by her work and her trajectory. Combining her technical competence, work ethic, always-ready positive attitude, and her organization and attention to detail, Rebecca has been an invaluable and rare contributor. I can rely on her to take a project, break it down, learn anything she needs to learn (and fast), and then run it to completion. All while staying positive and providing clear communication of what she’s working on and not dropping any details along the way.

In her short time here, she has also already taken a guidance role with other new contributors, improving our documentation and helping others get started. I don’t know how you found us, Rebecca, but I’m very glad you did!

And with that, I’ll pass it to Rebecca to speak about one of her first projects on Open Library: improving our translation/internationalization pipeline.

Improving Open Library’s Translation Pipeline

Picture this: you’re browsing around on a site, not a care in the world, and suddenly out of nowhere you are told you can “cliquez ici pour en savoir plus.” 

Maybe you know enough French to figure it out, maybe you throw it into Google Translate, maybe you can infer from the context, or maybe you just give up. In any of these cases, your experience of using the site just became that much less straightforward.

This is what the Open Library experience has been here and there for many non-English-speaking readers. All our translation is done by volunteers, so with over 300 site contributors and an average of 40 commits added to the codebase each week, there has typically been some delay between new text getting added to the site and that text being translated.

One major source of this delay was on the developer side of the translation process. To make translation of the site possible, the developers need to provide every translator with a list of all the phrases that will be visible to readers on-screen, such as the names of buttons (“Submit,” “Cancel,” “Log In”), the links in the site menu (“My Books,” “Collections,” “Advanced Search”), and the instructions for adding and editing books, covers, and authors. While updates to the visible text occur very frequently, the translation “template” which lists all the site’s visible phrases was previously only updated manually, a process that would  usually happen every 3-6 months. 

This meant that new text could sit on the site for weeks or months before our volunteer translators were able to access it for translation. There had to be a better way.

And there was! I’m happy to report that the Open Library codebase now automatically generates that template file every time a change is made, so translators no longer have to wait. But how does it work, and how did it all happen? Let’s get into some technical details.

How It Began

Back in February, one of the site’s translators requested an update to the template file so as to begin translating some of the new text. I’d done a little developer-side translation work on the site already, so I was assigned to the issue. 

I ran the script to generate the new file, and right away noticed two things:

  1. The process was very simple to run (a single command), and it ran very quickly.
  2. The update resulted in a 2,132-line change to the template file, which meant it had fallen very, very out of date.

I pointed this out to the issue’s lead, Drini, and he mentioned that there had been talk of finding a way to automate the process, but they hadn’t settled on the best way to do so.

I signed off and went to make some lunch, then ran back and suggested that the most efficient way to automate it would be to check whether each incoming change includes new/changed site text, and to run the script automatically if so. He liked the idea, so I wrote up a proposal for it, but nothing really came of it until:

The Hook

In March, Drini reached back out to me with an idea about a potentially simple way to do the automation. Whenever a developer submits a new change they would like to make to the code, we run a set of automatic tests, called “pre-commit hooks,” mostly to make sure that their submission does not contain any typos and would not cause any problems if integrated into the site. 

Since my automation idea had been to update the translation template each time a relevant change was made, Drini suggested that the most natural way to do that would be to add a quick template re-generation to the series of automated tests we already have.

The method seemed shockingly simple, so I went ahead and drafted an implementation of it. I tested it a few times on my own computer, found that it worked like a charm, and then submitted it, only to encounter:

The Infinite Loop of Failure

Here’s where things got interesting. The first version of the script simply generated a new template file whether or not the site’s text had actually been changed – this made the most sense since the process was so fast and if nothing actually had changed in the template, the developer wouldn’t notice a difference.

But strangely enough, even though my changes to the code didn’t include any new text, I was failing the check that I wrote! I delved into the code, did some more research into how these hooks work, and soon discovered the culprit. 

The process for a simple check and auto-fix usually works as follows:

  1. When the change comes in, the automated checks run; if the program notices that something is wrong (i.e. extra whitespace), it fixes any problems automatically if possible.
  2. If it doesn’t notice anything wrong and/or doesn’t make any changes, it will report a success and stop there. If it notices a problem, even if it already auto-fixed it, it will report a failure and run again to make sure its fix was successful.
  3. On the second run, if the automatic fix was successful, the program should not have to make any further changes, and will report a success. If the program does have to make further changes, or notices that there is still a problem, it will fail again and require human intervention to fix the problem.

This is the typical process for fixing small formatting errors that can easily be handled by an automation tool. But in this case, the script was running twice and reporting a failure both times.

