y0rune/vimwiki
1
0
Fork 0
Code Issues Pull Requests Releases Wiki Activity

Remove all protocols and tlds on URL-cleaning

Browse Source 
Also make removal of protocol, tld and 'www' more robust against
fails-positives by considering the position in the url.
This commit is contained in:
Michael F. Schönitzer 2018-06-17 16:10:44 +02:00
parent 4d4e88bfa2
commit 50c95f1bef
1 changed files with 10 additions and 13 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

23
autoload/vimwiki/base.vim
View File

@ -1858,20 +1858,17 @@ endfunction
function! s:clean_url(url) function! s:clean_url(url)
let url = split(a:url, '/\|=\|-\|&\|?\|\.') " remove protocol and tld
let url = substitute(a:url, '^\a\+://', '', '')
let url = substitute(url, '^\([^/]\+\).\a\{2,4}/', '\1/', '')
let url = split(url, '/\|=\|-\|&\|?\|\.')
let url = filter(url, 'v:val !=# ""') let url = filter(url, 'v:val !=# ""')
let url = filter(url, 'v:val !=# "www"') if url[0] == "www"
let url = filter(url, 'v:val !=# "com"') let url = url[1:]
let url = filter(url, 'v:val !=# "org"') endif
let url = filter(url, 'v:val !=# "net"') if url[-1] =~ '^\(htm\|html\|php\)$'
let url = filter(url, 'v:val !=# "edu"') let url = url[0:-2]
let url = filter(url, 'v:val !=# "http\:"') endif
let url = filter(url, 'v:val !=# "https\:"')
let url = filter(url, 'v:val !=# "file\:"')
let url = filter(url, 'v:val !=# "xml\:"')
let url = filter(url, 'v:val !=# "html"')
let url = filter(url, 'v:val !=# "htm"')
let url = filter(url, 'v:val !=# "php"')
" remove words consisting of only hexadecimal digits or non-word characters " remove words consisting of only hexadecimal digits or non-word characters
let url = filter(url, 'v:val !~ "^\\A\\{4,}$"') let url = filter(url, 'v:val !~ "^\\A\\{4,}$"')
let url = filter(url, 'v:val !~ "^\\x\\{4,}$" || v:val !~ "\\d"') let url = filter(url, 'v:val !~ "^\\x\\{4,}$" || v:val !~ "\\d"')