crawdad 0.0.1

data/lib/crawdad/prawn_tokenizer.rb

@@ -0,0 +1,139 @@
+# encoding: utf-8
+# Crawdad: Knuth-Plass linebreaking in Ruby.
+#
+# Copyright February 2010, Brad Ediger. All Rights Reserved.
+#
+# This is free software. Please see the LICENSE and COPYING files for details.
+
+require 'strscan'
+require 'enumerator'
+
+module Crawdad
+
+  # Ambassador to Prawn. Turns a paragraph into wrappable items.
+  #
+  class PrawnTokenizer
+
+    include Tokens
+
+    # Sets up a tokenizer for the given document (+pdf+).
+    #
+    def initialize(pdf)
+      @pdf = pdf
+    end
+
+    # Tokenize the given paragraph of text into a stream of items (boxes, glue,
+    # and penalties).
+    #
+    # +options+:
+    #
+    # +hyphenation+::
+    #   If provided, allow the given text to be hyphenated as needed to best
+    #   fit the available space. Requires the text-hyphen gem. Allowable values:
+    #   an ISO 639 language code (like 'pt'), or +true+ (synonym for 'en_us').
+    # +indent+::
+    #   If specified, indent the first line of the paragraph by the given
+    #   number of PDF points.
+    #
+    def paragraph(text, options={})
+      hyphenator = if options[:hyphenation]
+        begin
+          gem 'text-hyphen'
+          require 'text/hyphen'
+        rescue LoadError
+          raise LoadError, ":hyphenation option requires the text-hyphen gem"
+        end
+
+        language = ((lang = options[:hyphenation]) == true) ? 'en_us' : lang
+        @hyphenators ||= {}
+        @hyphenators[language] ||= Text::Hyphen.new(:language => language)
+      end
+
+      stream = []
+
+      if w = options[:indent]
+        stream << box(w, "")
+      end
+
+      # Interword glue can stretch by half and shrink by a third.
+      # TODO: optimal stretch/shrink ratios
+      space_width = @pdf.width_of(" ")
+      interword_glue = glue(space_width, 
+                            space_width / 2.0,
+                            space_width / 3.0)
+
+      # TODO: optimal values for sentence space w/y/z
+      sentence_space_width = space_width * 1.5
+      sentence_glue = glue(sentence_space_width,
+                           sentence_space_width / 2.0,
+                           sentence_space_width / 3.0)
+
+      # Break paragraph on whitespace.
+      # TODO: how should "battle-\nfield" be tokenized?
+      text.strip.split(/\s+/).each do |word|
+        w = StringScanner.new(word)
+
+        # For hyphenated words, follow each hyphen by a zero-width flagged
+        # penalty.
+        # TODO: recognize dashes in all their variants
+        while seg = w.scan(/[^-]+-/) # "night-time" --> "<<night->>time"
+          stream.concat add_word_segment(seg, hyphenator)
+        end
+
+        stream.concat(add_word_segment(w.rest, hyphenator))
+
+        # TODO: add ties (~) or some other way to denote a period that
+        # doesn't end a sentence.
+        if w.rest =~ /\.$/
+          stream << sentence_glue
+        else
+          stream << interword_glue
+        end
+      end
+
+      # Remove extra glue at the end.
+      stream.pop if token_type(stream.last) == :glue
+
+      # Finish paragraph with a penalty inhibiting a break, finishing glue (to
+      # pad out the last line), and a forced break to finish the paragraph.
+      stream << penalty(Infinity)
+      stream << glue(0, Infinity, 0)
+      stream << penalty(-Infinity)
+
+      stream
+    end
+
+    protected
+
+    # Returns a series of tokens representing the given word. Hyphenates using
+    # the given +hyphenator+, if provided. Appends a zero-width flagged penalty
+    # if the given word ends in a hyphen.
+    #
+    def add_word_segment(word, hyphenator)
+      tokens = []
+
+      if hyphenator
+        hyphen_width = @pdf.width_of('-')
+
+        splits = hyphenator.hyphenate(word)
+        # For each hyphenated segment, add the box with an optional penalty.
+        [0, *splits].each_cons(2) do |a, b|
+          seg = word[a...b]
+          tokens << box(@pdf.width_of(seg), seg)
+          tokens << penalty(50, @pdf.width_of('-'), true)
+        end
+
+        last = word[(splits.last || 0)..-1]
+        tokens << box(@pdf.width_of(last), last)
+      else
+        tokens << box(@pdf.width_of(word), word)
+      end
+
+      tokens << penalty(50, 0, true) if word =~ /-$/
+      tokens
+    end
+
+  end
+
+end
+

