Fulltext search and highlight - and return a jpeg

pdf-search.rb

require 'poppler'
require 'gdk_pixbuf2'

SCALE = 2

filename = "source.pdf"
doc = Poppler::Document.new(filename)
page = doc.get_page(0)

# render the page to an in-memory buffer
width, height = *page.size
buf = Gdk::Pixbuf.new(Gdk::Pixbuf::COLORSPACE_RGB, true, 8, width*SCALE, height*SCALE)
page.render(0, 0, width*SCALE, height*SCALE, SCALE, 0, buf)

# copy the rendered buffer into an pixmap for further editing
map = Gdk::Pixmap.new(nil, width*SCALE, height*SCALE, 24)
map.draw_pixbuf(nil, buf, 0, 0, 0, 0, -1, -1, Gdk::RGB::DITHER_NONE, 0, 0)

# setup highlight color and blend function
gc  = Gdk::GC.new(map) # graphics context
gc.rgb_fg_color = Gdk::Color.new(65535, 65535, 0)
gc.function = Gdk::GC::AND

# find each match and highlight it. The co-ordinate maths is ugly but
# necesary to convert from PDF co-ords to Pixmap co-ords
page.find_text("the").each do |match|
  matchx = match.x1 * SCALE
  matchy = (height - match.y2) * SCALE
  matchw = (match.x2-match.x1) * SCALE
  matchh = (match.y2-match.y1) * SCALE
  map.draw_rectangle(gc, true, matchx, matchy, matchw, matchh)
end

# save the buffer to a JPG
newbuf = Gdk::Pixbuf.from_drawable(nil, map, 0, 0, width*SCALE, height*SCALE)
newbuf.save("foo.jpg", "jpeg")