Posts

@Background Pony #1057
Well, ok, but while I have no idea what it has to do with the dev changes today, I still think if you want to find images with Lauren Faust on them, you can search the tag ‘Lauren Faust’.

If you want to use that to prove some other point, that’s fine, I guess. Whatever works for you. I just don’t know why you’re framing me as being opposed to whatever it is you’re proposing, or why that’s a perception you want to create.

Like, why would anyone give a shit if I thought the palette of the different tag categories was good or bad? How is presenting that as something I care about even useful for you in trying to advocate for your proposed change?

I don’t have anything to say about it. It’s not my site. I’m just a volunteer here.

Literally here and here.

I thought I saw it brought up more times than that, maybe I’m just remembering too intensely.

Literally what?

@🐴
On this website? tight flank

@Background Pony #2771
Also, how necessary is it to stick to a palette of 256 web-safe colors? It’s 2024, not 1996. More importantly, this is a place for images. Those images will use whichever colours they damn well please. I understand that colors should be perceptually distinct, but which part of the rainbow is most squeezed?

Tags like “alternate version” and “derpibooru exclusive” could have their own category (let’s call it origin-adjacent), but still keep the same color as the origin tags. If a different color is necessary, they could get a slightly different shade.

@byte[]

FWIW, -source_url:?* was the way to find absence of sources in the past (and more generally, to find absence of content for other string-type fields)

@The Smiling Pony
Ciaran’s lamentations about running out of distinct web-safe colors notwithstanding, this really means further division of tag categories is warranted.

Also, keep in mind that some characters have multiple mutual-implication-enforced tags, or that people use multiple tags for the same oc (until someone notices and posts about it in the tagging forum), and that “oc” and “oc only” count within oc_tag_count. (oc is automatically added whenever someone tags an oc:-prefixed tag, but may be missing for various reasons.)

@JP
That and so many other tags. Don’t get me started. Nevermind also whether you’d tag “oc only” on a picture where the oc has a figurine of a canon character in the background on a shelf, for example, and that you wouldn’t tag “oc only” if there’s an oc and something else that wouldn’t be character-categorized (eg. Optimus Prime).

@byte[]
Owo, thanks!

>oc_tag_count.gt:0, character_tag_count:0, -oc only
>121510 results
Oh, wow. People really do ignore the oc only tag.

@byte[]
For those trying it out, note that tags like “alternate version” or “derpibooru exclusive” count as origin tags, so results aren’t going to be 100% clean.

New search fields have been added

source_count: integer -> find images with the specified number of sources
source_count:0 - find images with no sources

body_type_tag_count, error_tag_count, character_tag_count, content_fanmade_tag_count, content_official_tag_count, oc_tag_count, rating_tag_count, species_tag_count, spoiler_tag_count: integer -> find images with the specified number of tags with the given category

origin_tag_count.lte:2 - find images with two, one, or zero artists

@The Smiling Pony
Frik! How did you know I was searching for Megaman?!

Thanks a bunch man.

@ЅℰℒᎯℙℋℐℰℒ
Like this?

@The Smiling Pony
How do you search up specific thread titles using just keywords?
All we get is the topic ID number.

@ЅℰℒᎯℙℋℐℰℒ
You mean like this?

I remember there used to be a way to search up posts using specific keywords under users or pull specific threads on the forum using said keywords.

@ЅℰℒᎯℙℋℐℰℒ
it is tag-based. Your most basic search is tag, tag, tag, -tag, -tag. All the syntactical stuff is for more advanced searching to narrow things down.

@ЅℰℒᎯℙℋℐℰℒ
You mean the tag search tool?

@The Smiling Pony
I remember it used to be a tag based search function, rather than the syntax one

@ЅℰℒᎯℙℋℐℰℒ
… which old search function?

The site’s only had two, and I really doubt you mean the first one…

I find that the old search function was much more user friendly than what we have now.

@Background Pony #1057
ok

The recent commit
“switch to maintained twitter scraper implementation” (#187) seems incomplete and potentially worse than nothing.

The part that’s worse than nothing:
It does not seem to fetch the best resolution available. A user uploading to the site would see that the site has successfully retrieved an image and be left unaware that there’s a different step they need to take (visiting the tweet, viewing the image, and copying that URL).

I do not know if this is a bug in the service this commit uses (which is FxTwitter) or an insurmountable limitation in the information Twitter makes available. If it’s a bug, philomena should hold off using that service until it is fixed; if it is insurmountable, it should not use that service to retrieve images.

(Curiously: the FxTwitter API endpoint provides in its JSON response the width and height pixel values of the full size image, but the associated link to the image on pbs.twimg.com retrieves one smaller than the stated pixel dimensions. Here’s an example, stating a 2048 x 1816 resolution yet linking to a 1200 x 1064 pixel image. Viewing the actual tweet instead, one indeed finds the full 2048 x 1816 image.)

The rest that’s incomplete:
The commit directly accesses https://d.fxtwitter.com/#{user}/status/#{status_id}.jpg which retrieves only the first image of a multi-image tweet. This is also nearly worse than nothing, because a user might forget that there are more images to go back for when they see only one image show up from fetching.

Additional images can be fetched from https://d.fxtwitter.com/#{user}/status/#{status_id}/photo/N where N is 1 through 4, although the endpoint currently returns image #1 if passed a number that the post does not contain. Alternately, api.fxtwitter.com returns post info including all associated media (with the aforementioned resolution problems.)

 
The API endpoint would also allow autofilling the tweet’s text as the post description, although an issue in the repo indicates it does not always return the full text (though I have not seen this first-hand.)

@JP
So long as I retain the cinnamon theme and can keep the sharp corners, I’m happy. I wouldn’t mind a readable ripoff of Ponerpics’ April Fool’s deer theme, though.

Double click the search bar to bring up your autocomplete selections, scroll/look for whatever you don’t want in there, hold down left click and tap the delete key.

That should clear those out for you.

@sxcbeast
Like Danbooru?