[Kavita] (User) Stats Page Improvement

[Ark1369]

Idea Description

The New Statistic Page Overhauls came with lots of cool features that make Stats a lot more useful than before, especially the User Stats pages on their profiles.
But this update has also included a feature cut of Mark as Read not counting towards any user statistic anymore.
I understand the motive of Statistics in Kavita to showcase what is being read in Kavita, but the removal of Mark as Read introduces issues like

• Latecomers to Kavita will never have their true Stats be populated because 100-1000s of series they might have read before coming to know of Kavita can't be associated with Stats, so you will get a skewed picture of Genres/Tags and other aspects.
• Content being Read in Kavita actively thro apps such as Mihon/Tachiyomi and forks is not counted towards any of the Statistics due to it again working on Mark as Read basis instead of page by page. Im aware of the fact PWA exists, but it has some limitations and a lot of users like to do their reading in one place. Thus another set of Stats data is lost for those who use it.

As such would like to propose a solution that respects the Kavita Reading Sessions, but also fulfills the other users needs - that is to have a small toggle on user statistic page (like a switch icon) - Which switches the user stat displays between Kavita Reading Session/Activity based (current - system) and Overall (which counts Mark as Read stats, thus fixing the already read stuff/stuff read via Mihon etc issues - like it was pre overhaul).

This way users will be able to get best of both worlds.

Incase a toggle like that would not be possible, I would like atleast Genre/Tags to be populated based on overall data instead of just Activity based one.

Thank You for considering this request :)

Idea Category

User Experience

Duration of Using Kavita

3 Years

Before submitting

• I've already searched for existing ideas before posting.

[robotzel-git]

I also support this idea. I'm basically only interested in the overall stats. Right now as a workaround I'm creating a reading session and scrolling to the end whenever I mark a volume as read, but it's a bit tedious...

[newyorker123]

Support this. The pdf reader in Kavita is sometimes buggy, so I prefer to read pdf in other reader, and Mark them read in kavita. It will be great if the stat can reflect that progress as well.

[majora2007]

I totally understand the gap for the end users and am committed to lessening it with a solution in v0.9.0, however I haven't fully realized what that solution is and how to easily bring it.

I'm toying with the idea to create a fake session when you mark something as read and use the average reading time to estimate the time to read it (capped at a full day).

What I need to work out is, do I want to build out new modification APIs around these sessions? Do I want to provide a UI to edit sessions?

I wont be offering a toggle on the stats page. It's way too much work to maintain 2 versions of that page with different queries.

[robotzel-git]

I think fake sesstions would be ideal, maybe with a specific attribute in case your apis need to offer an option to filter them. Editing sessions or anything more would be overkill...
For the reading time, I'm personally ok with it reflecting only my Kavita activity, so it could simply be 0, but average reading time is also fine. As for the capping, maybe it would be easier when marking an entire series as read to create sessions for each volume.

[Ark1369]

Not having separate Stats Pages is fine/understandable - I only really proposed it in case that Actual Reading sessions data needs to be separated out. Having all be together at one place is the ideal solution, so no issues whatsoever.

Fake Reading Sessions using average time would be perfectly valid and a nice solution, just unsure about the Hard Cap of 24 Hours
While not an issue always, there are quite a few long-running franchises in Manga/Manhwa that take significantly longer than that - One Piece, for instance

But even if we exclude the exceptional cases of long enough Manga, The biggest issue will come in Novel/Books Libraries as Web Novels and certain franchises can go significantly longer (sharing one of the more extreme cases as a reference, but most of Web Novels go in 100s of hours range)

Robotzel Solution of Creating Sessions per volume/chap can alleviate issues for Manga, but Books will suffer since they are more often than not a singular volume/entity (especially for WN, since having 1000s of individual chaps instead of one big merged EPUB will create a lot of other problems)

However, if other statistics like Tags/Genres/Word Count/Pages etc are properly added, Reading Time being capped is a reasonable limitation over the current state.

Editing Sessions could be useful in extreme cases above, but I think just not having any limits would be easier than that, both for Devs and Users (unless it's aimed to add start/finish read dates, no strong opinion on that kind of stuff)

[majora2007]

Cap was per reading session per entity. So no chapter can be over 24 hours.

If we are handling a mark for red on a series, we probably will need to spread it out over multiple days, if the underlying entities sum to meet such needs.

[Ark1369]

Ah I see, in that case Manga/Comics issues become a non factor since even biggest of Omnibuses wont come close to that mark, only Web Novels will be a problem (Since usually all of chaps in 1 merged EPUB), but I think there overall use is Niche enough (Only have met two others in Kavita discord having WN) to not overly plan around them + Word Count is more relevant statistic for books for most.

Series Level Mark as read, adding Sessions/Time equivalent to the sum total of individual entities would be perfect.

One thing I am curious however, is how Fake Sessions will handle data that exceeds 24 hours, such as when a user first installs or onboards Kavita; they will be marking a lot of series as read, which would easily exceed the 24-hour-a-day reading limit. Perhaps something like Mark as Read creates Activities Stacks starting a predetermined year? like 2020/2010 for instance, so they wouldn't fill up the present days and cause issues with being active reading and stay out of the way of proper stacks?
But you likely have more insight into this matter already than I do :)

Once again, thanks for considering this FR 👍

[majora2007]

Think of it like this, again off the top of my head thinking:

I have a series with 100 chapters, each chapter is estimated at 1 hour long.

I mark the whole series as read.

Kavita will create a reading session for the full day and attribute 24 chapters to it, then create a reading session for today - 1 day and attribute another 24 chapters to it.

Reading sessions in Kavita are always bound to one day. This will allow for you to have stats that reflect an estimation and will not be something crazy like you read 100 chapters in a single day.

[robotzel-git]

This sounds great. The only mild inconvenient would be for people that just installed Kavita and mark a lot of series as read, their stats will start a bit skewed with a lot of 24h sessions :) Maybe a toggle in the user profile (mark as read: use average duration/no duration) would help people that don't want the approximated stats (or want to track only the actual Kavita reading sessions).

[majora2007]

Yes, but the old stat system had the same problem, anytime you come in and mark as read, it skews the stats. I don't like having to duplicate mark as read/unread for all entities as a solution personally.

[majora2007]

This has been delivered as of v0.8.9.22. Please test it out and let us know your feedback.