tileinfo.rb @@ -11,18 +11,11 @@ end $title = ARGV[0] $number_of_pages = Integer(ARGV[1]) -unless FileTest.exists?( 'Blocks.txt' ) - unless system( "curl", "-s", "ftp://ftp.unicode.org/Public/5.1.0/ucd/Blocks.txt", "-o", "Blocks.txt" ) - STDERR.puts "Couldn't download Blocks.txt." - exit(-1) - end -end - -require 'individual-characters/tileinfo.rb' +require 'tileinfo.rb' $pages_hash = Hash.new -files = Dir['individual-characters/non-blank/U*-*-*-top.png'] +files = Dir['individual-characters/U*-*-*-top.png'] files.each do |png_file| @@ -107,11 +100,11 @@ $pages.each do |p| puts "#{p} (#{name_of_page(p)})" end -$info_hash = YAML.load( open( "individual-characters/non-blank/top-sizes.yaml", "r" ){ |f| f.read } ) +$info_hash = YAML.load( open( "top-sizes.yaml", "r" ){ |f| f.read } ) $info_hash.each_key do |png_file| # puts "key from $info_hash was: #{png_file}" - from_page = png_file.gsub(/\-\d{3}\-\d{8}-top\..*/,'') + from_page = png_file.gsub(/^.*(U[0-9A-Fa-f]+)\-\d{3}\-\d{8}-top\.png$/,'\1') $pages_hash[from_page].push(png_file) end @@ -215,7 +208,7 @@ class ImageToDraw end def draw_to( im, color, leave_header_space ) - im_tile = GD::Image.new_from_png( "individual-characters/non-blank/#{@image_file}" ) + im_tile = GD::Image.new_from_png( @image_file ) y = 0 im_tile.copy( im, x, y, 0, 0, im_tile.width, im_tile.height) im_tile.destroy compose-inline-breaks @@ -24,7 +24,7 @@ data.each do |e| unless system( "convert", "-scale", "50%", - "individual-characters/non-blank/" + name.gsub(".png","-top.png"), + "individual-characters/" + name.gsub(".png","-top.png"), destination_directory + "reduced-" + new_name ) puts "Failed to convert and copy: #{name}" exit(-1) copy-characters-to-webspace @@ -1,2 +1 @@ -Blocks.txt *.png.empty individual-characters/.gitignore @@ -1,6 +1,6 @@ #!/usr/bin/ruby -w -charts_index_page = "http://www.unicode.org/charts/PDF/" +require 'tileinfo.rb' # The directory with the helper files: helpers = "png-find-grid-revised" @@ -10,6 +10,8 @@ unless system("cd #{helpers} && make") exit(-1) end +charts_index_page = "http://www.unicode.org/charts/PDF/" + index_html = `curl -s #{charts_index_page}` unless $?.success? puts "Couldn't fetch the charts index page from: #{charts_index_page}" @@ -82,7 +84,6 @@ end puts "Globbing directory..." png_files = Dir['U*-*.png'] -# png_files = [ "U0000-002.png", "UFB50-004.png" ] some_done_done_from = Hash.new already_output_files = Dir['individual-characters/U*-*-*.png'] @@ -225,12 +226,140 @@ png_files.each do |png_file| end -unless system "cd individual-characters && ./link-non-blanks" - puts "The link-non-blanks script failed" - exit(-1) +def compare_tile_filenames( a, b ) + Integer( a.gsub( /^.*U([0-9A-F]+)\-([0-9A-F]+)\-([0-9A-F]+).*\.png$/, '0x\1\2\3' ) ) <=> + Integer( b.gsub( /^.*U([0-9A-F]+)\-([0-9A-F]+)\-([0-9A-F]+).*\.png$/, '0x\1\2\3' ) ) end -unless system "cd individual-characters/non-blank && ./extract-numbers" - puts "The extract-numbers script failed" - exit(-1) +files = Dir['individual-characters/U*-*-*.png'] + +# If this is a repeat run then we may have some leftover -top and +# -bottom files here: + +files.delete_if { |x| x =~ /top/ } +files.delete_if { |x| x =~ /bottom/ } + +files.sort! { |a,b| compare_tile_filenames( a, b ) } + +# ------------------------------------------------------------------------ + +last_block = -1 + +open( "top-sizes.yaml", "w" ) do |o_top_sizes| + + o_top_sizes.puts "---" + + open( "codepoints.yaml", "w" ) do |o_codepoints| + + files.each do |fname| + + fname =~ /U([0-9A-F]+)-([0-9A-F]+)/ + + block = Integer("0x"+$1) + + if block != last_block + puts "Extracting numbers from block: " + sprintf("0x%06X",block) + last_block = block + end + + block_name = name_of_page(fname.gsub(/^.