Tab-separated Value Reader

line_reader.v1.rb

module TSV
  class LineReader
    attr_reader :filepath
    def initialize(filepath)
      @filepath = filepath
    end

    def each
      buffer = ""
      open(filepath) do |f|
        while chunk = f.read(4096)
          buffer << chunk
          lines = buffer.split(/[\r\n]+/)
          buffer = lines.pop
          lines.each do |line|
            yield line
          end
        end
        yield buffer unless buffer.empty?
      end
    end
  end
end

line_reader.v2.rb

module TSV
  class LineReader
    attr_reader :filepath
    def initialize(filepath)
      @filepath = filepath
    end

    def each
      buffer = ""
      open(filepath) do |f|
        chunk = ""
        while chunk = f.read(4096, chunk)
          buffer << chunk
          lines = buffer.split(/[\r\n]+/)
          buffer = lines.pop
          lines.each do |line|
            yield line
          end
        end
        yield buffer unless buffer.empty?
      end
    end
  end
end

strict_tsv.rb

class StrictTsv
  include Enumerable

  attr_reader :filepath
  def initialize(filepath)
    @filepath = filepath
  end

  def each
    headers = nil
    row_enumerable.each do |line|
      unless headers
        headers = line.strip.split "\t"
        next
      end
      next if line.empty?
      fields = Hash[headers.zip(line.split "\t")]
      yield fields
    end
  end

  private
  def row_enumerable
    TSV::LineReader.new(filepath)
  end
end