# (RE)XML question

**URL:** <https://www.ruby-forum.com/t/re-xml-question/66043>\
**Category:** Ruby\
**Created:** [August 23, 2006, 12:16am UTC](https://www.ruby-forum.com/t/re-xml-question/66043 "2006-08-23T00:16:02Z")\
**Posts on this page:** 8\
**Page:** 1

<div class="post-metadata">

**Author:** ![unknown](https://www.ruby-forum.com/letter_avatar_proxy/v4/letter/u/4bbf92/32.png) [@unknown](https://www.ruby-forum.com/u/unknown)\
**Post date:** [August 23, 2006, 12:16am UTC](https://www.ruby-forum.com/t/re-xml-question/66043/1 "2006-08-23T00:16:02Z")

</div>

Question for you all. I want to treat HTML like XML  
(which is no big deal).

But I want to find certain “special” tags (not real  
HTML) and replace them with my own text.

It’s macro-type stuff. Basically I want to output  
the _same_ HTML except for the text that replaced  
the special tags.

I can’t find any examples of generating XML with  
REXML. It should be easy, I don’t want it to be  
too hard.

Contrived example below in case it helps.

How would you do this?

Thanks,  
Hal

Input:

Hi, there.

some more text

That's all.

Output:

Hi, there.

I found a foo tag enclosing 'some more text' with bar and bam values of 'this' and 'that'...

That's all.

---

<div class="post-metadata">

**Author:** ![unknown](https://www.ruby-forum.com/letter_avatar_proxy/v4/letter/u/4bbf92/32.png) [@unknown](https://www.ruby-forum.com/u/unknown)\
**Post date:** [August 23, 2006, 1:10am UTC](https://www.ruby-forum.com/t/re-xml-question/66043/2 "2006-08-23T01:10:11Z")

</div>

On Wed, Aug 23, 2006 at 07:15:09AM +0900, removed\_email\_address@domain.invalid wrote:

> Output:
> 
> Hi, there.
> 
> I found a foo tag enclosing 'some more text' with bar and bam values of 'this' and 'that'...
> 
> That's all.

So, in Hpricot:

doc = Hpricot("…")  
doc.search(“foo”).each do |ele|  
new\_ele = Hpricot   
‘

I found a ’ + ele.name + " tag enclosing ‘" +  
ele.inner\_html + "’ with " + ele.attributes.keys.join(’ and ‘) +  
" values of " + ele.attributes.values.map { |x| "’#{x}’" }.join(’  
and ') +  
“…

”  
ele.parent.replace\_child(ele, new\_ele.children.first)  
end  
puts doc

REXML has a replace\_child as well. But now you’ve motivated me to add  
Element#replace.

\_why

---

<div class="post-metadata">

**Author:** ![unknown](https://www.ruby-forum.com/letter_avatar_proxy/v4/letter/u/4bbf92/32.png) [@unknown](https://www.ruby-forum.com/u/unknown)\
**Post date:** [August 23, 2006, 2:30am UTC](https://www.ruby-forum.com/t/re-xml-question/66043/3 "2006-08-23T02:30:17Z")

</div>

why the lucky stiff wrote:

> ```
> ele.parent.replace_child(ele, new_ele.children.first)
> 
> ```
> 
> end  
> puts doc
> 
> REXML has a replace\_child as well. But now you’ve motivated me to add Element#replace.

Hmm, the right thing to do and a tasty way to do it.

This motivates me to download Hpricot for the first time  
and try it. Probably tomorrow as my brane is fride.

Thanks,  
Hal

---

<div class="post-metadata">

**Author:** ![unknown](https://www.ruby-forum.com/letter_avatar_proxy/v4/letter/u/4bbf92/32.png) [@unknown](https://www.ruby-forum.com/u/unknown)\
**Post date:** [August 23, 2006, 9:39pm UTC](https://www.ruby-forum.com/t/re-xml-question/66043/4 "2006-08-23T21:39:53Z")

</div>

unknown wrote:

> It’s macro-type stuff. Basically I want to output  
> the _same_ HTML except for the text that replaced  
> the special tags.

This is what XSLT was designed for and it may provide another option for  
you…

ilan

---

<div class="post-metadata">

**Author:** ![unknown](https://www.ruby-forum.com/letter_avatar_proxy/v4/letter/u/4bbf92/32.png) [@unknown](https://www.ruby-forum.com/u/unknown)\
**Post date:** [August 23, 2006, 10:46pm UTC](https://www.ruby-forum.com/t/re-xml-question/66043/5 "2006-08-23T22:46:34Z")

</div>

removed\_email\_address@domain.invalid wrote:

> Contrived example below in case it helps.

input = \<\<ENDHTML

Hi, there.

some more text

That's all.