*(U[0-9A-F]+).*$/,'\1')) + next unless block_name + + info = `#{helpers}/png-size #{fname}` + unless $?.success? + puts "png-size #{fname} failed." + next + end + + width = nil + height = nil + + info.chomp! + if info =~ /(\d+)x(\d+)/ + + width = Integer($1) + height = Integer($2) + + top_part_fname = fname.sub( /.png/, '-top.png' ) + bottom_part_fname = fname.sub( /.png/, '-bottom.png' ) + + guessed_text_size = 75 + + text_starts_at = height - guessed_text_size + + p = IO.popen( "#{helpers}/crop-images #{fname}", "w" ) + + p.puts( "#{top_part_fname} 0 0 #{width} #{text_starts_at}" ) + p.puts( "#{bottom_part_fname} 0 #{text_starts_at} #{width} #{guessed_text_size}" ) + + p.close + unless $?.success? + puts "#{helpers}/crop-images #{fname} failed." + exit(-1) + end + + top_width = nil + top_height = nil + + output = safe_backticks("#{helpers}/empty-image",top_part_fname) + if $?.success? + puts "Empty, skipping: #{fname}" + system("rm",top_part_fname) + system("rm",bottom_part_fname) + next + else + output.chomp! + if output =~ /^(\d+) (\d+)/ + top_width = Integer($1) + top_height = Integer($2) + end + end + + # Check that the bottom part isn't hashed out, and delete + # it if so: + + output = safe_backticks("#{helpers}/bottom-line-proportion",bottom_part_fname) + proportion = Float(output) + if proportion > 0.01 + puts "Probably cross-hatched, skipping: #{fname}" + system("rm",top_part_fname) + system("rm",bottom_part_fname) + next + end + + if top_width and top_height + o_top_sizes.puts("#{fname}: !ruby/object:TileInfo") + o_top_sizes.puts(" filename: #{fname}") + o_top_sizes.puts(" h: #{top_height}") + o_top_sizes.puts(" w: #{top_width}") + o_top_sizes.puts(" block: #{block_name}") + end + + # Now use OCR to try to parse the codepoint out of the bottom + # part: + + result = `pngtopnm #{bottom_part_fname} | ocrad -` + unless $?.success? + puts "png-topnm #{bottom_part_fname} | ocrad - failed." + exit(-1) + end + + result.chomp! + result.gsub!( /\s/, '' ) + result.gsub!( /[oO]/, '0' ) + result.gsub!( /a/, '8' ) + result.gsub!( /g/, '9' ) + + result.upcase! + + o_codepoints.puts "-" + o_codepoints.puts ' - "' + fname + '"' + o_codepoints.puts " - " + "0x" + result + + end + + end + + end + end make-unicode-poster individual-characters/link-non-blanks individual-characters/non-blank/.gitignore individual-characters/non-blank/extract-numbers individual-characters/tileinfo.rb bbb639dbb999107136ed9a96e88c32485ee7d990 Mark Longair mark@claret.study http://github.com/mhl/unicode-poster/commit/780d59492769f116eb69c48c3bfafa401f929144 780d59492769f116eb69c48c3bfafa401f929144 2008-07-18T03:34:40-07:00 2008-07-18T03:34:40-07:00 Remove link-non-blanks and non-blank/, tidy scripts and move into top-level directory 59c41f4c0dd172bf81f54d994b2fad5ccecfc9f7 Mark Longair mark@claret.study