Purpose: A Gnotime Activity Journal HTML parse to .csv shell script; I use it for copying/pasting to a timesheet template to save some typing. Editor used: vim/gvim

parse_act_journal.rb

require 'rubygems'
require 'nokogiri'
require 'fastercsv'
require 'htmlentities'

#puts $ARGV[0];

doc = Nokogiri::HTML( File.read($ARGV[0]) )
memo = []            #memo array
elapsed = []         #elapsed array
dates = []           #dates array
index = 0
this_date = ""

start_date = $ARGV[1]

# convert "01:30:00" to 1.5, "1:45:00" to 1.75, etc... 
# Credits/Ref: http://www.intura.co.uk/?p=23
def elapsed_to_decimal(elapsed)
  return 0 if elapsed.empty?
  left = elapsed[0,elapsed.index(":")].to_i       #hh
  middle = elapsed[elapsed.index(":")+1,2].to_i   #mm
  right = elapsed[elapsed.rindex(":")+1,2].to_i   #seconds field not really used (ignored)... 
  return left + (((middle * 5) / 3) / 100.to_f)
end

same_day_elapsed_time = false

doc.css('td').each do |node|
    if node['class'] =~ /gnotime-start-stop-table-start$/
    this_date = node.content()
    if this_date =~ /2011-/ 
      dates.push( this_date )
      #puts "pushed " + this_date
      same_day_elapsed_time = false
    elsif this_date == "\302\240"   #irb didn't show   but "\302\240"
      #puts "detects nbsp"
      same_day_elapsed_time = true
    end
  end
  if node['class'] =~ /(.*)-elapsed$/
    this_date = dates.last()
    if( same_day_elapsed_time == false )
      elapsed.push( elapsed_to_decimal( node.content() ) )
    else
      #puts "same_day_elapsed_time is #{same_day_elapsed_time}"
      elapsed.push( elapsed.pop() + elapsed_to_decimal( node.content() ) )
    end
  end
  #index += 1
end
#puts "dates are " + dates.join(',')
doc.css('div').each do |node|   #|div|
  #if div.attributes["class"].value =~ /(.*)-memo$/
  if node['class'] =~ /(.*)-memo$/
    this_memo_node_text = node.children.children.children.to_s()
    coder = HTMLEntities.new
    memo.push( coder.decode(this_memo_node_text) ) ; index += 1

    #check child dates for diff_day_same_memo condition. and then dup memo
    related_tds = node.next_element().next_element().css('td')
    td_index = 0
    #ad-hoc range limit due to my curernt Gnotime cross-day same Entry sit.
    while ( td_index < 16 && (a_memo_date_xml_text = related_tds[td_index]) )
      if a_memo_date_xml_text.content =~ /2011-/
        if td_index >= 5
          #puts "a different date: #{a_memo_date_xml_text.content}"
          #push same memo
          memo.push( coder.decode(this_memo_node_text) ); index += 1  
        end
      end
      td_index += 5
    end

  end
  
end

#keys = (memo.keys + elapsed.keys).uniq

#add Array.sum ability to Array class for summing up hours 
#class Array; def sum; inject( nil ) { |sum,x| sum ? sum+x : x }; end; end

FasterCSV.open("timesheet.csv", "w") do |csv|
  #csv << ["Date", "Hrs", "Description"]
  for memo_entry_index in 0...index
    reverse_index = index - memo_entry_index
    if( dates[reverse_index] && dates[reverse_index] >= start_date )
      csv << [ dates[reverse_index], elapsed[reverse_index], memo[reverse_index] ]
      csv << [ nil,nil,nil ]  # a blank row for better spacing
    end
  end
  #keys.each do |key|
  #  csv << [ key, elapsed[key], memo[key] ]
  #end
end