Just wanted to check if other people were getting 500 errors from every url when trying to implement the api? I'm very new to coding so its possible I'm doing something wrong, but I wanted to check.

Switch to a document database & keep mangasearch in memory.

To improve the performance of the search you could just load everything relevant into memory and access it in every search.

This is a calculation how much memory it will take to load every manga for the search, 60 chars on average for titles 20 tags/genres, 8 uploader ä artist etc and manga and oneshot. Iguessed that this site has arround 200k mangas. The amount of bytes per manga is probably way more than it actually is.

16 byte updated,created timestamp
4 bytes * 20 tags, genres
4 bytes * 8 uploader, artist, …
4 bytes* 2 types
1 licensed bool
4 bytes * 60 chars

(460+1+42+4* 28)*200000= 75400000byte => 75mb

It would probably make sense to load users and tags into a hashmap. If there would be 50000 tags/users with a namelengrh of 10 chars it would take up 2 mb. In comparison some chrome tabs take up more memory than that. The search would probably only take a view ms. This calculation is made based on the actual size of chars/ints in a low level language. If implemented in python it would probably be more

Since the mangas get looked up by the id there won't be much load on the db compared to a normal db.

Anyone using an app called Obsidian? Im hoping that i can use mangaupdates api however i don't know what to do because im not a code expert. I only know that i need the script of this web api so i can use it in quickadd plugin but... all i can found was jikan api script but its not as detailed when it comes to manga, manhwa compared to your incredible website!! Sigh Im so upset...

Quote from vishnubly

Hello, I believe I have found a bug, in "categories/search", we cannot see any more than 2500 results, i.e for per page 100 results, after page 25, we cannot see any results, the same behavior is observed from UI as well: https://www.mangaupdates.com/categories.html?page=26&perpage=100&am ...

Wanted to echo this! Is there a way around this to see more than 2.5k results for a genre?

The current implementation of the Categories page uses an elastic aggregation which is limited by size. It's not something we're willing to increase at this point due to performance concerns.

Thank you for the response!

Another question! I am making a request on 'https://api.mangaupdates.com/v1/series/search' using these parameters: {'genre':["Yuri"],'page':1,'perpage':5} but the results are respecting the genre parameter. Is this a bug or did I do this incorrectly?
Thanks!

Can you change the single quotes to double quotes like this:

{
"genre": [
"Yuri"
],
"page": 1,
"perpage": 5
}

It worked for me.

Why is there a limit when there is no search criteria? Only order_by date_added or nothing at all shouldn’t take up much performance so why is it capped to 10k?

It's based on elasticsearch which has a built in 10k limit

Is there a search_after or scroll api simple search for a single genre/no search at all? https://www.elastic.co/guide/en/elasticsearch/reference/current/paginate-search-results.html#search-after

I don't recall this parameter existing when we did the original elasticsearch 5 implementation. Regardless, we're more focused on the UI right now, so I can't guarantee any major refactors like that in the short term.

Hello, I have a question about API. I have noticed that API is tend to raise 503 errors and I can do noting with that. Can you please say me what can be a reason of that?

You are being rate limited by the API

A just ruler amongst tyrants

Got it, what number of requests is rate limit for API? And one more question isn't that better to raise 422 status in that situation?