Mirrors/gallery-dl
1
0
Fork 0
mirror of https://github.com/mikf/gallery-dl.git synced 2025-01-31 11:41:35 +01:00
Code Issues Releases Wiki Activity

[weasyl] add 'favorite' extractor (#1032)

Browse Source
This commit is contained in:
Mike Fährmann 2020-09-26 02:11:37 +02:00
parent e9cc719497
commit cc4ac80302
No known key found for this signature in database
GPG Key ID: 5680CA389D365A88
2 changed files with 42 additions and 1 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

2
docs/supportedsites.rst
View File

@ -129,7 +129,7 @@ Twitter https://twitter.com/ |twitter-C|
VSCO https://vsco.co/ Collections, individual Images, User Profiles
Wallhaven https://wallhaven.cc/ individual Images, Search Results Optional (`API Key <configuration.rst#extractorwallhavenapi-key>`__)
Warosu https://warosu.org/ Threads
Weasyl https://www.weasyl.com/ Folders, Journals, Submissions
Weasyl https://www.weasyl.com/ Favorites, Folders, Journals, Submissions
Webtoon https://www.webtoons.com/ Comics, Episodes
Weibo https://www.weibo.com/ Images from Statuses, User Profiles
WikiArt.org https://www.wikiart.org/ Artists, Artist Listings, Artworks

41
gallery_dl/extractor/weasyl.py
View File

@ -193,3 +193,44 @@ class WeasylJournalsExtractor(WeasylExtractor):
for journalid in text.extract_iter(page, 'href="/journal/', '/'):
data = self.retrieve_journal(journalid)
yield Message.Url, data["html"], data
class WeasylFavoriteExtractor(WeasylExtractor):
subcategory = "favorite"
directory_fmt = ("{category}", "{owner_login}", "Favorites")
pattern = BASE_PATTERN + r"favorites\?userid=(\d+)&feature=submit"
test = ("https://www.weasyl.com/favorites?userid=184616&feature=submit", {
"count": ">= 5",
})
def __init__(self, match):
WeasylExtractor.__init__(self, match)
self.userid = match.group(1)
def items(self):
owner_login = lastid = None
url = self.root + "/favorites"
params = {
"userid" : self.userid,
"feature": "submit",
}
while True:
page = self.request(url, params=params).text
pos = page.index('id="favorites-content"')
if not owner_login:
owner_login = text.extract(page, '<a href="/~', '"')[0]
yield Message.Directory, {"owner_login": owner_login}
for submitid in text.extract_iter(page, "/submissions/", "/", pos):
if submitid == lastid:
continue
lastid = submitid
submission = self.request_submission(submitid)
if self.populate_submission(submission):
yield Message.Url, submission["url"], submission
if "&amp;nextid=" not in page:
return
params["nextid"] = submitid