Change how artist URLs are normalized in artist entries. Don't try to secretly
convert image URLs to profile URLs in artist entries. For example, if someone puts a
Pixiv image URL in an artist entry, don't secretly try to fetch the source and
convert it into a profile URL in the `normalized_url` field.
We did this because years ago, it was standard practice to put image URLs in artist
entries. Pixiv image URLs used to contain the artist's username, so we used to put
image URLs in artist entries for artist finding purposes. But Pixiv changed it so
that image URLs no longer contained the username, so we dealt with it by adding a
`normalized_url` column to artist_urls and secretly converting image URLs to profile
URLs in this field. But this is no longer necessary because now we don't normally put
image URLs in artist entries in the first place.
Now the `profile_url` method in `Source::URL` is used to normalize URLs in artist
entries. This lets us parse various profile URL formats and normalize them into a
single canonical form.
This also removes the `normalize_for_artist_finder` method from source strategies.
Instead the `profile_url` method is used for artist finding purposes. So the profile
URL returned by the source strategy needs to be the same as the URL in the artist
entry in order for artist finding to work.
Add a method for converting a source URL into a profile URL. This will
be used for normalizing profile URLs in artist entries.
Also add the ability to parse a few more profile URL formats.
This turns out to be a little simpler than keeping them separate. The
only thing special we have to do for Sta.sh is use the Sta.sh page when
we have a DeviantArt image with a Sta.sh referer.
Bug: We assumed the referer URL was from the same site as the target
URL. We tried to call methods on the referer only supported by the
target URL.
Fix: Ignore the referer URL when it's from a different site than the
target URL.
Remove the `preview_urls` method from strategies. The only place this was used was
when doing IQDB searches, to download the thumbnail image from the source instead of
the full image.
This wasn't worth it for a few reasons:
* Thumbnails on other sites are sometimes not the size we want, which could affect
IQDB results.
* Grabbing thumbnails is complex for some sites. You can't always just rewrite the
image URL. Sometimes it requires extra API calls, which can be slower than just
grabbing the full image.
* For videos and animations, thumbnails from other sites don't always match our
thumbnails. We do smart thumbnail generation to try to avoid blank thumbnails, which
means we don't always pick the first frame, which could affect IQDB results.
API changes:
* /iqdb_queries?search[file_url] now downloads the URL as is without any modification.
Before it tried to change thumbnail and sample size image URLs to the full version.
* /iqdb_queries?search[url] now returns an error if the URL is for a HTML page that
contains multiple images. Before it would grab only the first image and silently
ignore the rest.
Remove the `image_url` method from source strategies. This method would
return only the first image if a source had multiple images. The
`image_urls` method should be used instead. Tests were the main place
that still used `image_url` instead of `image_urls`.
Also make post replacements return an error if replacing with a source
that contains multiple images, instead of just blindly replacing the
post with the first image in the source.
Fix this Javascript warniing:
Alpine Warning: Unable to initialize. Trying to load Alpine before
`<body>` is available. Did you forget to add `defer` in Alpine's
`<script>` tag?
Fix `image_urls` returning `[nil]` when fetching data for a image URL
that was bad_id. In that case `original_urls` is empty, so we fall back
to using the deleted image URL as-is.
Fix this deprecation warning:
DEPRECATION WARNING: ActiveSupport::TimeWithZone.name has been deprecated
and from Rails 7.1 will use the default Ruby implementation. You can set
`config.active_support.remove_deprecated_time_with_zone_name = true` to
enable the new behavior now.
Triggered by the XML serializer in the API.
* Fix error when uploading non-ugoira files.
* Fix sample image URLs not being rewritten to full images correctly. We
have to get the full image URL from the API because given an
/img-master/ URL, we don't know what the original file extension is.
* Drop support for preview_urls. This means that IQDB lookups may be
slower, especially for ugoiras, since we have to download the full
ugoira now. However, ugoira lookups should produce better results,
since the ugoira thumbnail chosen by Pixiv wasn't necessarily the same
as the thumbnail chosen by Danbooru.
* Drop support for uploading single manga pages:
http://www.pixiv.net/member_illust.php?mode=manga_big&illust_id=18557054&page=2
Previously uploading an URL like this would only upload a single image
out of a multi-image work. Now it will upload all images in the work.
Pixiv no longer supports URLs like this, so we don't either.
* Add support for parsing URLs like this:
https://i.pximg.net/c/360x360_70/custom-thumb/img/2022/03/08/00/00/56/96755248_p0_custom1200.jpg
Apparently artists can choose a custom thumbnail now (not like anyone
will try to upload one though).
Add upload support for Pixiv Sketch. Fetch tags, commentary, and artist,
and rewrite sample images to full images.
Authentication isn't required. R18 images are hidden in the browser but
visible in the API.
Forcibly redirect users to the name change page if their name is
invalid. This means user with invalid names can't do anything or view
any pages until they change their name. API requests are still allowed.
Add `foreman` to the base Docker image. This way you can do this:
docker run --rm -it -v $PWD:/danbooru ghcr.io/danbooru/danbooru foreman start
to start everything needed to run Danbooru in development mode (except
for the Postgres database). This will start everything listed in the
Procfile:
bin/rails server
bin/good_job start
bin/rails danbooru:cron
bin/webpack-dev-server
Use the GoodJob job adapter instead of the default Rails async job
adapter in development mode.
The default async adapter runs jobs in a background thread in the
`bin/rails server` process, but this sometimes has problems with jobs
blocking the main server thread. The job queue interface at `/jobs` also
didn't work with this.
This means that now you have to run `bin/good_job start` in development
mode in order to work background jobs. This is required for uploads to
work.
Add stricter username rules:
* Only allow usernames to contain basic letters, numbers, CJK characters, underscores, dashes and periods.
