Filter by chapter count/page count when searching for manga [complicated, see dev response]

[LeoTal]

I would like an option to the advanced search which allows you to filter by a minimum (or maximum) number of chapters.
If "only with translate chapters in language" is selected, then only counting the number of chapters in that language.
Alternatively the metric counted could be pages.
I want this feature primarily to sort away manga with few chapters, I usually want to read something with a substantial number of published chapters, so this feature would save me a lot of time, and I hope other readers here would appreciate it too.

 

[Raebit]

+1
Sorting by more to less chapter is great too

 

[tristan9]

Sorting by chapter counts may be possible, but the problem is that it isn’t very clear what that means in technical terms. So if a title has:

• vol 1 ch 1 by group A in english
• vol 1 ch 1 by group B in english
• vol 1 ch 1 by group C in french
• vol 1 ch 3 by group D in chinese

Then it’s not entirely obvious what the chapter count is, without starting to store a somewhat ludicrous amount of different counts for various meanings of chapter count:

• absolute chapter count,
• or deduped per volume x ch number permutation,
• or deduped per volume x ch number AND per language,
• maybe filtered by a given language
• or filtered by a set of languages,
• what about alt. versions when they exist (censored vs uncensored for example),
• and probably a few cases I forget

Basically it should be doable but there’s quite a few questions about what exactly of these we should apply, and how much data that represents and how to organize it for fast searches.

Now for page counts, that’s the same problem but worse so unlikely to ever happen tbh.

 

[LeoTal]

Sorting by chapter counts may be possible, but the problem is that it isn’t very clear what that means in technical terms. So if a title has:

• vol 1 ch 1 by group A in english
• vol 1 ch 1 by group B in english
• vol 1 ch 1 by group C in french
• vol 1 ch 3 by group D in chinese

Then it’s not entirely obvious what the chapter count is, without starting to store a somewhat ludicrous amount of different counts for various meanings of chapter count:

• absolute chapter count,
• or deduped per volume x ch number permutation,
• or deduped per volume x ch number but per language,
• maybe filtered by a given language
• or filtered by a set of languages,
• what about alt. versions when they exist (censored vs uncensored for example),
• and probably a few cases I forget

Basically it should be doable but there’s quite a few questions about what exactly of these we should apply, and how much data that represents and how to organize it for fast searches.

Now for page counts, that’s the same problem but worse so unlikely to ever happen tbh.

In the worst case, you can simply count by the largest number (group D) so long as there is some metric by which to compare length of content between different works, even if it is imprecise that is fine.

Obviously I do not mean by my suggestion that a work with the same number of chapters, but with twice as many groups working on translations should be counted as twice as long, in terms of content.

If it is possible to store the de-duplicated value per language, then as an extension, it would be additionally useful to be able to sort according the the number of chapters in the language specified in the advanced search option
"only with translate chapters in language", but that's not necessary, since just sorting away the very short works regardless of translation status would be helpful for me.

 

[tristan9]

In the worst case, you can simply count by the largest number (group D) so long as there is some metric by which to compare length of content between different works, even if it is imprecise that is fine.

The problem is some groups finding it funny to do things like scanlating the last chapter of a series only. Or, when it's not malicious, it can also just be the fact that scanlation is regularly not done in-order.

As in, you will regularly see a group start a very long series with the last few chapters, and then work on backfilling from ch1 as they follow up with the last ones. Think the following order: ch80, ch81, ch1, ch82, ch2, ch83, ch3, ...

So just taking the highest will lead to the number being a bit meaningless in too many cases, probably. Maybe it'd be niche enough to still be worth it though; hard to say for certain. Still feels like it'd be too fragile of an approach.

If it is possible to store the de-duplicated value per language, then as an extension, it would be additionally useful to be able to sort according the the number of chapters in the language specified in the advanced search option

That is indeed the likely "best" approach, though it is expensive in terms of the amount of data we'll need to compile per-manga as a result. But if/when we do it, yeah that is likely what we'd go with

"only with translate chapters in language", but that's not necessary, since just sorting away the very short works regardless of translation status would be helpful for me.

You'd be surprised how misleading it could be though; 10 chapters of 50 pages each vs 10 chapters of a 1-page-twitter-4-koma series is... Let's just say that from your description so far, it seems like the latter is exactly what you try to avoid 😅 So while we'll likely get there eventually, I can't guarantee that the result will truly be what you want.

 

[Raebit]

In worst case maybe like novel updates? By frequency update?