data/lib/crawdad/tokens.rb

@@ -0,0 +1,48 @@
+
+module Crawdad
+
+  module Tokens
+
+    def token_type(token)
+      token[0]
+    end
+
+    def box(width, content)
+      [:box, width, content]
+    end
+
+    def box_content(token)
+      token[2]
+    end
+    
+    def glue(width, stretch, shrink)
+      [:glue, width, stretch, shrink]
+    end
+
+    def glue_stretch(token)
+      token[2]
+    end
+
+    def glue_shrink(token)
+      token[3]
+    end
+
+    def penalty(penalty, width=0, flagged=false)
+      [:penalty, width, penalty, flagged]
+    end
+
+    def penalty_penalty(token)
+      token[2]
+    end
+
+    def penalty_flagged?(token)
+      token[3]
+    end
+
+    def token_width(token)
+      token[1]
+    end
+
+  end
+
+end

metadata

@@ -0,0 +1,75 @@
+--- !ruby/object:Gem::Specification 
+name: crawdad
+version: !ruby/object:Gem::Version 
+  version: 0.0.1
+platform: ruby
+authors: 
+- Brad Ediger
+autorequire: 
+bindir: bin
+cert_chain: []
+
+date: 2010-03-19 00:00:00 -05:00
+default_executable: 
+dependencies: []
+
+description: Crawdad is an implementation of Knuth-Plass linebreaking (justification) for Ruby.
+email: brad.ediger@madriska.com
+executables: []
+
+extensions: 
+- Rakefile
+extra_rdoc_files: []
+
+files: 
+- lib/crawdad
+- lib/crawdad/prawn_tokenizer.rb
+- lib/crawdad/breakpoint.rb
+- lib/crawdad/native.rb
+- lib/crawdad/ffi.rb
+- lib/crawdad/paragraph.rb
+- lib/crawdad/compatibility.rb
+- lib/crawdad/tokens.rb
+- lib/crawdad/ffi
+- lib/crawdad/ffi/paragraph.rb
+- lib/crawdad/ffi/tokens.rb
+- lib/crawdad/ffi/breakpoint_node.rb
+- lib/crawdad.rb
+- ext/crawdad/breakpoint.h
+- ext/crawdad/tokens.c
+- ext/crawdad/paragraph.h
+- ext/crawdad/Makefile
+- ext/crawdad/tokens.h
+- ext/crawdad/paragraph.c
+- Rakefile
+has_rdoc: true
+homepage: http://github.com/madriska/crawdad
+post_install_message: 
+rdoc_options: 
+- --title
+- Crawdad Documentation
+- -q
+require_paths: 
+- lib
+- ext
+required_ruby_version: !ruby/object:Gem::Requirement 
+  requirements: 
+  - - ">="
+    - !ruby/object:Gem::Version 
+      version: "0"
+  version: 
+required_rubygems_version: !ruby/object:Gem::Requirement 
+  requirements: 
+  - - ">="
+    - !ruby/object:Gem::Version 
+      version: "0"
+  version: 
+requirements: []
+
+rubyforge_project: 
+rubygems_version: 1.3.1
+signing_key: 
+specification_version: 2
+summary: Knuth-Plass linebreaking for Ruby
+test_files: []
+