cobra
/
MeMe
1
1
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity

Fix meme pages that are not a part of a series 

Signed-off-by: Skylar "The Cobra" Widulski <cobra@vern.cc>
This commit is contained in:
Skylar "The Cobra" Widulski 2023-12-07 11:46:48 -05:00
parent 11d77d435e
commit b174194068
Signed by: cobra
GPG Key ID: 4FD8F812083FF6F9
1 changed files with 13 additions and 5 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

18
meme/scraping.scm
View File

@ -190,6 +190,7 @@
(define lst1 (list ""))
(define skip #f)
(while (not (null-pointer? chld))
(name chld)
(cond
(skip
(set! skip #f)))
@ -254,8 +255,15 @@
"/html/body/div[3]/div/article/header/section/h1/text()" body #f)
(get-xpath-string
"/html/body/div[3]/div/article/header/a/img/@src" body #t)
(dump-xpath-xml
"/html/body/div[3]/div/article/header/section/div/h5" body)
(if (null-pointer?
(caddr (parse-c-struct
(get-xpath-nodeset
"/html/body/div[3]/div/article/header/section/div/h5"
body)
xml-nodeset)))
""
(dump-xpath-xml
"/html/body/div[3]/div/article/header/section/div/h5" body))
(dump-xpath-xml
"/html/body/div[3]/div/article/header/section/div/footer/p[1]" body)
(dump-xpath-xml
@ -326,7 +334,7 @@
(list
(dump-xpath-xml "/html/body/div[3]/div/div[1]/header/hgroup/h1" body)
(if (null-pointer?
(caddr (parse-c-struct
(caddr (parse-c-struct
(get-xpath-nodeset
"/html/body/div[3]/div/div[1]/header/hgroup/p"
body)
@ -463,7 +471,7 @@
body #t)
(get-xpath-string
(string-append
"/html/body/div[3]/aside/div["
"/html/body/div[3]/aside/div["
(number->string i)
"]/table/tbody/tr[2]/td[2]/a/img/@data-src")
body #t)
@ -516,7 +524,7 @@
(get-xpath-string
"/html/body/div[3]/aside/div[4]/table/tbody/tr[2]/td[3]/a/img/@data-src"
body #t)
(get-xpath-string
"/html/body/div[3]/aside/div[4]/table/tbody/tr[3]/td[1]/a/@href"
body #t)