Or

"chapter translated : 1, 5 , 30 , 35"
There are 4 chapter translated?

Sad, I wish this feature available since I saw many old titles manga (thinking since it was old they have a lot of chapter) 1 chapter translated only... then no more updates 😭

 

[tristan9]

In worst case maybe like novel updates? By frequency update?

Can you expand on how precisely their system works? How do they arrive at a precise number to then sort with?

"chapter translated : 1, 5 , 30 , 35"
There are 4 chapter translated?

This ignores just about what I said earlier tho 😅 How exactly do you expect these to be counted or not based on diff groups maybe duping each other, and about languages, and how does that play with a given user's blocked groups or uploaders?

 

[OisE]

how does that play with a given user's blocked groups or uploaders?

His problem no ? And there is a button showing X numbers of chapters are hide no ?

to be counted or not based on diff groups

I don't understand why it's important to differentiate groups.
In my opinion just counting the chapters without decimal for each languages :

• EN : 150 ch
• FR : 7 ch
• RU : 42 ch
• PO : .....

 

[tristan9]

His problem no ? And there is a button showing X numbers of chapters are hide no ?

Yes and no. If:

• we still return it, but hide them, the chapter # is different from what the user sees, which is confusing for them
• we still return it, but show a "X chapters filtered", the chapter # is more understandable, but less useful for the user
• we don't return it, that fixes the chapter # for the user and makes it actionable, but incurs quite a bit of extra work on our end (means we can't precompute a title's per-language chapter count, basically)

I don't understand why it's important to differentiate groups.

Idk. It's a genuine question; means that you see chapter # as an evaluation of the amount of unique content available, while maybe another user thinks of it as "the number of chapters uploaded". Different people's different needs and wants, kind of.

In my opinion just counting the chapters without decimal for each languages :

What do you mean "without decimal" here? How do we know for sure that a group's "chapter 6" should be deduped with another group's "chapter 6.1, chapter 6.2, ..."? Maybe the other group uses decimals for omakes, which are real content that we likely should count there, no?

 

[OisE]

Yes and no. If:

If he really wants to read the manga he can just unblock the group.

Because, in my opinion, the primary use is to know how much is the advancement of the translation.

 

[LeoTal]

The problem is some groups finding it funny to do things like scanlating the last chapter of a series only.

Hilarious.

As in, you will regularly see a group start a very long series with the last few chapters, and then work on backfilling from ch1 as they follow up with the last ones. Think the following order: ch80, ch81, ch1, ch82, ch2, ch83, ch3, ...

Does this represent a significant percentage of works?

So just taking the highest will lead to the number being a bit meaningless in too many cases, probably. Maybe it'd be niche enough to still be worth it though; hard to say for certain. Still feels like it'd be too fragile of an approach.

I would prefer to have the option even if it fails in many edge cases (is fragile as you say), since my purpose is to remove a large number of works within a certain category, even halving or quartering the results for my present search would be very helpful, even if more than half of the remaining results are still "false positives" such as edge-cases you mention. I would appreciate being able to use this option in the advanced search filter.

That is indeed the likely "best" approach, though it is expensive in terms of the amount of data we'll need to compile per-manga as a result. But if/when we do it, yeah that is likely what we'd go with

I agree that the most robust and bulletproof solution is the best in the long run, however I would still derive most of the value from my suggestion with the "fragile" functionality above and would benefit from being able to use a WiP feature months earlier.
I think there are other readers who feel the same.

You'd be surprised how misleading it could be though; 10 chapters of 50 pages each vs 10 chapters of a 1-page-twitter-4-koma series is... Let's just say that from your description so far, it seems like the latter is exactly what you try to avoid 😅 So while we'll likely get there eventually, I can't guarantee that the result will truly be what you want.

I am glad that the advanced search currently allows you to filter out 4Koma, which I do

 

[tristan9]

If he really wants to read the manga he can just unblock the group.

lol I mean I would concur with you, but MD generally tries to do the right ™️ thing, even if we don't personally agree with the reasons why (a lot of, in this case) people see a feature as necessary 😅

Hilarious.

ikr

Does this represent a significant percentage of works?

quite a bit yeah; sometimes it also happens because the group that did the first few chapters decides to have them removed from MD for example, for whatever reason

I would prefer to have the option even if it fails in many edge cases

Fair point yeah

I am glad that the advanced search currently allows you to filter out 4Koma, which I do

fair enough 😄

 

[LeoTal]

To clarify the primary purpose of such a feature:

Filter away the absolutely huge volume of manga which have <10 chapters posted total, which aren't marked as on hiatus.

