module Dstat

Constants

VERSION

Public Instance Methods

flatten_headers(csv) click to toggle source 
# File lib/dstat.rb, line 4
def flatten_headers(csv)
  headers = csv[0]
  csv[0] = []

  good = headers[0]
  headers.each do |header| 
    good = header unless header.nil?
    csv[0] << good
  end

  csv[1] = csv[1].map do |child_header| 
    parent_header = csv[0][csv[1].index(child_header)]
    parent_header != child_header ? "#{parent_header} #{child_header}" : parent_header
  end
  csv[1..-1]
end
import_dstat_csv(file) click to toggle source

Given a dstat csv file, this will import the data, columnwise, into a Hash. dstat CSV files are often of the form:

top_level_header_1,top_level_header_2,,, child_header_1,child_header_2,child_header3,child_header4

or a more concrete example:

epoch,memory usage,,, epoch,used,buff,cach,free 1121245,12456,12856,20059,2999095 1121246,12457,12855,20068,2996885

Where used, buff, cache, and free all correspond to memory usage. An index is created for each (parent header -> child header), unless parent header == child header.

Using the concrete example above, the hash returned from this method would be indexed like so:

hash => [1121245, 1121246] hash => {:used => [12456, 12457], 

            :buff => [12856, 12855], 
            :cach => [20059, 20068], 
            :free => [2999095, 2996885]
}

Written another way:

puts hash[:used]

> “[12456, 12457]” ↑

The hash also has aliases for the keys, so you can access the values multiple ways:

puts hash[:used]

> “[12456, 12457]” ↑

puts hash[“used”]

> “[12456, 12457]” ↑

puts hash

> “[12456, 12457]” ↑

# File lib/dstat.rb, line 84
def import_dstat_csv(file)
  csv = parse_from_file(file)
  csv = normalize_top_level_header(csv)
  csv = csv.transpose
  csv_hash = Hash.new
  csv.each do |column|
    parent_header = column[0].gsub(/\s+/, "_").strip.to_sym
    child_header = column[1].gsub(/\s+/, "_").strip.to_sym
    if (parent_header != child_header)
      csv_hash[parent_header] = Hash.new if csv_hash[parent_header].nil?
      csv_hash[parent_header.to_s] = Hash.new if csv_hash[parent_header.to_s].nil?
      csv_hash[parent_header][child_header] = column[2..-1]
      csv_hash[parent_header.to_s][child_header.to_s] = column[2..-1]
      csv_hash["#{parent_header}_#{child_header}"] = column[2..-1]
    else
      csv_hash[parent_header] = column[2..-1]
      csv_hash[parent_header.to_s] = column[2..-1]
    end
  end
  csv_hash
end
normalize_top_level_header(csv) click to toggle source 
# File lib/dstat.rb, line 21
def normalize_top_level_header(csv)
  headers = csv[0]
  csv[0] = []

  good = headers[0]
  headers.each do |header| 
    good = header unless header.nil?
    csv[0] << good
  end
  csv
end
parse_from_file(file) click to toggle source 
# File lib/dstat.rb, line 33
def parse_from_file(file)
file = File.new(file) unless file.is_a? File
# Split on the double newlines that separate the dstat header from the csv data
file = file.read.split("\n\n")
file = file[1].nil? ? file[0] : file[1]
CSV.parse(file)
end