lucene-stats.rb

require 'rubygems'
require 'ruport'
require 'pp'

path = ARGV[0]
stats = {}

types = {
  'gen'   => 'Segments File: stores information about segments',
  'lock'  => 'Lock File: the Write lock prevents multiple IndexWriters from writing to the same file.',
  'cfs'   => 'Compound File: an optional "virtual" file consisting of all the other index files for systems that frequently run out of file handles.',
  'fnm'   => 'Fields: stores information about the fields',
  'fdx'   => 'Field Index: contains pointers to field data',
  'fdt'   => 'Field Data: the stored fields for documents',
  'tis'   => 'Term Infos: part of the term dictionary, stores term info',
  'tii'   => 'Term Info Index: the index into the Term Infos file',
  'frq'   => 'Frequencies: contains the list of docs which contain each term along with frequency',
  'prx'   => 'Positions: stores position information about where a term occurs in the index',
  'nrm'   => 'Norms: encodes length and boost factors for docs and fields',
	'tvx'   => 'Term Vector Index: stores offset into the document data file',
 	'tvd' 	=> 'Term Vector Documents: contains information about each document that has term vectors',
  'tvf'   => 'Term Vector Fields: the field level info about term vectors',
  'del'   => 'Deleted Documents: info about what files are deleted'
}

Dir.glob(path + "*").each do |file|
  type = file.split('.').last 
  
  if stats[type]
    stats[type][:size] += File.size(file)
  else
    stats[type] = {:size => File.size(file)}
  end
end

total = stats.values.inject(0){|t,v|t+=v[:size]}
stats = stats.collect do |file, stat| 
  percent = format("%.2f", ((stat[:size].to_f / total) * 100)).to_f
  [file, format("%.2f", stat[:size].to_f / (1024*1024)), percent, types[file]]
end.sort{|a, b| b[-2] <=> a[-2]}

report = Ruport::Data::Table.new({
  :column_names => ['filetype', 'size (MB)', '%', 'description'],
  :data => stats 
  })

puts report.to_s