ansaurus

Question

Strip text from HTML document using Ruby

Answer 1

+3 A:

You can scan the string to create an array of "tokens", and then only select those that are html tags:

>> some_html
=> "<div>foo bar</div><p>I like <em>this</em> stuff <a href='http://foo.bar'&gt; long time</a></p>"
>> some_html.scan(/<\/?[^>]+>|[\w\|`~!@#\$%^&*\(\)\-_\+=\[\]{}:;'",\.\/?]+|\s+/).select { |t| t =~ /<\/?[^>]+>/ }.join("")
=> "<div></div><p><em></em><a href='http://foo.bar'&gt;&lt;/a&gt;&lt;/p&gt;"

==Edit==

Or even better, just scan for html tags ;)

>> some_html.scan(/<\/?[^>]+>/).join("")
=> "<div></div><p><em></em><a href='http://foo.bar'&gt;&lt;/a&gt;&lt;/p&gt;"

hgimenez 2009-09-30 12:13:35

Answer 2

A:

To grab everything not in a tag, you can use nokogiri like this:

doc.search('//text()').text

Of course, that will grab stuff like the contents of <script> or <style> tags, so you could also remove blacklisted tags:

blacklist = ['title', 'script', 'style']
nodelist = doc.search('//text()')
blacklist.each do |tag|
  nodelist -= doc.search('//' + tag + '/text()')
end
nodelist.text

You could also whitelist if you preferred, but that's probably going to be more time-intensive:

whitelist = ['p', 'span', 'strong', 'i', 'b']  #The list goes on and on...
nodelist = Nokogiri::XML::NodeSet.new(doc)
whitelist.each do |tag|
  nodelist += doc.search('//' + tag + '/text()')
end
nodelist.text

You could also just build a huge XPath expression and do one search. I honestly don't know which way is faster, or if there is even an appreciable difference.

Pesto 2009-09-30 14:03:59

Answer 3

+12 A:

This works too:

doc = Nokogiri::HTML(your_html)
doc.xpath("//text()").remove

andre-r 2009-09-30 14:29:14

Nice! Thanks for making the rest of us look stupid :-)

Jörg W Mittag 2009-09-30 19:28:01

Oh wow - I'm going to check this out.

davidsmalley 2009-10-02 10:34:01

Answer 4

A:

I just came up with this, but @andre-r's solution is soo much better!

#!/usr/bin/env ruby

require 'nokogiri'

def strip_text doc
  Nokogiri(doc).tap { |doc|
    doc.traverse do |node|
      node.content = nil if node.text?
    end
  }.to_s
end

require 'test/unit'
require 'yaml'
class TestHTMLStripping < Test::Unit::TestCase
  def test_that_all_text_gets_strippped_from_the_document
    dirty, clean = YAML.load DATA
    assert_equal clean, strip_text(dirty)
  end
end
__END__
---
- |
  <!DOCTYPE html>
  <html xmlns='http://www.w3.org/1999/xhtml' xml:lang='en' lang='en'>
    <head>
        <meta http-equiv='Content-type'     content='text/html; charset=UTF-8' />
        <title>Test HTML Document</title>
        <meta http-equiv='content-language' content='en' />
    </head>
    <body>
        <h1>Test <abbr title='Hypertext Markup Language'>HTML</abbr> Document</h1>
        <div class='main'>
            <p>
                <strong>Test</strong> <abbr title='Hypertext Markup Language'>HTML</abbr> <em>Document</em>
            </p>
        </div>
    </body>
  </html>
- |
  <!DOCTYPE html>
  <html xmlns="http://www.w3.org/1999/xhtml" xml:lang="en" lang="en">
  <head>
  <meta http-equiv="Content-Type" content="text/html; charset=UTF-8">
  <title></title>
  <meta http-equiv="content-language" content="en">
  </head>
  <body><h1><abbr title="Hypertext Markup Language"></abbr></h1><div class="main"><p><strong></strong><abbr title="Hypertext Markup Language"></abbr><em></em></p></div></body>
  </html>

Jörg W Mittag 2009-09-30 14:55:52

ansaurus

tags:

views:

answers:

Strip text from HTML document using Ruby

related questions