Collects tweets posted by specified user.

TwitterBackup.rb

# -*- coding: utf-8-unix -*-
#
# Tested on :
# - Ruby 1.9.2-p180 (2011-02-18 revision 30909) [x86_64-darwin10.7.0]
#
# Usage : $ ruby TwitterBackup.rb [user name]
#
require 'date'
require 'net/http'
require 'rexml/document'

class TwitterStatuses

  NSTATUS_PER_REQ = 20
  NSTATUS_MAXIMUM = 3200
  SLEEP_SECOND = 30

  attr_reader :statuses_count

  def initialize( screen_name )
    @screen_name = screen_name
    @statuses_count = nil
  end

  def update_statuses_count
    uri = "http://api.twitter.com/1/users/show.xml?screen_name=#{@screen_name}"
    resp = nil
    begin
      resp = Net::HTTP.get_response( URI.parse(uri) )
    rescue
      raise RuntimeError, "Connection error."
    end

    users_show_xmldoc = REXML::Document.new( resp.body )
    @statuses_count = users_show_xmldoc.root.elements["statuses_count"].text.to_i

    @statuses_count = NSTATUS_MAXIMUM if @statuses_count > NSTATUS_MAXIMUM
  end


  def get_user_timeline( page )
    uri = "http://api.twitter.com/1/statuses/user_timeline.xml?screen_name=#{@screen_name}&include_rts=1&include_entities=1&page=#{page}"

    resp = nil
    begin
      resp = Net::HTTP.get_response( URI.parse(uri) )
    rescue
      raise RuntimeError, "Connection error."
    end

    return REXML::Document.new( resp.body )
  end


  def fetch_statuses( pages_fetch = nil )
    backup_filename = "TwitterBackup_#{@screen_name}-[#{Time.now.to_datetime}].xml"

    backup_xmldoc = REXML::Document.new( "<?xml version='1.0' encoding='UTF-8'?>" )
    statuses_aggregate = backup_xmldoc.add_element( 'statuses', {'type' => 'array'} )

    partial_xmldoc = nil

    if ( pages_fetch != nil )
      pages = pages_fetch
    else
      pages = @statuses_count / NSTATUS_PER_REQ
      pages += @statuses_count & NSTATUS_PER_REQ != 0 ? 1 : 0
    end

    begin
      pages.times do |page|
        partial_xmldoc = get_user_timeline( page + 1 )
        status_count =  partial_xmldoc.root.elements.size
        statuses_aggregate.add_element( partial_xmldoc.root )
        print( "Fetched page# #{page+1}/#{pages} (status count=#{status_count}). Waiting #{SLEEP_SECOND} seconds.\n" )
        sleep( SLEEP_SECOND ) if page + 1 != pages
      end
    rescue => e
      p e
    ensure
      backup_xmldoc.write( File.new(backup_filename, 'w') )
    end
  end

end


if __FILE__ == $0
  ts = TwitterStatuses.new( ARGV[0] )
  ts.update_statuses_count
  ts.fetch_statuses
end