cleaning up msword HTML export
webdev_aw_ucsb
I am trying to cleanup an MSWord HTML export to something much cleaner
for consumption by TinyMCE.
I would like to take the orignal dirty html file remove some generic
MSWord classes -- and in a few instances create my own wrapper
classes ...
Here are my 2 problems:
1) I want to cleanup the HTML by removing unwanted attributes & styles
and then output a simplified HTML page:
unwanted_class_node = doc.search("//p[@class ='MsoNormal']")
unwanted_class_node.each do |n|
n.remove_attribute('class')
end
@builder = Nokogiri::HTML::Builder.new do |dest|
dest.html {
dest.head {}
dest.script {}
dest.style {}
dest.body {
dest.text(
#doc.xpath('//body').children.to_html
doc.xpath('//body').inner_html
)
}
}
end
File.open('cleaned.html', 'w') {|f| f.write(@builder.to_html) }
But in the above example code the resulted cleaned.html is contains
HTML-literals < instead of < within the destination files <body>
tag.
2) For some block element(s) in the source HTML there is:
<p>blah blhal bhalb blah blah<br><span style="mso-spacerun:
yes"> </span>foo foo foo foo foo foo foo</p>
I would like to transform the above <p> to essentially:
<p>blah blhal bhalb blah blah<br><span class="poetry-line-odd">foo foo
foo foo foo foo foo foo</span></p>
But I'm having trouble accomplishing (2) as this is my first exposure
to XPath / Nokogiri -- but I believe the tools are in my hands to get
this done ... and I think this "might work" (TM).
Thanks for your time,
David
Aaron Patterson
Yes, when you call the text method, it assumes that you want to insert
text and will escape characters for you. I would recommend copying
the nodes from one document to another like this:
doc1 = Nokogiri::HTML(<<-eohtml)
<html>
<body><h1>Hello World</h1> how are you?</body>
</html>
eohtml
dest = Nokogiri::HTML::Builder.new do |b|
b.html do
b.head do
b.script
b.style
end
b.body
end
end.doc
# Get the destination body
body = dest.at('body')
# Add the children of the source body to the destination body
doc1.at('body').children.each { |c| body << c }
puts dest
> 2) For some block element(s) in the source HTML there is:
> <p>blah blhal bhalb blah blah<br><span style="mso-spacerun:
> yes"> </span>foo foo foo foo foo foo foo</p>
>
> I would like to transform the above <p> to essentially:
> <p>blah blhal bhalb blah blah<br><span class="poetry-line-odd">foo foo
> foo foo foo foo foo foo</span></p>
>
> But I'm having trouble accomplishing (2) as this is my first exposure
> to XPath / Nokogiri -- but I believe the tools are in my hands to get
> this done ... and I think this "might work" (TM).
If you're more comfortable with CSS, you should use CSS. Here is a
CSS query that says "find all span tags that have an attribute named
style whose value is 'mso-spacerun: yes'":
doc.css('span[style = "mso-spacerun: yes"]').each do |span|
span['style'] = 'poetry-line-odd'
end
Here is the same query using XPath:
doc.xpath('//span[@style = "mso-spacerun: yes"]').each do |node|
span['style'] = 'poetry-line-odd'
end
Hope that helps!
--
Aaron Patterson
http://tenderlovemaking.com/
Mike Dalessio
http://github.com/flavorjones/loofah
If Loofah doesn't do exactly what you want, please let me know, I'm happy to improve it!
--
mike dalessio
mi...@csa.net