SOHEI
09/30/19 05:16PM
Small feature suggestion
I came up with an idea that might help with image curation and visibility. I'm just shooting the breeze here, so take it with a grain of salt.

So, I sometimes just hit the "random" button for a while to see what I get, and I notice that some images have very few tags. As a result, these images are ignored by the search engine. I'd add some tags myself, but they're too hard to find reliably.

I know that that the "tagme" tag exists, but I think something that would be more helpful is to add a list that sorts images by LEAST amount of tags, so a person can easily just load up a queue of images that need tagging, and tag characters, series, artists etc. that they recognize.

I don't know how your site's code works, so maybe this would be too difficult. Also, maybe this is just me and most of the users here wouldn't find this an interesting feature or use it. If this feature just isn't plausible or wanted, that's fine, it's just something I thought might help with the curation on the site.

So that's my two cents, thank you for reading, and have a nice day!

(P.S. Thank you for adding the tag prediction on the search engine, I really appreciate it)
GazapoDeMuerte
10/11/19 01:58AM
little known little little
Honestly, the admin would have to be the ones to explain the search coding, but our framework is based on the old framework from Danbooru.donmai . The search may find and aggrogate images into groups and produce a result that way, but that's if the word exists within that image/document. If the word does not exist in the image/document, that image/document is simply skipped. ...
So..
trying to find a search script for "(input1=null)+(input2=null)+(...etc..)" could go on forever.
.... my brain is about to explode from thinking about the possibilities too hard. But whether the input reads 'null' or 'yes' , it's all the same because the search is pinging every single image for the value that the user inputs into the search parameter (IE: tagvalue:=x<10, so that tagvalue=n+9(1[e-1])...value9 eventually drops to zero.
The progression of that execute would then look something like:
SRC="tagvalue"true value/pingvalue=10
SRC id:Id: 1823537
(ping1)value=true
(ping2)value=true
(ping3)value=true
(ping4)value=true
(ping5)value=true
(ping6)value=true
(ping7)value=false
(ping8)value=false
(ping9)value=false
(ping10)value=false
Postscript value"Id: 1823537"=(ping0)(ping1)(ping2)(ping3)(ping4)(ping5)(ping6)
Id:1823537 value6

Normal pings probably look more like
SRC="tagvalue"true value/"minato"
SRC id:Id: 1823537
(ping1)=false
(ping2)=false
(ping3)=false
(ping4)=false
(ping5)=false
(ping6)=true
Postscript value"Id: 1823537"=(ping0)(ping1)(ping2)(ping3)(ping4)(ping5)(ping6)
Id: 1823537 value6

Note that the first script used 4 more resources than the second script in this example.
1


Reply | Forum Index