Updates to Lessons, Reviews, and Extra Study

[Kumirei]

I just noticed that reordering this way actually clears the item stats, so any answers already submitted (correct or incorrect) for uncompleted items are wiped. Any way we could avoid that? It really hampers the usability of reordering scripts.

For people following along at home, unless we can keep the item stats on reorder “back to back” is not viable

[Sinyaven]

Version 0.5 fixes the dumb bug I introduced in version 0.4. structuredClone() does not clone getters.

EDIT: Shouldn’t these two data attributes be able to replicate the most important parts of back to back?

@tofugu-scott

• The data attribute data-quiz-queue-complete-subjects-in-order-value is a boolean set to "true" when you want to complete each subject in the order of the given queue (both reading and meaning are answered before moving on). This is used in extra study.
• The data attribute data-quiz-queue-question-order-value can be set to either readingFirst or meaningFirst and will override the random question selection with the given preference.

I’m not sure if you get notified when I edit the reply, so I’m also tagging you @Kumirei

[Tantivy]

This is actually pretty cool and I like most of the changes. The only thing I’m not a huge fan of is the huge space between the kanji and the word below it. I think it looks pretty closer together. Other than that this is pretty cool though

[Kumirei]

It can handle the “repeat until correct” type of back to back, but it’s better when you can shuffle an item when you get it wrong. I also have the “true back to back” option, which forces you to get both questions right back to back

[Sinyaven]

@Kumirei

it’s better when you can shuffle an item when you get it wrong

You are right – and if I understand correctly, the problem is that if you reorder the queue after a wrong answer, the record about the wrong answer is wiped so that the final result will always be submitted as passed. The only idea I currently have is to listen to the didAnswerQuestion event and manipulate the answer stats there:

window.addEventListener(`didAnswerQuestion`, e => e.detail.subjectWithStats.stats.meaning.incorrect = 1);

This should at least allow you change the submitted result to “not passed”. This could probably solve the “true back to back” problem, but I don’t think it will help with the shuffle option.

---

Anyway, I have added wkQueue.completeSubjectsInOrder and wkQueue.questionOrder to modify the corresponding data attributes.

[Gorbit99]

While doing lessons, there’s a significant pause when switching from one subject to the next. What could be causing this?

Also, I assume this is part of the session removal, but the number of items completed in that session is gone. I personally used this to pace my reviews, doing x an hour. Could this maybe get the same treatment as the “last 10 reviewed items”, so that the number items done in the last hour is shown?

[rfindley]

@tofugu-scott

The rationale for this approach has to do with the order that the stimulus lifecycle callbacks fire.

Scott, do you see any issues with something like this:

let quiz_queue = Stimulus.getControllerForElementAndIdentifier(
                 document.querySelector('#quiz-queue'),'quiz-queue');

do_some_sort_of_reordering(); // ...by manipulating the tags containing the subjects/subjectIDs
quiz_queue.completeSubjectsInOrderValue = true;
quiz_queue.questionOrderValue = 'readingFirst';
quiz_queue.initialize();
quiz_queue.nextItem();

I’ve tested this, it works, and I haven’t noticed any side-effects yet, but I’m new to Stimulus.

As far as I can tell, all the resources from prior calls to initialize(), such as QuizAPI and QuizQueue, should be junk-collected.

Of course, if a userscript does this after the user has already answered some items, it should also keep track of the answered items and make sure they are removed from the initial <script> tags, and I think the other scripters are already aware of that.

[seanblue]

@tofugu-scott

Just want to make sure you are aware that the queue order for lessons is really easy to manage as it is powered completely by the queue url parameter. e.g. https://preview.wanikani.com/subjects/1/lesson?queue=1-2-3-4-5-6 will show lessons for subjects with ids 1, 2, 3, 4, 5, 6 in that order (if you want a different lesson order just change the queue order).

Is there a way to change the lesson order after the lesson page has loaded (not the quiz) or must it be done on page load?

Also, does that URL take all the lessons possible or just the ones for a single batch? That is, if my batch size is 5, will the URL always have 5 subject ids or will it have the ids for all the lessons available?

@tofugu-scott

The queue I auto generate is based on the users lesson order preferences and their lesson batch size preferences (the default size is 5). There is nothing to stop the queue from being manipulated (for now). Please just use it responsibly.

Is there a way to manipulate the batch size?

In [Userscript] WaniKani Lesson Filter I allow the user to filter lessons by type (radical, kanji, vocab), change the batch size, and shuffle the lessons. All of this is done on the lesson page (not the quiz). I’m trying to figure out what is still possible on the lesson page itself versus if I’ll have to change the script to execute prior to the lesson page by opening the lesson URL with the necessary query params up front.

[tofugu-scott]

Thanks for the heads up. I have managed to get the scroll to disappear in Firefox but not Edge. I wont spend anymore time on this particular issue as the plan is to have a better UI for this on narrow screens in the future.

[tofugu-scott]

Thanks for the helpful info, both in terms of the summary page and the home button. I think we can do a better job of highlighting the wrap up which would negate the need for a confirmation dialog when you click the home button. To a veteran user, having to confirm when you click home would be annoying.

That said, I would consider this for future updates and not address these points this time round. Your points have been noted though.

[tofugu-scott]

Thanks for your feedback on this, and while I agree that the meaning and reading could be better, the focus at this point was to bring the styling inline between subject pages and the item info. The thought is that any improvements made in one place will benefit the other.

