sxml: avoiding intermixed text?

sxml: avoiding intermixed text?

[Aleix Conchillo Flaqué]

Hi!

is there an easy way to avoid intermixed text from xml? For example, 
given this XML:

<?xml version="1.0"?>

<test>
   <check><value>empty</value></check>
   <check>
     <value>hi!
     this is just a test!
     </value>
   </check>
</test>

the following code:

-------------
(use-modules (sxml simple))
(use-modules (sxml xpath))

(let ((sxml (with-input-from-file "test.xml" xml->sxml)))
   (map
    (lambda (c) (pk (cdr c)))
    ((sxpath '(// check)) sxml)))
-------------

Returns:

-------------
;;; (((value "empty")))

;;; (("
     " (value "hi!
     this is just a test!
     ") "
   "))
-------------

I'd like to get rid of the (car) of the second print. Or saying it in 
another way, I want to get rid of all text that is not a leaf.

I'm using xpath to find my elements, which is what I found easiest from 
all the sxml stuff. May be I should do it some other way.

Any help would be welcome!

Cheers,

Aleix

[SOLVED] Re: sxml: avoiding intermixed text?

[Aleix Conchillo Flaqué]

When I asked the question I thought I could use some sxml function to do 
this job. Particularly, I found out a way to do it with the SXML 
modifications explained here:

http://www.modis.ispras.ru/Lizorkin/sxml-tutorial.html#hevea:modif

Unfortunately this is not available in Guile. So, This function removes 
whitespace text nodes from an sxml.

(define (remove-whitespace-nodes sxml)
   (define (node-fix node)
     (cond ((symbol? node) (list node))
           ((string? node) (if (string-null? (string-trim node))
                               #nil
                               (list node)))
           (else (list (remove-whitespace-nodes node)))))
   (let loop ((node sxml) (result '()))
     (cond ((null? node) result)
           (else (loop (cdr node)
                       (append result (node-fix (car node))))))))

I'm new to scheme, so a code review is really welcome!

Cheers,

Aleix

Re: [SOLVED] Re: sxml: avoiding intermixed text?

[Thien-Thi Nguyen]

() Aleix Conchillo Flaqué <aconchillo@gmail.com>
() Thu, 13 Jan 2011 11:47:26 +0100

   I'm new to scheme, so a code review is really welcome!

I would replace

     (let loop ((node sxml) (result '()))
       (cond ((null? node) result)
             (else (loop (cdr node)
                         (append result (node-fix (car node)))))))

with

     (map node-fix sxml)

Re: [SOLVED] Re: sxml: avoiding intermixed text?

[Aleix Conchillo Flaqué]

On 01/13/2011 11:53 AM, Thien-Thi Nguyen wrote:
> I would replace
>
>       (let loop ((node sxml) (result '()))
>         (cond ((null? node) result)
>               (else (loop (cdr node)
>                           (append result (node-fix (car node)))))))
>
> with
>
>       (map node-fix sxml)

This would only change whitespaces for #nil... but...

(delete #nil (map node-fix sxml))

works solves this.

Thank you very much!

Aleix