jaculabilis/danbooru
1
0
Fork 0
Code Issues Pull Requests Projects Releases Wiki Activity

Merge pull request #4800 from nonamethanks/fix-skeb

Browse Source 
Skeb: add support for multi-image posts and fix tests
This commit is contained in:
evazion
2021-05-17 12:16:21 -05:00
committed by GitHub
parent efcfb9a440 1bc266a3db
commit 12eacbe76f
2 changed files with 5 additions and 3 deletions
Download Patch File Download Diff File Expand all files Collapse all files

6
app/logical/sources/strategies/skeb.rb
View File

@@ -3,6 +3,7 @@
# * https://skeb.imgix.net/requests/199886_0?bg=%23fff&auto=format&w=800&s=5a6a908ab964fcdfc4713fad179fe715
## Watermarked:
# * https://skeb.imgix.net/requests/73290_0?bg=%23fff&auto=format&txtfont=bold&txtshad=70&txtclr=BFFFFFFF&txtalign=middle%2Ccenter&txtsize=150&txt=SAMPLE&w=800&s=4843435cff85d623b1f657209d131526
# * https://skeb.imgix.net/uploads/origins/04d62c2f-e396-46f9-903a-3ca8bd69fc7c?bg=%23fff&auto=format&w=800&s=966c5d0389c3b94dc36ac970f812bef4 (new format)
## Full Size (found in commissioner_upload):
# * https://skeb.imgix.net/requests/53269_1?bg=%23fff&fm=png&dl=53269.png&w=1.0&h=1.0&s=44588ea9c41881049e392adb1df21cce
#
@@ -13,6 +14,7 @@
# * https://skeb.jp/@OrvMZ/works/1 (separated request and client's message after delivery. We can't get the latter)
# * https://skeb.jp/@asanagi/works/16 (age-restricted, watermarked)
# * https://skeb.jp/@asanagi/works/6 (private, returns 404)
# * https://skeb.jp/@nasuno42/works/30 (multi-image post)
#
# Profile URLS
# Since skeb forces login through twitter, usernames are the same as twitter
@@ -25,7 +27,7 @@ module Sources
PAGE_URL = %r{#{PROFILE_URL}/works/(?<illust_id>\d+)}i
IMAGE_URL = %r{https?://(?:www\.)?skeb\.imgix\.net/requests/[\d_]+\?.*}i
IMAGE_URL = %r{https?://(?:www\.)?skeb\.imgix\.net/(requests|uploads/origins)/.*}i
def domains
["skeb.jp"]
@@ -47,7 +49,7 @@ module Sources
# Heavy heuristic to extract the uncropped image among the nighmare that is the skeb minified json
candidates = page&.css("script")&.map { |script| script.text&.scan(/(https:\\u002F\\u002Fskeb\.imgix\.net.*?)(?:"|,|\s)/) }
candidates = candidates.to_a.flatten.compact.uniq.reject { |match| match.include? "crop=" }
[candidates.first.gsub("\\u002F", "/")]
candidates.map { |img| img.gsub("\\u002F", "/") }
else
[]
end

2
test/unit/sources/skeb_test.rb
View File

@@ -33,7 +33,7 @@ module Sources
end
should "get the image url" do
assert_equal("https://skeb.imgix.net/requests/229088_2?bg=%23fff&auto=format&w=800&s=9cac8b76c0838f2df4f19ebc41c1ae0a", @site.image_url)
assert_equal("https://skeb.imgix.net/requests/229088_2?bg=%23fff&auto=format&txtfont=bold&txtshad=70&txtclr=BFFFFFFF&txtalign=middle%2Ccenter&txtsize=150&txt=SAMPLE&w=800&s=32a275893cf5362d51e5744ff5d8f88b", @site.image_url)
end
should "get the canonical url" do