By comparing the versions of the template, I discovered that the problem was very simple: the hook is designed, as described above, to report a failure and re-run if it has made any changes to the code. The template includes a timestamp that automatically lists when it was last updated down to the second. When running online, because more pre-commit checks are run than when running locally, pre-commit takes long enough that by the time it runs again, enough seconds have elapsed that it generates a new timestamp, causing it to notice a one-line difference between the current and previous templates (the timestamp itself), and so it fails again. I.e.:

  1. The changes come in, and the program auto-updates the translation template, including the timestamp.
  2. It notices that it has made a change (the timestamp and any new/changed phrases), so it reports a failure and runs again.
  3. The program auto-updates the translation template again, including the timestamp.
  4. It notices that it has made a change (the timestamp has changed), and reports a second failure.

And so on. An infinite loop of failure!

We could find no way to simply remove the timestamp from the template, so to get out of the infinite loop of failure, I ended up modifying the script so that it actually checks whether the incoming changes would affect the template before updating it. Basically, the script gathers up all the phrases in the current template and compares them to all the incoming phrases. If there is no difference, it does nothing and reports a success. If there is a difference, i.e. if the changes have added or changed the site’s text, it updates the template and reports a failure, so that now:

  1. The changes come in, and the program checks whether an auto-update of the template would have any effect on the phrases. 
  2. If there are no phrase changes, it decides not to update the template and reports a success. If there are phrase changes, it auto-updates the template, reports a failure and runs again.
  3. The program checks again whether an auto-update would have any effect, and this time it will not (since all the new phrases have been added), so it does not update the template or timestamp, and reports a success.

What it looks like locally:

A screen recording of the new translation script in action. A developer adds the word "Your" to the phrase "Delete Your Account" and submits the change. The automated tests run; the translation test fails, and updates the template. The developer submits the updated template change, and the automated tests run again and pass.

I also added a few other options to the script so that developers could run it manually if they chose, and could decide whether or not to see a list of all the files that the script found translatable phrases in.

The Rollout

To ensure we were getting as much of the site’s text translated as possible, I also proposed and oversaw a bulk formatting of a lot of the onscreen text which had previously not been findable by the template-updating function. The project was heroically taken on by Meredith (@merwhite11), who successfully updated the formatting for text across almost 100 separate files. I then did a full rewrite of the instructions for how to format text for translation, using the lessons we learned along the way.

When the translation automation project went live, I also wrote a new guide for developers so they would understand what to expect when the template-updating check ran, and answered various questions from newer developers re: how the process worked.

The next phase of the translation project involved using the same automated process we figured out to update the template to notify developers if their changes include text that isn’t correctly formatted for translation. Stef (@pidgezero-one) did a fantastic job making that a reality, and it has allowed us to properly internationalize upwards of 500 previously untranslatable phrases, which will make internationalization much easier to keep track of for future developers.

When I first updated the template file back in February of this year, it had not been updated since March of the previous year, about 11 months. The automation has now been live since May 1, and since then the template has already been auto-updated 35 times, or approximately every two to three days. 

While the Open Library translation process will never be perfect, I think we can be very hopeful that this automation project will make une grosse différence.

🎉 2023 Open Library Community Celebration 🎃

Posted on by

🇳🇿🇦🇺🇳🇱🇺🇸🇨🇦🇸🇬🇩🇪🇮🇹🇮🇷🇪🇸🇨🇭🇺🇦🇨🇴🇲🇾🇮🇳🇦🇹🇪🇬

Back in 2020, we started the tradition of hosting an annual Community Celebration to honor the efforts of volunteers across the globe who help make the Open Library project possible.

Tomorrow, Tuesday, October 31st at 9am Pacific, we warmly invite the public to join us in a small gathering to celebrate the hardworking humans who keep the website going, see demonstrations of their accomplishments, and get a glimpse into our direction for 2024 — Halloween Edition!

Join us tomorrow:

During this online celebration, you may look forward to:

  • Announcements of Our Latest Developments: Discover the impact of our recent initiatives and how they’re making a difference.
  • Opportunities to Participate: Learn how you can get involved and become an active member of our volunteer community.
  • A Sneak Peek Into Our Future: Get an exclusive glimpse of what lies ahead in 2024 and how we’re shaping the future together.

For all the latest updates leading up to the event, be sure to follow us on Twitter by visiting https://twitter.com/openlibrary. Looking for ways to get involved?

Mark your calendars, spread the word, and get ready for an event that’s all about our incredible community. We can’t wait to see you there!

