Need to implement a library loader

[monkstone]

An alternative library loader is required, if we no-longer assume we have a sketchbook (or similar mac/windows) to install libraries. A central location (and vanilla processing compatible path is required).
It is thus proposed that libraries such as video and sound should live in eponymous folders in a libraries folder in the ~/.jruby_art folder. See fb3fe74 for first hack, works on linux with video library.

require 'jruby_art'

#Dir[File.join('library/video', '*.jar')].each do |jar|
#  require_relative jar
#end

class AsciiVideoCapture < Processing::App
  load_library :video
  attr_reader :bright, :char, :cheat_screen, :font, :font_size, :letters, :video
  # All ASCII characters, sorted according to their visual density
  LETTER_STRING = %q{ .`-_':,;^=+/\"|)\\<>)iv%xclrs{*}I?!][1taeo7zjLunT#JCwfy325Fp6mqSghVd4EgXPGZbYkOA&8U$@KHDBWNMR0Q}
  LETTER_ORDER = LETTER_STRING.scan(/./)

  def setup
    size(640, 480)
    init_video
    @font_size = 1.5
    @font = load_font(data_path('UniversLTStd-Light-48.vlw'))
    # for the 256 levels of brightness, distribute the letters across
    # the an array of 256 elements to use for the lookup
    @letters = (0...256).map do |i|
      LETTER_ORDER[map1d(i, (0...256), (0...LETTER_ORDER.length))]
    end
    # current brightness for each point
    @bright = Array.new(video.width * video.height, 128)
  end

  def init_video
    # This the default video input, see the test_capture
    # example if it creates an error
    @video = Java::ProcessingVideo::Capture.new(self, 160, 120)
    # Start capturing the images from the camera
    video.start
    @cheat_screen = false
  end

  def capture_event(c)
    c.read
    background 0
  end

  def draw
    return unless (video.available == true)
    capture_event(video)
    push_matrix
    hgap = width / video.width
    vgap = height / video.height
    scale([hgap, vgap].max * font_size)
    text_font(font, font_size)
    index = 0
    video.load_pixels
    (0...video.height).each do
      # Move down for next line
      translate(0,  1.0 / font_size)
      push_matrix
      (0...video.width).each do
        pixel_color = video.pixels[index]
        # Faster method of calculating r, g, b than red(), green(), blue()
        r = pixel_color >> 16 & 0xff
        g = pixel_color >> 8 & 0xff
        b = pixel_color & 0xff
        # Another option would be to properly calculate brightness as luminance:
        # luminance = 0.3*red + 0.59*green + 0.11*blue
        # Or you could instead red + green + blue, and make the the values[] array
        #256*3 elements long instead of just 256.
        pixel_bright = [r, g, b].max
        # The 0.1 value is used to damp the changes so that letters flicker less
        diff = pixel_bright - bright[index]
        bright[index] += diff * 0.1
        fill(pixel_color)
        text(letters[bright[index]], 0, 0)
        # Move to the next pixel
        index += 1
        # Move over for next character
        translate(1.0 / font_size, 0)
      end
      pop_matrix
    end
    pop_matrix
    # image(video, 0, height - video.height)
    # set() is faster than image() when drawing untransformed images
    set(0, height - video.height, video) if cheat_screen
  end

  MESSAGE = <<-EOS
  Controls are:
  g to save_frame, f & F to set font size
  c to toggle cheat screen display
  EOS

  #
  # Handle key presses:
  # 'c' toggles the cheat screen that shows the original image in the corner
  # 'g' grabs an image and saves the frame to a tiff image
  # 'f' and 'F' increase and decrease the font size
  #
  def key_pressed
    case key
    when 'g' then save_frame
    when 'c' then @cheat_screen = !cheat_screen
    when 'f' then @font_size *= 1.1
    when 'F' then @font_size *= 0.9
    else
      warn MESSAGE
    end
  end
end

AsciiVideoCapture.new(title: 'Video Capture Sketch')

[monkstone]

Latest commit dbe9044 caters for loading jars from a local library. @todo add ruby library functionality.