ENDHTML

require ‘rexml/document’  
doc = REXML::Document.new( input )  
doc.root.each\_element( ‘//foo’ ){ |e|  
new\_para = REXML::Element.new( ‘p’ )  
new\_para.text = “I found a foo tag enclosing ‘#{e.text}’ with bar and  
bam values of ‘#{e.attributes[‘bar’]}’ and ‘#{e.attributes[‘bam’]}’…”  
e.parent.replace\_child( e, new\_para )  
}  
puts doc

#=\>   
#=\>   
#=\>

Hi, there.

  
#=\> 

I found a foo tag enclosing ‘some more text’ with bar  
and bam values of ‘this’ and ‘that’…

  
#=\> 

That’s all.

  
#=\>   
#=\>

---

<div class="post-metadata">

**Author:** ![unknown](https://www.ruby-forum.com/letter_avatar_proxy/v4/letter/u/4bbf92/32.png) [@unknown](https://www.ruby-forum.com/u/unknown)\
**Post date:** [August 23, 2006, 10:50pm UTC](https://www.ruby-forum.com/t/re-xml-question/66043/6 "2006-08-23T22:50:35Z")

</div>

Ilan B. wrote:

> > It’s macro-type stuff. Basically I want to output  
> > the _same_ HTML except for the text that replaced  
> > the special tags.
> 
> This is what XSLT was designed for and it may provide another option for  
> you…

That makes sense. I’ve never used XSLT, but I’m sure that’s  
a viable solution.

\_Why’s Hpricot example worked perfectly for me, BTW.

So, a related question.

Suppose I wanted to “nest” macros of this kind. Something like:

\<mac1 foo=“1” bar="2\>My name is  
seed-value  
today.

Forgive the nonsense example.

Could XSLT handle this easily? Could Hpricot (\_why)?

Thanks,  
Hal

---

<div class="post-metadata">

**Author:** ![unknown](https://www.ruby-forum.com/letter_avatar_proxy/v4/letter/u/4bbf92/32.png) [@unknown](https://www.ruby-forum.com/u/unknown)\
**Post date:** [August 23, 2006, 11:00pm UTC](https://www.ruby-forum.com/t/re-xml-question/66043/7 "2006-08-23T23:00:36Z")

</div>

removed\_email\_address@domain.invalid wrote:

> I can’t find any examples of generating XML with
> 
> Output:
> 
> Hi, there.
> 
> I found a foo tag enclosing 'some more text' with bar and bam values of 'this' and 'that'...
> 
> That's all.

require ‘xml-split.rb’

tag = ‘foo’  
DATA.read.xml\_split(tag).each {|stuff|  
if stuff.class == String  
print stuff  
else  
attr = stuff[0].xml\_parse  
puts “

I found a #{tag} tag enclosing ‘#{stuff[1]}’ with”  
print “#{attr.keys.join(’ and ‘)} values of "  
print "’#{attr.values.join(”’ and ‘")}’…

"  
end  
}

**END**

Hi, there.

some more text

That's all.

---- output ----

Hi, there.

I found a foo tag enclosing 'some more text' with bam and bar values of 'that' and 'this'...

That's all.

---

<div class="post-metadata">

**Author:** ![unknown](https://www.ruby-forum.com/letter_avatar_proxy/v4/letter/u/4bbf92/32.png) [@unknown](https://www.ruby-forum.com/u/unknown)\
**Post date:** [August 24, 2006, 3:36pm UTC](https://www.ruby-forum.com/t/re-xml-question/66043/8 "2006-08-24T15:36:30Z")

</div>

unknown wrote:

> \_Why’s Hpricot example worked perfectly for me, BTW.
> 
> So, a related question.
> 
> Suppose I wanted to “nest” macros of this kind. Something like:
> 
> \<mac1 foo=“1” bar="2\>My name is  
> seed-value  
> today.
> 
> Forgive the nonsense example.
> 
> Could XSLT handle this easily? Could Hpricot (\_why)?
> 
> Thanks,  
> Hal

Yes, both techniques could handle nested elements, I don’t know what XML  
tools you are using, but many come with XSLT support built in. XSLT  
allows any XML(XHTML) doc to be transformed into any other. At one time  
it was slated to replace .CSS but that never seeemed to materialize.  
Now days, it’s mostly used in report generation and xml rpc filtering  
but it ofcourse has many uses. The disadvantages of XSLT is that it can  
be rather challenging to debug and it can grow to be very verbose in non  
trivial transformations. The advantage is that it is a W3C standard and  
practically every platform/language has support for it in one form or  
another.

I have no experience of Hpricot but if you are already using Ruby as  
your main processor then I would probably stick with Hpricot as the  
solutions above look much cleaner than an XSLT solution 🙂 Oh… and  
lastly, if you don’t use XSLT/XPath on a regular basis, you can easily  
forget it’s symantics and have to keep referring back to the docs or at  
least I have to.

ilan