Skip to content

Latest commit

 

History

History
135 lines (96 loc) · 3.98 KB

0178-scriba.org

File metadata and controls

135 lines (96 loc) · 3.98 KB

scriba

The previous post was about Geneva - the CL documentation system. Vasily Postnicov mentioned on twitter another Lisp documentation tool called Codex.

Codex is based on Scriba markup and today we’ll look at how to use it.

POFTHEDAY> (scriba.parser:parse-string "Blah minor")
(:DOCUMENT "Blah minor")


POFTHEDAY> (scriba.parser:parse-string "@b(Blah) minor")
(:DOCUMENT (:NAME "b" :ATTRS NIL :BODY ("Blah")) " minor")

But this is an internal AST representation. Scriba is based on other Fernando Boretti’s library - CommonDoc and they should be used in tandem:

POFTHEDAY> (make-instance 'scriba:scriba)
#<SCRIBA:SCRIBA {10023FD193}>

POFTHEDAY> (defparameter *format* (make-instance 'scriba:scriba))
*FORMAT*

POFTHEDAY> (common-doc.format:parse-document *format* "@b(Hello) @i(World)!")
#<COMMON-DOC:DOCUMENT "">

POFTHEDAY> (common-doc.format:emit-to-string *format* *)
"@title()

@b(Hello)@i(World)!
"

There is also another CL library, also made by Fernando - pandocl. It can be used when you need to convert the document into HTML or another format.

For example, here how we can render our hello world into the HTML:

POFTHEDAY> (pandocl:parse-string
            "@b(Hello) @i(World)!"
            :scriba)
#<COMMON-DOC:DOCUMENT "">

POFTHEDAY> (pandocl:emit  * "hello.html")
#<COMMON-DOC:DOCUMENT "">

POFTHEDAY> (alexandria:read-file-into-string "hello.html")
"<!DOCTYPE html>
 <html>
   <head><title></title></head>
   <body>
     <b>Hello</b><i>World</i>!
   </body>
 </html>"

But what I’m really interested in is Scriba’s extensibility. The text consists of inline and multiline blocks. Each block has a name and optional attributes.

Let’s pretend, everyday we are writing texts, mentioning different twitter users and want a shorthand syntax for them! When rendering into HTML, these tags should be transformed into the link and real user name.

Naive approach does not work, because we did nothing to extend the protocol:

POFTHEDAY> (common-doc.format:parse-document *format*
                                             "Hello @twitter(bob)!")
; Debugger entered on #<SIMPLE-ERROR "No node with name twitter" {100324FCC3}>

But I found the way to do this. Thanks to the CLOS!

What we need, is to define the twitter node, using this CommonDoc macro:

POFTHEDAY> (common-doc:define-node twitter (common-doc:markup)
                                   ()
                                   (:tag-name "twitter"))

POFTHEDAY> (common-html.emitter::define-emitter (node twitter)
             (let* ((username (common-doc:text (first (common-doc:children node))))
                    (url (format nil "https://twitter/~A" username))
                    (name (format nil "@~A" username)))
               (common-html.emitter::with-tag ("a" node :attributes `(("href" . ,url)))
                                              (write-string name
                                                            common-html.emitter::*output-stream*))))

POFTHEDAY> (common-doc.format:parse-document *format* "Hello @twitter(bob)!")
#<COMMON-DOC:DOCUMENT "">

POFTHEDAY> (pandocl:emit  * "hello.html")
#<COMMON-DOC:DOCUMENT "">

POFTHEDAY> (alexandria:read-file-into-string "hello.html")
"<!DOCTYPE html>
 <html>
    <head><title></title></head>
    <body>
       Hello <a href=\"https://twitter/bob\">@bob</a>!
    </body>
</html>"

As you can see, I’ve used a bunch of internal symbols, to extend Common HTML and make it work the way I need. Probably it will be a good idea to make this API public.

Anyway, I like Scriba and Common Doc because it was relatively easy to hack and do what I need.