Nokogiri recursively get all children

wmarbut picture wmarbut · Apr 9, 2012 · Viewed 9.5k times · Source

The Problem

I am running some statistics against various URLS. I want to find the top level element with the most concentrated number of children. The method that I would like to follow is to identify all top level elements and then determine what percentage of all the elements on the page belong to it.

Goal

  • Recursively get all children of a given element.

Inputs: a Nokogiri Element

Outputs: an array of Nokogiri Elements OR the count of total number of children

Setup

  • Ruby 1.9.2
  • Nokogiri gem

What I ended up coming up with (this works, but isn't as pretty as my chosen answer below)

getChildCount(elem)
    children = elem.children
    return 0 unless children and children.count > 0
    child_count = children.count
    children.each do |child|
        child_count += getChildCount(child)
    end
    child_count
end

Answer

John Douthat picture John Douthat · Apr 9, 2012

the traverse method yields the current node and all children to a block, recursively.

# if you would like it to be returned as an array, rather than each node being yielded to a block, you can do this
result = []
doc.traverse {|node| result << node }
result

# or, 
require 'enumerator'
result = doc.enum_for(:traverse).map