CODEWITHSUNDEEP
ruby-on-railsrubyxmlregexnokogiri
ylluminate
ylluminate

Reputation: 12369

Parsing contents of paragraph elements with Nokogiri

I'd like to know the proper way to parse a block of contents with Nokogiri:

I have some documents to parse where they originally contained a format where each main container was a <p>. The main pieces of information within each one are divided up, oddly, with <font> tags.

Effectively a stock sample of <p> contents contains the following and is a typical example (some have a lot more content, some a lot less):

<p>
  <font size="5" face="Arial, Helvetica, sans-serif" color="#00CCAA" class="">
    <font color="#AAFF33" class="">
      October 10, 1990 - Maybe a Title
    </font>- 
    <font size="4" class="">
      Some long text here.         
      <font color="#66CC00" class="">
        <a href="SourceTitle/date.pdf">[Blah Blah, October 27, 1982 p. 2</a>
        ]
      </font>. 
      More content. 
      <font color="#00FF33" class="">[Another Source, 1971, issue 01/4]
      </font>. 
    </font>
    <font size="5" face="Arial, Helvetica, sans-serif" color="#00CCAA" class="">
      <font color="#AAFF33" class=""><font size="4" color="#00CCAA" class="">
        Another fantastic article. 
        <a href="SourceTitle/Date.pdf">[Some Source, October 4, p.6]</a>
      </font>
    </font>
  </font>
</font>
</p>

Essentially the "font size" attribute is what sets each component apart in the article. The main points to extract are the FIRST <font size ="5"... (that is the article date and main title, if a title is given) tags, then the actual content.

Presently I have all paragraph chunks coming out with: doc.xpath('//p').each do |node|

However I am not sure if I should pass it through Nokogiri again to parse out it's contents or if I should just run it all through a regex. Was hoping for a small example of doing this "properly" with, I'm assuming, using an embedded xpath discovery within the initial block that pulls the elements out. I assume that there is a way to pull out the sub components based on the font size demarcation, but I've simply not seen a specific example of this yet.

Upvotes: 0

Views: 1260

Answers (1)

Michael Kohl
Michael Kohl

Reputation: 66837

Does that help you get started?

>> doc.xpath('//p').each do |node|
..     puts node.xpath("font[@size='5']/font").first.content.strip
..   end #=> 0
October 10, 1990 - Maybe a Title

Build similar expressions for the other parts you need and you are done :-)

Upvotes: 2

Related Questions