* Don't allow names to start or end with punctuation.
* Don't allow names to have multiple underscores in a row.
* Don't allow active users to have names that look like deleted users (e.g. "user_1234").
* Don't allow emoji or any other Unicode characters except for Chinese, Japanese, and Korean
characters. CJK characters are currently grandfathered in but will be disallowed in the future.
Users with an invalid name will be shown a permanent sitewide banner until they change their name.
The median username length is 8 characters. The 99% percentile is 18
characters. The 99.9% percentile is 24 characters. About 750 users have
a name more than 24 characters long.
This doesn't do anything about existing users with long usernames.
Note that this is the length in Unicode codepoints, not grapheme
clusters. Some Unicode characters and emoji may be a single glyph but
composed of multiple codepoints.
Fix an exception on the error page when a controller index action raised
an PG::AmbiguousColumn error because the model `search` method generated
SQL with an ambiguous column reference. In this case the error page
tried to generate data attributes for the <body> tag, but this failed
because evaluating the `current_item` raised an exception again.
Fix `Cannot write log file 'ffmpeg2pass-0.log' for pass-1 encoding: Permission denied` error
when uploading ugoira files. Caused by the fact that 2-pass encoding tries to write a log file in
the current directory by default, which fails in production because the default working directory in
the Docker image is /danbooru, which is read-only.
Add a limit so that users can't upload more if they already have more
than 250 images queued for upload.
For example, if you upload a Pixiv post that has 200 images, then you'll
have 200 queued images for upload. This will go down as the images are
processed. If you exceed the limit, then trying to create new uploads
will return an error.
This is to prevent single users from overwhelming the site by uploading
too many images at once, thereby preventing other users from uploading
because the job queue is backed up and can't process new uploads by
other users until existing uploads are finished.
Fix it so that upvoting or downvoting a revealed thresholded comment
doesn't hide it again.
The fix is to explicitly store a `data-show-thresholded` flag on the
comment, instead of manually hiding elements with jQuery, and to morph
the comment HTML instead of replacing it so that the state isn't lost
after voting. Alpine.js is used for this, which isn't strictly necessary,
but is useful to test the library before adopting it on a wider scale.
https://alpinejs.dev/start-here
Also fixes the uploader uploading all images when trying to upload only a
single image in a multi-image work. Caused by `image_urls` incorrectly
returning all images when the source strategy was given a url for a
single image.
Add `#basename`, `#filename`, and `#file_ext` utility methods to
Danbooru::URL and change a few places to use them. Simplifies parsing
filenames in source URLs in various places.
Introduce a Source::URL class for parsing URLs from source sites. Refactor the Twitter
source strategy to use it.
This is the first step towards factoring all the URL parsing logic out of source
strategies and moving it to subclasses of Source::URL. Each site will have a subclass
of Source::URL dedicated to parsing URLs from that site. Source strategies will use
these classes to extract information from URLs.
This is to simplify source strategies. Most sites have many different URL formats we have
to parse or rewrite, and handling all these different cases tends to make source
strategies very complex. Isolating the URL parsing logic from the site scraping logic
should make source strategies easier to maintain.
Fix certain ugoiras having very low quality webm samples. This was
because we had a target bitrate of 5 Mbps, but this wasn't enough for
videos that were high resolution or that had choppy, hard-to-compress
motion, such as post 5081776 (nsfw).
NicoSeiga changed it so that on every login, you must enter a 2FA code
sent by email. This broke the NicoSeiga strategy. The fix is to just use
a static session cookie instead (and hope it doesn't expire, and isn't
tied to an IP).
The `nico_seiga_login` and `nico_seiga_password` config settings have
been removed from config/danbooru_default_config.rb and replaced by
`nico_seiga_user_session`. If you run your own Danbooru instance, you
will have to update your config file manually.
* Remove unnecessary trailing slashes when artist URLs are saved.
* Automatically add `http://` to new artist URLs if it's missing (before
this was an error; now it's automatically fixed).
Introduce a Danbooru::URL class for dealing with URLs. This is a wrapper
around Addressable::URI that adds some additional helper methods. Most
significantly, the `parse` method only allows valid http/https URLs, and
it returns nil instead of raising an exception when the URL is invalid.
Fixes a bug where the Foundation source strategy failed because http.rb
automatically sent a `Content-Length: 0` header with all GET requests,
which caused Foundation to return a 400 Bad Request error. This behavior
was fixed in http.rb 5.x.
http.rb 5.x has a breaking change where it now includes the request object
inside the response object, which we have to handle in a few places.
Allow uploading multiple files from your computer at once.
The maximum limit is 100 files at once. There is still a 50MB size limit
that applies to the whole upload. This limit is at the Nginx level.
The upload widget no longer shows a thumbnail preview of the uploaded
file. This is because there isn't room for it in a multi-file upload,
and because the next page will show a preview anyway after the files are
uploaded.
Direct file uploads are processed synchronously, so they may be slow.
API change: the `POST /uploads` endpoint now expects the param to be
`upload[files][]`, not `upload[file]`.
Use a spinner icon instead of the word "Loading" for thumbnails that are
being processed in the background in a batch upload.
Also use morphdom to update thumbnails so we only update the parts of
the DOM that actually changed.
Add data attributes to thumbnails on the /uploads, /upload_media_assets,
and /media_assets pages. Add a `data-is-posted` attribute for styling
thumbnails based on whether they've already been posted.
Followup to 093a808a3. Using a NOT EXISTS clause is much faster than the
`LEFT OUTER JOIN posts WHERE posts.id IS NULL` clause generated by
`.where.missing(:post)`.
Change the loading indicator from a progress bar to a spinner. Fixes
issue with the <progress> element having a different appearance on
different browsers.
evazion
evazion
