github/gallery-dl
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

[kemonoparty] extract inline images (fixes #1286)

Browse Source
This commit is contained in:
Mike Fährmann
2021-02-05 20:10:20 +01:00
parent a041a017d1
commit 2cc1e04fe5
1 changed files with 34 additions and 21 deletions
Download Patch File Download Diff File Expand all files Collapse all files

55
gallery_dl/extractor/kemonoparty.py
View File

@@ -10,6 +10,7 @@
from .common import Extractor, Message from .common import Extractor, Message
from .. import text from .. import text
import re
class KemonopartyExtractor(Extractor): class KemonopartyExtractor(Extractor):
@@ -21,6 +22,8 @@ class KemonopartyExtractor(Extractor):
archive_fmt = "{user}_{id}_{filename}.{extension}" archive_fmt = "{user}_{id}_{filename}.{extension}"
def items(self): def items(self):
find_inline = re.compile(r'src="(/inline/[^"]+)').findall
for post in self.posts(): for post in self.posts():
files = [] files = []
@@ -28,6 +31,9 @@ class KemonopartyExtractor(Extractor):
files.append(post["file"]) files.append(post["file"])
if post["attachments"]: if post["attachments"]:
files.extend(post["attachments"]) files.extend(post["attachments"])
for path in find_inline(post["content"] or ""):
files.append({"path": path, "name": path})
post["date"] = text.parse_datetime( post["date"] = text.parse_datetime(
post["published"], "%a, %d %b %Y %H:%M:%S %Z") post["published"], "%a, %d %b %Y %H:%M:%S %Z")
yield Message.Directory, post yield Message.Directory, post
@@ -68,27 +74,34 @@ class KemonopartyPostExtractor(KemonopartyExtractor):
"""Extractor for a single kemono.party post""" """Extractor for a single kemono.party post"""
subcategory = "post" subcategory = "post"
pattern = r"(?:https?://)?kemono\.party/([^/?#]+)/user/(\d+)/post/(\d+)" pattern = r"(?:https?://)?kemono\.party/([^/?#]+)/user/(\d+)/post/(\d+)"
test = ("https://kemono.party/fanbox/user/6993449/post/506575", { test = (
"pattern": r"https://kemono\.party/files/fanbox" ("https://kemono.party/fanbox/user/6993449/post/506575", {
r"/6993449/506575/P058kDFYus7DbqAkGlfWTlOr\.jpeg", "pattern": r"https://kemono\.party/files/fanbox"
"keyword": { r"/6993449/506575/P058kDFYus7DbqAkGlfWTlOr\.jpeg",
"added": "Wed, 06 May 2020 20:28:02 GMT", "keyword": {
"content": str, "added": "Wed, 06 May 2020 20:28:02 GMT",
"date": "dt:2019-08-11 02:09:04", "content": str,
"edited": None, "date": "dt:2019-08-11 02:09:04",
"embed": dict, "edited": None,
"extension": "jpeg", "embed": dict,
"filename": "P058kDFYus7DbqAkGlfWTlOr", "extension": "jpeg",
"id": "506575", "filename": "P058kDFYus7DbqAkGlfWTlOr",
"num": 1, "id": "506575",
"published": "Sun, 11 Aug 2019 02:09:04 GMT", "num": 1,
"service": "fanbox", "published": "Sun, 11 Aug 2019 02:09:04 GMT",
"shared_file": False, "service": "fanbox",
"subcategory": "post", "shared_file": False,
"title": "c96取り置き", "subcategory": "post",
"user": "6993449", "title": "c96取り置き",
}, "user": "6993449",
}) },
}),
# inline image (#1286)
("https://kemono.party/fanbox/user/7356311/post/802343", {
"pattern": r"https://kemono\.party/inline/fanbox"
r"/uaozO4Yga6ydkGIJFAQDixfE\.jpeg",
}),
)
def __init__(self, match): def __init__(self, match):
KemonopartyExtractor.__init__(self, match) KemonopartyExtractor.__init__(self, match)