Mirrors/gallery-dl
1
0
Fork 0
mirror of https://github.com/mikf/gallery-dl.git synced 2024-11-22 10:42:34 +01:00
Code Issues Releases Wiki Activity

[naver] recognize '.naver' URLs

Browse Source 
https://blog.naver.com/PostView.naver?…
This commit is contained in:
Mike Fährmann 2024-03-05 22:30:29 +01:00
parent 36fc510d3a
commit 7b28418f69
No known key found for this signature in database
GPG Key ID: 5680CA389D365A88
2 changed files with 18 additions and 5 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

11
gallery_dl/extractor/naver.py
View File

@ -26,7 +26,8 @@ class NaverPostExtractor(NaverBase, GalleryExtractor):
"{post[date]:%Y-%m-%d} {post[title]}") "{post[date]:%Y-%m-%d} {post[title]}")
archive_fmt = "{blog[id]}_{post[num]}_{num}" archive_fmt = "{blog[id]}_{post[num]}_{num}"
pattern = (r"(?:https?://)?blog\.naver\.com/" pattern = (r"(?:https?://)?blog\.naver\.com/"
r"(?:PostView\.nhn\?blogId=(\w+)&logNo=(\d+)|(\w+)/(\d+)/?$)") r"(?:PostView\.n(?:aver|hn)\?blogId=(\w+)&logNo=(\d+)|"
r"(\w+)/(\d+)/?$)")
example = "https://blog.naver.com/BLOGID/12345" example = "https://blog.naver.com/BLOGID/12345"
def __init__(self, match): def __init__(self, match):
@ -73,7 +74,8 @@ class NaverBlogExtractor(NaverBase, Extractor):
subcategory = "blog" subcategory = "blog"
categorytransfer = True categorytransfer = True
pattern = (r"(?:https?://)?blog\.naver\.com/" pattern = (r"(?:https?://)?blog\.naver\.com/"
r"(?:PostList.nhn\?(?:[^&#]+&)*blogId=([^&#]+)|(\w+)/?$)") r"(?:PostList\.n(?:aver|hn)\?(?:[^&#]+&)*blogId=([^&#]+)|"
r"(\w+)/?$)")
example = "https://blog.naver.com/BLOGID" example = "https://blog.naver.com/BLOGID"
def __init__(self, match): def __init__(self, match):
@ -81,12 +83,11 @@ class NaverBlogExtractor(NaverBase, Extractor):
self.blog_id = match.group(1) or match.group(2) self.blog_id = match.group(1) or match.group(2)
def items(self): def items(self):
# fetch first post number # fetch first post number
url = "{}/PostList.nhn?blogId={}".format(self.root, self.blog_id) url = "{}/PostList.nhn?blogId={}".format(self.root, self.blog_id)
post_num = text.extract( post_num = text.extr(
self.request(url).text, 'gnFirstLogNo = "', '"', self.request(url).text, 'gnFirstLogNo = "', '"',
)[0] )
# setup params for API calls # setup params for API calls
url = "{}/PostViewBottomTitleListAsync.nhn".format(self.root) url = "{}/PostViewBottomTitleListAsync.nhn".format(self.root)

12
test/results/naver.py
View File

@ -24,6 +24,12 @@ __tests__ = (
"#sha1_metadata": "a6e23d19afbee86b37d6e7ad934650c379d2cb1e", "#sha1_metadata": "a6e23d19afbee86b37d6e7ad934650c379d2cb1e",
}, },
{
"#url" : "https://blog.naver.com/PostView.naver?blogId=rlfqjxm0&logNo=221430673006",
"#category": ("", "naver", "post"),
"#class" : naver.NaverPostExtractor,
},
{ {
"#url" : "https://blog.naver.com/gukjung", "#url" : "https://blog.naver.com/gukjung",
"#category": ("", "naver", "blog"), "#category": ("", "naver", "blog"),
@ -42,4 +48,10 @@ __tests__ = (
"#count" : 12, "#count" : 12,
}, },
{
"#url" : "https://blog.naver.com/PostList.naver?blogId=gukjung",
"#category": ("", "naver", "blog"),
"#class" : naver.NaverBlogExtractor,
},
) )