Update

You can watch the recording of the 2023 Open Library Community Celebration here!

How one volunteer is sharing a better reading experience with all of us

Posted on by

For nearly 15 years Open Library has been giving patrons free access to information about books in its catalog, direct to their computers. But for millions of readers across the globe who rely on their phones for access, this hasn’t always presented the ideal mobile reading experience.

This year, a volunteer within the Open Library community named Mark developed an independent mobile app, an unofficial companion to the website called the Open Library Reader. This lite app, which is available for free from the Apple store and Play store, emphasizes the mobile reading experience and showcases the books within a patron’s Open Library reading log. It’s a great way to take your personal library with you on the go.

While Open Library Reader is an unofficial app which is not maintained or supported by the staff at Internet Archive, we’re ecstatic that talented volunteers within our community are stepping up to design new experiences they wish existed for themselves and others. We applaud Mark, not only for the time he invested and showing what’s possible with our APIs, but — true to the spirit of Open Library — for sharing his app for free with patrons, in such a way which seems to respect patron privacy.

We sat down with Mark for an interview to learn why he created the Open Library Reader and which of its features may be appreciated by book lovers who are on the go.

A picture of a patron’s personal library when logged in to the Open Library Reader app

Open Librarian: “Why did you find the need to build an Open Library Reader?”

Mark: I read a lot of books on my iPad, especially old, hard-to-find mystery novels. Open Library has a lot of great reads, but I was getting frustrated trying to manage my Reading Log and read books in the tablet browser. There was a lot of scrolling and clicking around, a tap in the wrong place could send me off somewhere else, and the book I was reading was always surrounded by browser and bookreader controls. I just wanted to sit down and read, and not have to be reminded of the fact that I was looking at a website through a browser.

Open Librarian: What were some of the approaches Open Library Reader used to solve these problems?

Mark: I thought about some of the good tablet-based reading experiences I’ve had, and imagined what it could look like if the interface were centered around the individual reader and the small set of tools they need to find, manage, and read books. So the Reading Log shelves and the reading interface are the core of the app, and everything else kind of happens at the edges. Everything you need is just one tap away. The reading interface is still the familiar Internet Archive BookReader, but I’ve overlaid some additional functionality. You can hide all the controls with the single tap, and the book expands to completely fill the screen. I also added a swipe gesture, so it’s easy to turn pages if you’re holding your device with one hand on the couch.

Open Librarian: What does it feel like to use? Can we have a tour?

Mark:

Open Librarian: What is your favorite part of the app? I like how it shows the return time

Mark: That is cool — that’s another example of centering the needs of the reader. It’s hard to pick a favorite part. Every feature is the result of me reading in the app every day for months before I released it. Periodically, I’d think “that’s kind of annoying” or “I wish I could…” and I’d go code for a while until I was happy with the experience. But the full-screen reading mode is probably my favorite. With the high-resolution page scans expanded to fill the screen, it’s almost like reading a physical book.

Open Librarian: What was your experience like developing the Reader?

Mark: I’m a retired web developer, so interface design, user experience, APIs and that sort of thing are nothing new, but I’ve never built a native app. After some reading, I picked Google’s Flutter tool, which allows easy cross-platform app development. I was amazed at how fast it was to assemble a simple app with just a few lines of code, and then it was just a matter of layering on the functionality I wanted. I spent a lot of time exploring the Open Library and Internet Archives APIs to figure out the best way to get at the data I needed, and even submitted a few updates to the Open Library codebase to support features I wanted to build. The Open Library team was extremely welcoming and supportive, and really made this app possible.

How can you support Mark’s work?

First, try downloading the Open Library Read App from the Apple store or Play store. If you have a suggestion, question, or feedback for Mark, send him an email to olreader@loomis-house.com. If you appreciate his work, consider rating the app on the app stores and leaving a review so others may discover and enjoy it too. To learn more about Mark and the Open Library Reader, look out for his upcoming interview on the Open Library Community Podcast.

Want to contribute to Open Library too?

See all the ways you can volunteer within the Open Library community!

An old brass key dropped on the ground in the woods.

Open Library Tags Explained—for Readers Seeking Buried Treasure

Posted on by

As part of an open-source project, the Open Library blog has a growing number of contributors: from librarians and developers to designers, researchers, and book lovers. Each contributor writes from their perspective, sharing contributions they’re making to the Open Library catalog. As such, the Open Library blog has a versatile tagging system to help patrons navigate such a diverse and wide range of content.

Continue reading →