From 5c64ced22dcf5579a502955ae1bbb013bfe121e7 Mon Sep 17 00:00:00 2001
From: Tinmarino <tinmarino@gmail.com>
Date: Fri, 14 Aug 2020 01:26:55 -0400
Subject: [PATCH] Feature: Utility function: linkify to extract link title from
 url (PR #982 from @jeromg)

---
 autoload/vimwiki/base.vim | 73 +++++++++++++++++++++++++++++++++++++++
 doc/vimwiki.txt           | 16 +++++++++
 2 files changed, 89 insertions(+)

diff --git a/autoload/vimwiki/base.vim b/autoload/vimwiki/base.vim
index 167d85d..8e2ae71 100644
--- a/autoload/vimwiki/base.vim
+++ b/autoload/vimwiki/base.vim
@@ -2828,6 +2828,79 @@ function! vimwiki#base#search(search_pattern) abort
   endtry
 endfunction
 
+" used by function linkify to extract web page <title>
+function! s:get_title(match) abort
+    " Do not overwrite if g:page_title is already set
+    " when there are multiple <title> tags, only use the first one
+    " this is a side effect of the substitute's 'n' flag (count number of
+    " occurences and evaluate \= for each one
+    if (g:page_title !=# '')
+        return
+    endif
+    let l:title = a:match
+
+    " cleanup title so it's compatible with vimwiki links
+    let l:title = substitute(l:title, '\\', '', 'g')
+    let l:title = substitute(l:title, '\[', '(', 'g')
+    let l:title = substitute(l:title, '\]', ')', 'g')
+
+    " cosmetic cleanup (html entities), maybe more to add
+    let l:title = substitute(l:title, '&lt;', '<', 'g')
+    let l:title = substitute(l:title, '&gt;', '>', 'g')
+    let l:title = substitute(l:title, '&nbsp;', ' ', 'g')
+
+    " store title in global var
+    let g:page_title = l:title
+endfunction
+
+" transform the url under the cursor to a wiki link
+function! vimwiki#base#linkify() abort
+    let g:page_title = ''
+
+    " save existing value of @u and delete url under the cursor into @u
+    let l:save_reg = @u
+    exe 'normal! "udiW'
+
+    " create a scratch buffer and switch to it
+    let current_buf = bufnr('')
+    let scratch_buf = bufnr('scratch',1)
+    exe 'sil! ' . scratch_buf . 'buffer'
+
+    " load web page into scratch buffer using Nread with mode=2
+    " FIXME: on Windows, with vim 7/8 (not with nvim), makes the cmd.exe window show up (annoying)
+    exe 'sil! :2Nread ' . @u
+
+    " extract title from html
+    " Note: if URL cannot be downloaded the buffer is empty or contains a single
+    " line: 'Not found'
+    let page_ok=0
+    if (wordcount().chars !=0 && getline(1) !=? 'Not found')
+        let page_ok=1
+        " regex seems to work fine, but may not cover all cases
+        exe 'sil! :keepp %s/\v\<title.{-}\>((.|\r)+)\<\/title\>/\=s:get_title(submatch(1))/n'
+    endif
+
+    " wipeout scratch buffer and switch to current
+    exe scratch_buf . 'bwipeout'
+    exe current_buf . 'buffer'
+
+    if (page_ok)
+        " use template [[URL|DESCRIPTION]]
+        let template = g:vimwiki_global_vars.WikiLinkTemplate2
+        let link = substitute(template, '__LinkUrl__', @u, '')
+        let link = substitute(link, '__LinkDescription__', g:page_title==#'' ? @u : g:page_title, '')
+        exe 'normal! i' . link
+    else
+        "if URL could not be downloaded, undo and display message
+        "TODO: other behaviours may be possible (user options?)
+        exe 'normal! u'
+        echomsg 'Error downloading URL: ' . @u
+    endif
+
+    " restore initial value of @u
+    let @u = l:save_reg
+endfunction
+
 
 " -------------------------------------------------------------------------
 " Load syntax-specific Wiki functionality
diff --git a/doc/vimwiki.txt b/doc/vimwiki.txt
index 0e1c85e..1b0fd6b 100644
--- a/doc/vimwiki.txt
+++ b/doc/vimwiki.txt
@@ -1189,6 +1189,19 @@ as a wiki page.
 To scan the page for new or changed definitions for reference-links, simply
 re-open the page ":e<CR>".
 
+Extract title from external links~
+
+It is possible to automagically extract the title of a web page to create a
+link. The function #vimwiki#base#linkify() will get the URL under the cursor
+and replace the url with a markdown link.
+
+If the URL cannot be retrieved, nothing is changed.
+If the <title> html tag cannot be extracted, the url is used as the link
+title.
+
+Suggested mapping: >
+  nnoremap <silent> <leader>uu :call vimwiki#base#linkify()<cr>
+
 ------------------------------------------------------------------------------
 5.3. Headers                                          *vimwiki-syntax-headers*
 
@@ -3762,6 +3775,7 @@ Contributors and their Github usernames in roughly chronological order:
     - Ryan Winograd
     - Birger Niklas (@BirgerNi)
     - Chip Senkbeil (@chipsenkbeil)
+    - Jerome Gay (@jeromg)
 
 
 ==============================================================================
@@ -3774,6 +3788,8 @@ http://code.google.com/p/vimwiki/issues/list. They may be accessible from
 https://github.com/vimwiki-backup/vimwiki/issues.
 
 New:~
+    * Feature: #837 extract web page <title> from URL under cursor and create
+      a nice wiki link
     * Feature: #922 #928: g:vimwiki_tag_format to change the tag format
     * Feature: Support Emoji (Conceal and Complete)
     * Issue #209: Feature: Markdown: Support SetExt Heading