Skip to content

Commit 1dc7955

Browse files
committed
[simpcity] extract "click to load media" URLs (#8609)
1 parent 25ac509 commit 1dc7955

File tree

2 files changed

+18
-5
lines changed

2 files changed

+18
-5
lines changed

gallery_dl/extractor/simpcity.py

Lines changed: 10 additions & 5 deletions
Original file line numberDiff line numberDiff line change
@@ -30,11 +30,14 @@ def items(self):
3030
self.login()
3131

3232
extract_urls = text.re(
33-
r'(?s)<(?:'
34-
r'video (.*?\ssrc="[^"]+".*?)</video>'
35-
r'|a [^>]*?href="(?:https://[^"]+)?(/attachments/[^"]+".*?)</a>'
36-
r'|div [^>]*?ata-src="(?:https://[^"]+)?(/attachments/[^"]+".*?)/>'
37-
r'|(?:a [^>]*?href|iframe [^>]*?src)="([^"]+)'
33+
r'(?s)(?:'
34+
r'<video (.*?\ssrc="[^"]+".*?)</video>'
35+
r'|<a [^>]*?href="'
36+
r'(?:https://[^"]+)?(/attachments/[^"]+".*?)</a>'
37+
r'|<div [^>]*?data-src="'
38+
r'(?:https://[^"]+)?(/attachments/[^"]+".*?)/>'
39+
r'|(?:<a [^>]*?href="|<iframe [^>]*?src="|'
40+
r'''onclick="loadMedia\(this, ')([^"']+)'''
3841
r')'
3942
).findall
4043

@@ -53,6 +56,8 @@ def items(self):
5356
data["num"] += 1
5457
data["num_external"] += 1
5558
data["type"] = "external"
59+
if ext.startswith("//"):
60+
ext = "https:" + ext
5661
yield Message.Queue, ext, data
5762

5863
elif video:

test/results/simpcity.py

Lines changed: 8 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -205,6 +205,14 @@
205205
},
206206
},
207207

208+
{
209+
"#url" : "https://simpcity.cr/threads/lustn4lexi-hot4lexi-lexi-2-legit-hott4lexi-lexi.175167/post-2512729",
210+
"#comment" : "'Click here to load redgifs media' (#8609)",
211+
"#class" : simpcity.SimpcityPostExtractor,
212+
"#auth" : True,
213+
"#results" : "https://redgifs.com/ifr/unusedsubmissivemullet",
214+
},
215+
208216
{
209217
"#url" : "https://simpcity.cr/threads/alua-tatakai.89490/",
210218
"#class" : simpcity.SimpcityThreadExtractor,

0 commit comments

Comments
 (0)