ffprobe output parser for deprecated ffprobe versions without json output.

ffprobe_parser.rb

require 'pathname'
require 'open3'

# ffprobe output parser for deprecated ffprobe versions.
class FfprobeParser

  # Get media properties for given file.
  #
  # @example Get properties for a given file.
  #   FfprobeParser.properties('/srv/file/video.mp4')
  #     => { legth: 4232, width: 576, height: 1024 }
  #
  # @param file [String] media file to get information for
  # @return properties [Hash] parsed properties
  def self.properties(file)
    ffprobe_output = ffprobe(check_filepath(file))
    parse(ffprobe_output)
  end

  # Parse ffprobe output string.
  #
  # @todo
  #   Ugly regex parsing. Implement better string parser or even better switch
  #   to ffprobe json output if possible.
  # @param string [String] to parse
  # @return properties [Hash]
  def self.parse(string)
    # check string encoding
    string = check_giving_encoding(string)
    # let's have a hash to store some information
    hash = {}
    # matching 00:00:00.00 after 'Duration: '
    length = /[Duration]+:\s+(\d+:\d+:\d+\.\d+)/.match(string)[1]
    hash[:length] = convert_length_to_seconds(length)

    # only video files have width and height attributes
    if string =~ /Stream\s#.*\sVideo/
      hash[:video] = true
      # matching 123x456 and split to an array
      match = /,\s(\d+x\d+)/.match(string)[1].split('x')

      hash[:width]  = match.first
      hash[:height] = match.last
    end

    hash
  end

  class << self

    private

    # Convert ffprobe length into seconds.
    #
    # @param string [String] ffproble output length
    # @return length [Fixnum] converted length
    def convert_length_to_seconds(string)
      hours, minutes, seconds = string.split(':')
      # remove microseconds from seconds e.g. 42.23
      seconds = /(\d{2})\.\d{2}/.match(seconds)[1].to_i

      (hours.to_i * 60 * 60)+(minutes.to_i * 60)+seconds
    end

    # Check that we can use string operations on given string and encoding.
    #
    # @param string [String] to check
    # @return string [String] with valid™ encoding format
    def check_giving_encoding(string)
      begin
        string.match(/i like utf-8/)
      rescue
        string.force_encoding('ISO-8859-1')
      end

      string
    end

    # Call ffprobe system command and return output as string.
    #
    # @param file [Pathname] to open
    # @return string [String] ffprobe output
    def ffprobe(file)
      stdin, stdout, stderr = Open3.popen3("ffprobe '#{file}'")
      stderr.read
    end

    # Convert file path to pathname object if the file exist
    #
    # @param file [String] pathname to check and convert to pathname object
    # @return file [Pathname] object
    def check_filepath(file)
      file = Pathname(file)

      unless file.exist?
        STDERR.puts "File #{file} does not exist!"
        exit 1
      end

      file
    end
  end
end