Recursively drilling down the Nokogiri Tree

Question

I want to drill down the tree, and store all the levels:

search_q = Regex.new("Some search regex here")
#something like: page.search('body').first.children.select {|x| x.text[search_q]}.first.children.select {|x| x.text[search_q]}.first......ad infinitum.

I've done a hack:

arbitrarily_long_number = 100
drill = []
(0..arbitrarily_long_number).collect do |n|
  begin
    drill << eval("page.search('body')"+".first.children.select {|x| x.text[search_q]}" * n)
  rescue
    break
  end
end

The problem is that this drills only through the "first" selection. Is there a way to make it drill through every node? I'm thinking of some sort of inject function, but I still haven't wrapped my head around it. Any help would be appreciated.

Output:

pp drill[-4]
puts
pp drill[-3]
puts
pp drill[-2]
#=>[#(Element:0x3fc2324522b4 {
   name = "u",
   children = [
     #(Element:0x3fc232060b60 {
       name = "span",
       attributes = [
         #(Attr:0x3fc2320603e0 {
           name = "style",
           value = "font-size: large;"
           })],
       children = [ #(Text "Ingredients:")]
       })]
   })]

[#(Element:0x3fc232060b60 {
   name = "span",
   attributes = [
     #(Attr:0x3fc2320603e0 { name = "style", value = "font-size: large;" })],
   children = [ #(Text "Ingredients:")]
   })]

[#(Text "Ingredients:")]

Notes: I'm using the mechanize gem, which leverages off of Nokogiri. http://mechanize.rubyforge.org/Mechanize/Page.html#method-i-search http://nokogiri.org/Nokogiri/XML/Node.html#method-i-search

What does "store all the levels" mean? Show us code where you are using Nokogiri and a sample of your HTML or XML. As is your question doesn't contain enough to work with. — the Tin Man, Sep 14 '12 at 16:23
Just wondering if there's a clean way with inject. I can do a loop of loops, but doesn't sound like the ruby way. — Mr. Demetrius Michael, Sep 14 '12 at 16:39
It's hard to tell what you are asking. Ideally you would supply a sample of the HTML, what code you want to run, and what output you would like to see. — Phrogz, Sep 14 '12 at 17:39

score 2 · Answer 1 · answered Sep 15 '12 at 00:25

2

To me it sounds like you want traverse:

doc.traverse do |node|
  drill << node
end

answered Sep 15 '12 at 00:25

pguardiario

53,827
19
119
159

Close, but traverse flattens the node tree – Mr. Demetrius Michael Sep 16 '12 at 04:02

score 1 · Answer 2 · answered Sep 14 '12 at 17:38

Your question is not clear.

If, by

I want to drill down the tree, and store all the levels:

you mean you want to traverse all the nodes, tell Nokogiri to do that.

require 'nokogiri'

doc = Nokogiri::XML(<<EOT)
<a>
  <b>
    <c>1</c>
  </b>
</a>
EOT

doc.search('*').each do |n|
  puts n.name
end

Pasting that into IRB and grabbing the output:

irb(main):011:0* doc.search('*').each do |n|
irb(main):012:1*   puts n.name
irb(main):013:1> end
a
b
c

I used XML, and you're using HTML, but that won't matter. You'll have to change doc to page to fit Mechanize's way, but that is easy.

Recursively drilling down the Nokogiri Tree

2 Answers2