 

[Xwui]

So far, from what I can see, when you use the Search or Advanced Search option, the only details displayed are:

• Country of Origin/Original's Language
• Genre
• Rating
• Number of Times Bookmarked
• Number of Reads

However, the number of Chapters available is not displayed unless you specifically click a result to go to its page. While I understand that it may be difficult since there are multiple uploaders for chapters + multiple languages, but wouldn't it be a bit more convenient for the current number of chapters to be displayed? For example, searching for stories by inputting the genre will show all the above details + the latest/highest chapter uploaded. I just think this would make searching for stuff to read more convenient, since often times I would find something that seems interesting, only for it to only have 1 or 2 chapters available which isn't very satisfying when I'm really in the mood to read something.

mod note: merged into existing thread

 

[stealthynut]

How many of you are browsing for a new manga, read through a bunch of overviews, find one that sounds awesome, but then it only has 1-3 chapters? What if there was a way to filter out those mangas with only a few number of chapters so you could find that manga that you can really get into? Or just a sort by feature that allows you to sort by number of chapters, that'd work too if it was easier.

mod note - merged into existing thread

 

[TheLastSpartan42]

Sorting by chapter counts may be possible, but the problem is that it isn’t very clear what that means in technical terms. So if a title has:

• vol 1 ch 1 by group A in english
• vol 1 ch 1 by group B in english
• vol 1 ch 1 by group C in french
• vol 1 ch 3 by group D in chinese

Then it’s not entirely obvious what the chapter count is, without starting to store a somewhat ludicrous amount of different counts for various meanings of chapter count:

• absolute chapter count,
• or deduped per volume x ch number permutation,
• or deduped per volume x ch number AND per language,
• maybe filtered by a given language
• or filtered by a set of languages,
• what about alt. versions when they exist (censored vs uncensored for example),
• and probably a few cases I forget

Basically it should be doable but there’s quite a few questions about what exactly of these we should apply, and how much data that represents and how to organize it for fast searches.

Now for page counts, that’s the same problem but worse so unlikely to ever happen tbh.

The quickest most raw form of this feature is just the count of what is displayed on the manga page. However many links exist is the count. This feature is desperately needed. Use an agile approach get a simple feature in and iterate. No need to do it all at once. Just this would filter soooo much and make a huge time save.

Hopefully a count of links(or W/e is used to make them) is performant. Zero idea of the code base but this is as simple as I can think to take this feature.

This count could also be added to the manga cards on the various search’s to give users quick information too.

 

[Sw1tchblade]

lol I mean I would concur with you, but MD generally tries to do the right ™️ thing, even if we don't personally agree with the reasons why (a lot of, in this case) people see a feature as necessary 😅


ikr


quite a bit yeah; sometimes it also happens because the group that did the first few chapters decides to have them removed from MD for example, for whatever reason


Fair point yeah


fair enough 😄

I admit i haven't checked if you have mentioned somewhere what database you use, but if i put it in sql terms,
couldn't you just return the result from something like

select count(*) from chapters where language = 'EN'

for each language and just include it in the search results? Edge case or no, it's still better than nothing.
I assume you have the data tagged in some clever way.

 

[TheLastSpartan42]

I admit i haven't checked if you have mentioned somewhere what database you use, but if i put it in sql terms,
couldn't you just return the result from something like

select count(*) from chapters where language = 'EN'

for each language and just include it in the search results? Edge case or no, it's still better than nothing.
I assume you have the data tagged in some clever way.

Looking at the urls they use they have guids, A title guid has to have some connection to the chapter guids.

Tbh even removing the language filter would be fine. I think just being able to tell if some manga has more than 3 chapters or something fast will save me time in reading the title and synopsis.

 

[tristan9]

I admit i haven't checked if you have mentioned somewhere what database you use, but if i put it in sql terms,
couldn't you just return the result from something like

You're missing a lot of context as to why this isn't possible for us indeed... We're aware of basic SQL and we would have long done this if it was (a) advisable and (b) this easy... 😅

 

[DrBuns]

You're missing a lot of context as to why this isn't possible for us indeed... We're aware of basic SQL and we would have long done this if it was (a) advisable and (b) this easy... 😅

Do this, but on the search page. I don't know why you are making a big fuss about it not being possible.