Quick and dirty ugly script to get YouTube History playlist into a JSON. You need to extract the history page as an HTM file with Chrome and rename it 'history.htm' in same folder of script. Note that this is for French version of YouTube. Tweak it to fill your need, needed to have a quick result.

get_youtube_history.rb

#!/usr/bin/env ruby

require 'json'
require 'nokogiri'

html = File.read('history.htm')
page = Nokogiri::HTML(html)
uniq_results = []
results = []

page.css('#video-title').each do |video|
  element = Nokogiri::HTML(video.to_s).at_css('a')
  label = element['aria-label'].to_s.reverse!.match(/seuv(?<upload_date>.*?)a y li(?<author>.*?)ed(?<title>.*)/)
  if label.nil?
    label = element['aria-label'].to_s.reverse!.match(/(?<upload_date>.*?)a y li(?<author>.*?)ed(?<title>.*)/)
    upload_date = label['upload_date'].to_s.strip.reverse!
    views = 0
  else
    views = element['aria-label'].to_s.split('il y a')[1].to_s.split('vues')[0].split(' ').reverse![0].to_s.strip.gsub(/[[:space:]]/, '')
    upload_date = label['upload_date'].to_s.strip.reverse!.split(' ')[0...-1].join(' ')
  end
  link = element['href'].to_s.split('&t=')[0]
  title = label['title'].to_s.strip.reverse!
  author = label['author'].to_s.strip.reverse!
  data = {
    'link' => link,
    'title' => title,
    'author' => author,
    'upload_date' => upload_date,
    'views' => views
  }
  @collection = data
  results.push(@collection)
end

results.each_with_index do |result, index|
  occurence = results.count(result)
  result["number_of_personnal_views"] = occurence
  uniq_results.push(result)
end

results = uniq_results.uniq
sorted_results = results.sort_by{ |e| e['number_of_personnal_views'].to_i }
puts sorted_results.map { |o| Hash[o.each_pair.to_a] }.to_json