Refactor processing, make script standalone with initial env, add command-line argument

ask-wikipedia.rb

#!/usr/bin/env ruby
require 'open-uri'
require 'json'

language = 'en'

article = ARGV[0] ||
  begin
    print 'What do you need to know? : '
    URI::encode gets.chomp
  end

request_url = "http://#{language}.wikipedia.org/w/api.php?action=parse&page=#{article}&format=json&prop=text&section=0"

open(request_url) do |file|
  puts JSON.parse(file.read())['parse']['text'].first[1]
    .gsub(/<\/?[^>]+>/, '') # strip tags
    .gsub(/[[:space:]]+/, ' ') # strip whitespace
    .gsub(/&#[0-9]+;/,'') # strip encoded
    .gsub(/\[[0-9]+\]/,'') # strip referencing
end

I would only want this for the command-line, not to open in Vim. Also Wiki syntax isn't really MD. It could be interesting if you color-formatted it for command-line output, not sure how though.

Strip tags with regexp?? The horror ("the center cannot hold" etc :)

If you did it with Node you could pull in Caja and detag it properly and safely... just saying...

https://www.npmjs.org/package/sanitizerhttps://www.npmjs.org/package/sanitizer

@schmerg The original arose in the Ruby G+ community, but Node ports welcome. I agree DOM munging would work nicely and could be achieved in a Ruby context with Nokogiri.

@mahemoff -- how about port as a ruby gem?