Extract useful data from HTML and XML with ease!

Nikkou adds additional methods to Nokogiri to make extracting commonly-used data from HTML and XML easier. It lets you transform HTML into structured data very quickly, and it integrates nicely with Mechanize.

Add Nikkou to your Gemfile:

gem 'nikkou'

Here's a summary of the methods Nikkou provides (see "Methods" for details):

parse_text - Parses the node's text as XML and returns it as a Nokogiri::XML::NodeSet

time(options={}) - Intelligently parses the time (relative or absolute) of either the text or a specified attribute; accepts a time_zone option

url(attribute='href') - Converts the href (or other specified attribute) into an absolute URL using the document's URI; <a href="/p/1">Link</a> yields http://mysite.com/p/1

attr_equals(attribute, string) - Finds nodes where the attribute equals the string

attr_includes(attribute, string) - Finds nodes where the attribute includes the string

attr_matches(attribute, pattern) - Finds nodes where the attribute matches the pattern

*drill(methods) - Nil-safe method chaining

find(path) - Same as search but returns the first matched node

text_equals(string) - Finds nodes where the text equals the string

text_includes(string) - Finds nodes where the text includes the string

text_matches(pattern) - Finds nodes where the text matches the pattern

Returns a Time object (in UTC) by automatically parsing the text or specified attribute of the node.

# <a href="/p/1">3 hours ago</a>
doc.search('a').first.time

attribute

The attribute to parse:

# <a href="/p/1" data-published-at="2013-05-22 02:42:34">My link</a>
doc.search('a').first.time(attribute: 'data-published-at')

time_zone

The document's time zone (the time will be converted from that to UTC):

# <a href="/p/1">3 hours ago</a>
doc.search('a').first.time(time_zone: 'America/New_York')

Returns an absolute URL; useful for parsing relative hrefs. The document's uri needs to be set for Nikkou to know what domain to add to relative paths.

# <a href="/p/1">My link</a>
doc.uri = 'http://mysite.com/mypage'
doc.search('a').first.url # "http://mysite.com/p/1"

If Mechanize is being used, the uri doesn't need to be manually set.

attribute

The attribute to parse:

# <a href="/p/1" data-comments-url="/p/1#comments">My Link</a>
doc.uri = 'http://mysite.com/mypage'
doc.search('a').first.url('data-comments-url') # "http://mysite.com/p/1#comments"

Selects nodes where the specified attribute equals the string.

# <div data-type="news">My Text</div>
doc.attr_equals('data-type', 'news').first.text # "My Text"

Selects nodes where the specified attribute includes the string.

# <div data-type="major-news">My Text</div>
doc.attr_includes('data-type', 'news').first.text # "My Text"

Selects nodes with an attribute matching a pattern. The pattern's matches are available in Node#matches.

# <span data-tooltip="3 Comments">My Text</span>
doc.attr_matches('data-tooltip', /(\d+) comments/i).first.text # "My Text"
doc.attr_matches('data-tooltip', /(\d+) comments/i).first.matches # ["3 Comments", "3"]

Nil-safe method chaining. Replaces this:

node = doc.find('.count')
if node
  attribute = node.attr('data-count')
  if attribute
    return attribute.to_i
  end
end

With this:

return doc.drill([:find, '.count'], [:attr, 'data-count'], :to_i)

Same as search, but returns the first matched node. Replaces this:

nodes = node.search('h4')
if nodes
  return nodes.first
end

With this:

return node.find('h4')

Selects nodes where the text includes the string.

# <div data-type="news">My Text</div>
doc.text_includes('Text').first.text # "My Text"

Selects nodes with text matching a pattern. The pattern's matches are available in Node#matches.

# <a href="/p/1">3 Comments</a>
doc.text_matches(/^(\d+) comments$/i).first.attr('href') # "/p/1"
doc.text_matches(/^(\d+) comments$/i).first.matches # ["3 Comments", "3"]

Nikkou is released under the MIT License. Please see the MIT-LICENSE file for details.