Commit Graph

13 Commits

Author SHA1 Message Date
evazion
57a92ad336 Fix #5072: Fandom source normalization is wrong 2022-04-22 01:27:17 -05:00
nonamethanks
3b055138ff Fix normalization for fandom sources 2022-04-22 03:27:05 +02:00
evazion
d9d3c1dfe4 sources: rename Sources::Strategies to Source::Extractor.
Rename Sources::Strategies to Source::Extractor. A Source::Extractor
represents a thing that extracts information from a given URL.
2022-03-24 03:49:44 -05:00
evazion
34aa22f90b sources: fix fandom.com page urls.
Fix it so that sources like this:

* https://vignette.wikia.nocookie.net/valkyriecrusade/images/c/c5/Crimson_Hatsune_H.png/revision/latest?cb=20180702031954

link to this:

* https://valkyriecrusade.fandom.com/?file=Crimson_Hatsune_H.png

instead of this

* https://valkyriecrusade.fandom.com/wiki/File:Crimson_Hatsune_H.png

The `/wiki/File:$name` URL redirects to whatever wiki page contains the
image instead of showing the file itself.
2022-03-23 23:38:06 -05:00
evazion
4ef8178bd1 sources: remove canonical_url method.
Refactor source strategies to remove the `canonical_url` method.

`canonical_url` returned the URL that should be used as the source of
the post after upload. Now we simply use `Source::URL#page_url` to
determine the source after upload. If the source is an image URL that is
convertible to a page URL, then the image URL is used as the source. If
the source is an image URL that is not convertible to a page URL, then
the page URL is used as the source.

This simplifies source strategies so that all they have to care about is
implementing the `Source::URL#page_url` and `Sources::Strategies#page_url`
methods, and the preferred source will be chosen for posts automatically.
2022-03-23 23:38:06 -05:00
evazion
3aa5cab2aa sources: refactor normalize_for_source.
`normalize_for_source` was used to convert image URLs to page URLs when displaying sources
on the post show page. Move all the code for converting image URLs to page URLs from
`Sources::Strategies#normalize_for_source` to `Source::URL#page_url`.

Before we had to be very careful in source strategies not to make any network calls in
`normalize_for_source`, since it was used in the view for the post show page. Now all the
code for generating page URLs is isolated in Source::URL, which makes source strategies
simpler. It also makes it easier to check if a source is an image URL or page URL, and if
the image URL is convertible to a page URL, which will make autotagging bad_link or
bad_source feasible.

Finally, this fixes it to generate better page URLs in a handful of cases:

* https://www.artstation.com/artwork/qPVGP instead of https://anubis1982918.artstation.com/projects/qPVGP
* https://yande.re/post/show?md5=b4b1d11facd1700544554e4805d47bb6s instead of https://yande.re/post?tags=md5:b4b1d11facd1700544554e4805d47bb6
* http://gallery.minitokyo.net/view/365677 instead of http://gallery.minitokyo.net/download/365677
* https://valkyriecrusade.fandom.com/wiki/File:Crimson_Hatsune_H.png instead of https://valkyriecrusade.wikia.com/wiki/File:Crimson_Hatsune_H.png
* https://rule34.paheal.net/post/view/852405 instead of https://rule34.paheal.net/post/list/md5:854806addcd3b1246424e7cea49afe31/1
2022-03-23 01:34:04 -05:00
evazion
28971fe103 sources: factor out site_name method. 2022-03-11 23:20:53 -06:00
evazion
2f61486ac6 sources: remove image_url method from base strategy.
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.
2022-03-11 01:59:21 -06:00
evazion
043c08eb05 sources: factor out Source::URL::TwitPic. 2022-02-23 23:49:31 -06:00
nonamethanks
b3f92dd2c7 Fix gelbooru source normalization 2021-01-03 20:12:40 +01:00
nonamethanks
307df3b3e4 Refactor source normalization
* Move the source normalization logic out of the post model
  and into individual sources' strategies.
* Rewrite normalization tests to be handled into each source's test,
  and expand them significantly. Previously we were only testing
  a very small subset of domains and variants.
* Fix up normalization for several sites.
* Normalize fav.me urls into normal deviantart urls.
2020-05-21 22:46:51 +02:00
Albert Yi
6612aa1af9 fix unit tests 2019-04-25 18:13:59 -07:00
evazion
d693f01dde Fix #3859: Related tag and find artist don't run when fetch data fails.
Fixes an exception in the artist finder caused by searching for a nil profile_url.
2018-09-01 11:48:42 -05:00