Your feature requests have been noted and may be addressed in the next batch of changes I will be doing after this change is on production.

Thanks for taking the time to provide your feedback here, and if you have further suggestions and feature requests please do let customer support know by emailing them at hello@wanikani.com

[tofugu-scott]

@Kumirei

Second I am wondering if there would be any way for us to alter the item answer stats? I erase answer stats for items for a particular setting of my “back to back” feature so that you have to redo previously correctly answered questions. Totally understand if there is not, I’ll just sunset that feature

There is no way to manipulate the answer stats sorry.

@Sinyaven

You are right – and if I understand correctly, the problem is that if you reorder the queue after a wrong answer, the record about the wrong answer is wiped so that the final result will always be submitted as passed. The only idea I currently have is to listen to the didAnswerQuestion event and manipulate the answer stats there:

window.addEventListener(`didAnswerQuestion`, e => e.detail.subjectWithStats.st

The suggestion by @Sinyaven above will no longer work either. I have addressed this loophole. My reason is that you should not rely on the code I consider internal. I want to be free to manipulate and change this however I see fit without the need to notify the user script community of every change. I feel that If I change the events then I can let you know, but if the internal structure of my code changes, it shouldn’t break anything you add. This helps reduce the burden on customer support as users of the scripts often don’t know if the script is causing a problem, or if it is Wanikani (not all users, but some…)

@Kumirei

Third, is there any way for us to change the preferred voice actor (in a session)? Being able to randomize the voice actor is really valuable

@Hubbit200

Just want to say I agree with what @Kumirei says above about the voice actor thing - I find having it randomized for each audio is essential (to avoid be getting too used to hearing words pronounced always exactly the same) and if it can’t be an option built into WaniKani then it would be great if you could at least make sure scripts can still do it!

The above is great feedback, and should be submitted as a feature request. I do however intend to reduce the payload of the subjects sent to the client by removing the non-preferred voice actor. If you feel that having randomness of voice actor is a great feature could you please raise a feature request by sending customer support an email?! We can then evaluate and prioritise with the other features that we want to deliver.

@Kumirei

It can handle the “repeat until correct” type of back to back, but it’s better when you can shuffle an item when you get it wrong. I also have the “true back to back” option, which forces you to get both questions right back to back

Maybe this should be a feature request for extra study too.

[tofugu-scott]

@Gorbit99

While doing lessons, there’s a significant pause when switching from one subject to the next. What could be causing this?

Can you please clarify what is a significant pause? It is loading a new page from the server when you move between subjects. There are optimisations that I could do to preload the next subjects, but preloading should be done mindfully as the extra bandwidth may not be necessary.

@Gorbit99

Also, I assume this is part of the session removal, but the number of items completed in that session is gone. I personally used this to pace my reviews, doing x an hour. Could this maybe get the same treatment as the “last 10 reviewed items”, so that the number items done in the last hour is shown?

Could you please request this as a feature by sending an email to customer support so it can be evaluated and fed into the backlog as appropriate?

[Gorbit99]

@tofugu-scott

Can you please clarify what is a significant pause? It is loading a new page from the server when you move between subjects. There are optimisations that I could do to preload the next subjects, but preloading should be done mindfully as the extra bandwidth may not be necessary.

Looking at the network tab, it varies pretty wildly, but really anywhere from 200ms at best to a whole second at worst.

@tofugu-scott

Could you please request this as a feature by sending an email to customer support so it can be evaluated and fed into the backlog as appropriate?

Sure, I’ll link back to the post I’ve made originally, I assume that works fine.

[tofugu-scott]

I would heavily advise against this approach as you are going to be to reliant on code that is likely to change. If, for instance I decide that the initialise code should be done on the connect callback, then your code will break. You are of course free to do what ever you like, but I think that approach is much more brittle than the alternative.

However, I will admit that I don’t know what any of the user scripts do in detail, so if you need to do this, I will just say “here be dragons”.

[tofugu-scott]

@Gorbit99

Looking at the network tab, it varies pretty wildly, but really anywhere from 200ms at best to a whole second at worst.

Once this goes out to production I will monitor the performance and timings. We have tools on production to help us do that.

The production hardware is different to that used for preview also, so I will defer to the performance optimisations until then (if they are still necessary).

[ikarishinji]

So far everything looks great. I’ve found one thing tho.
In the vocabulary examples for a kanji, before I could see the reading by hovering the word, now, it doesn’t show. That helps with understanding how a kanji will sounds in a word, and even if there’s other readings different from the one showing.

See the screenshot below.

Thanks

[ikarishinji]

Now, I found another one.
When teaching the kanji there were two readings, I picked あた for this one. Turns out I shouldn’t.

See the kanji page below.

[seanblue]

あた is a kun’yomi reading, so it seems like the shake is correct but the lesson info is wrong.

[tofugu-scott]

@ikarishinji

In the vocabulary examples for a kanji, before I could see the reading by hovering the word, now, it doesn’t show. That helps with understanding how a kanji will sounds in a word, and even if there’s other readings different from the one showing.

This now works as it did before.

@ikarishinji

When teaching the kanji there were two readings, I picked あた for this one. Turns out I shouldn’t.

like @seanblue helpfully identified, this was an issue with the lesson slide and not the lesson quiz. I have fixed this issue now.

Thank you I really appreciate you taking the time to report these issues.