How to match this string with elisp regexp?

Question

I'm writing an extension to open a custom link type in Org-mode with Kiwix URL. (Kiwix is an offline Wikipedia program). Here is my kiwix.el source code.

I created a wiki: link type for Org-mode.

Here is the link formats:

• [[wiki:(library):query]]
• [[wiki:query]]

links:

• wiki:(zh):%E7%A6%85%E5%AE%97
• wiki:(en):linux
• wiki:linux

And here is the elisp regexp:

"\\(wiki\\):\\((\\(.*\\)):\\)?\\(.*\\)"

Here is my current function to handle the link:

(defun org-wiki-link-open (link)
  "Open LINK in external wiki program."
  (cond ((string-match "\\(wiki\\):\\((\\(.*\\)):\\)?\\(.*\\)"  link) ; [[wiki:(library):query]]
         (let* ((type (match-string 1 link))
                ;; convert between libraries full name and abbrev.
                (library (cdr (assoc
                               (match-string 3 link)
                               kiwix-librarie-abbrev-list)))
                (query (match-string 4 link))
                (url (concat kiwix-server-url library "/A/" (url-encode-url (capitalize query)) ".html")))
           (browse-url url)))
        ((string-match "\\(wiki\\):\\(.*\\)"  link) ; [[wiki:query]]
         (let* ((type (match-string 1 link))
                (query (match-string 4 link))
                (url (concat kiwix-server-url kiwix-default-library "/A/" (url-encode-url (capitalize query)) ".html")))
           (browse-url url)))))

(defun org-wiki-link-export (link description format)
  "Export the wiki LINK with DESCRIPTION for FORMAT from Org files."
  (let* ((type (when (string-match "\\(.+\\):(\\(.+\\)?):\\(.*\\)" link)
                 (match-string 1 link)))
         (library (when (string-match "\\(.+\\):(\\(.+\\)?):\\(.*\\)" link)
                    (match-string 2 link)))
         ;; query need to be convert to URL encoding: "禅宗" https://zh.wikipedia.org/wiki/%E7%A6%85%E5%AE%97
         (query (url-encode-url
                 (or description
                     (when (string-match "\\(.+\\):(\\(.+\\)?):\\(.*\\)" link)
                       (match-string 3 link)))))
         ;; "http://en.wikipedia.org/wiki/Linux"
         ;;         --
         ;;          ^- library: en, zh
         (path (concat "http://" library ".wikipedia.org/wiki/" query))
         (desc query))
    (when (stringp path)
      (cond
       ((eq format 'html) (format "<a href=\"%s\">%s</a>" path desc))
       ((eq format 'latex) (format "\\href{%s}{%s}" path desc))
       (t path)))))

(org-add-link-type "wiki" 'org-wiki-link-open 'org-wiki-link-export)

But this regexp is wrong, can't open those example links correctly.

• [[wiki:(en):Linux][Linux (wiki)]]
• [[wiki:Linux][Linux]]
• [[wiki:(zh):%E7%A6%85%E5%AE%97][禅宗]]

So I hope someone can help me to improve this regexp pattern and my function code.

Answer 1

Your regexp should be \\(wiki\\):\\(?:(\\([^:]*\\)):\\)?\\([^] \n\t\r]*\\).

You have with that:

• group 1: wiki
• non capturing group (\(?:...\)) for optional library
• group 2: library (en or zh)
• group 3: link? (match everything but ], space, tab, carriage return, linefeed by using [^] \n\t\r]*)

So, you have one group less (your code should be adapted to that).

As an optimization you should string-match once and use subsequent match-string for each group of interest.