Added follosing formats: rst (reStructuredText), odt (OpenOffice text…

… document).

.gitignore

@@ -1 +1,2 @@
 doc
+*.gem

README

@@ -1,4 +1,3 @@
-Pandoku
-=======
+= Pandoku
 
 Pandoku is loose Ruby interface for Pandoc, the most powerful markup processor.

lib/pandoku.rb

@@ -3,7 +3,7 @@
 
 # %w<markdown rst html latex
 # context man mediawiki texinfo docbook opendocument odt s5 rtf>.each do |f|
-%w<markdown html>.each do |f|
+%w<markdown rst html odt>.each do |f|
   require "#{File.dirname(__FILE__)}/pandoku/formats/#{f}"
 end
 

lib/pandoku/document.rb

@@ -1,7 +1,7 @@
 require File.dirname(__FILE__) + '/format'
 
 module Pandoku
-  PANDOC_PATH = 'pandoc'
+  PANDOC_PATH = ENV['PANDOC_PATH'] || 'pandoc'
 
   class Document
     attr_reader :format, :text
@@ -16,16 +16,18 @@ def initialize(format, text)
 
     # Makes a command string to execute Pandoc.
     def command_for(format)
-      <<-CMD
-        #{PANDOC_PATH} -f #{self.format.class.name}
-                       -t #{format.class.name}
-                       #{self.format.cliopts}
-                       #{format.cliopts}
-      CMD
+      commands = [PANDOC_PATH, '-f', self.format.class.name,
+                              '-t', format.class.name]
+      commands += self.format.cliopts
+      commands += format.cliopts
+      escapeshellarg = lambda do |arg|
+        "'" + arg.to_s.gsub(/[^\\]'/) {|s| %<#{s.chars.first}\\'> } + "'"
+      end
+      commands.select {|v| v }.collect(&escapeshellarg).join(' ')
     end
 
     # Compiles the document to given +format+.
-    # If a second argument +io+ is true, returns +IO+ instead of +String+.
+    # If a second argument +io+ is +true+, returns +IO+ instead of +String+.
     def compile(format, io = false)
       unless format.is_a?(OutputFormat)
         raise TypeError, 'format must be OutputFormat'

lib/pandoku/format.rb

@@ -21,13 +21,9 @@ def initialize(options = {})
     end
 
     def cliopts
-      escapeshellarg = lambda do |arg|
-        "'" + arg.gsub(/[^\\]'/) {|s| %<#{s.chars.first}\\'> } + "'"
-      end
-      @options.select {|k, v| v } \
-              .collect {|p| %<--#{p[0].gsub('_', '-')}> +
-                            (p[1] != true ? %<=#{p[1]}> : '') } \
-              .join(' ')
+      self.options.select {|k, v| v } \
+                  .collect {|p| %<--#{p[0].to_s.gsub('_', '-')}> +
+                                (p[1] != true ? %<=#{p[1]}> : '') }
     end
   end
 
@@ -39,7 +35,7 @@ def parse(text)
 
   module OutputFormat
     # Compiles the given +document+ to the format.
-    # If a second argument +io+ is true, returns +IO+ instead of +String+.
+    # If a second argument +io+ is +true+, returns +IO+ instead of +String+.
     def compile(document, io = false)
       cin, cout, cerr = Open3.popen3(document.command_for(self))
       cin.print(document.text)

lib/pandoku/formats/html.rb

@@ -5,49 +5,53 @@ module Pandoku::Formats
   #
   # == Available Options
   #
-  # * <tt>:standalone</tt> - Produce output with an appropriate header
-  #   and footer.
-  # * <tt>:parse_raw</tt> - Parse untranslatable HTML codes as raw HTML,
-  #   instead of ignoring them.
-  # * <tt>latexmathml</tt> - Use LaTeXMathML to display embedded TeX math in
-  #   HTML output. To insert a link to a local copy of the <b>LaTeXMathML.js</b>
-  #   script, provide a URL string. If just <tt>true</tt> is provided, the
-  #   contents of the script will be inserted directly into the HTML header.
-  # * <tt>:jsmath</tt> - Use jsMath to display embedded TeX math in HTML output.
-  #   The URL should point to the jsMath load script; if provided,
-  #   it will be linked to in the header of standalone HTML documents.
-  # * <tt>:gladtex</tt> - Enclose TeX math in +<eq>+ tags in HTML output.
-  #   These can then be processed by gladTeX to produce links
-  #   to images of the typeset formulas.
-  # * <tt>:mimetex</tt> - Render TeX math using the mimeTeX CGI script.
-  #   If URL is not specified, it is assumed that the script is at
-  #   <b>/cgi-bin/mimetex.cgi</b>.
-  # * <tt>:no_wrap</tt> - Disable text wrapping in output.
-  #   (Default is to wrap text.)
-  # * <tt>:sanitize_html</tt> - Sanitizes HTML using a whitelist.
-  #   Unsafe tags are replaced by HTML comments; unsafe attributes are omitted.
-  #   URIs in links and images are also checked against a whitelist of URI
-  #   schemes.
-  # * <tt>:email_obfuscation</tt> -
-  #   (<tt>nil</tt>|<tt>:javascript</tt>|<tt>:references</tt>)
-  #   Specify a method for obfuscating <tt>mailto:</tt> links in HTML documents.
-  #   * <tt>:nil</tt> leaves <tt>mailto:</tt> links as they are.
-  #   * <tt>:javascript</tt> obfuscates them using javascript.
-  #   * <tt>:references</tt> obfuscates them by printing their letters
-  #     as decimal or hexadecimal character references.
-  #     If <tt>:strict => true</tt> is specified, references is used regardless
-  #     of the presence of this option.
-  # * <tt>:toc</tt> - Include an automatically generated table of contents.
-  # * <tt>:css</tt> -  Link to a CSS style sheet.
-  #   The pathname of the style sheet goes value.
-  # * <tt>:include_in_header</tt> - Include contents of the given filename
-  #   at the end of the header. Implies <tt>:standalone</tt>.
-  # * <tt>:include_before_body</tt> - Include contents of the given filename
-  #   at the beginning of the document body.
-  # * <tt>:include_after_body</tt> - Include contents of the given filename
-  #   at the end of the document body.
-  # * <tt>:custom_header</tt> - Use contents of the given filename as the
-  #   document header. Implies <tt>:standalone</tt>.
+  # <tt>:standalone</tt>:: Produce output with an appropriate header and footer.
+  # <tt>:parse_raw</tt>:: Parse untranslatable HTML codes as raw HTML,
+  #                       instead of ignoring them.
+  # <tt>latexmathml</tt>::
+  #     Use LaTeXMathML to display embedded TeX math in HTML output.
+  #     To insert a link to a local copy of the <b>LaTeXMathML.js</b> script,
+  #     provide a URL string.  If just <tt>true</tt> is provided, the contents
+  #     of the script will be inserted directly into the HTML header.
+  # <tt>:jsmath</tt>::
+  #     Use jsMath to display embedded TeX math in HTML output.
+  #     The URL should point to the jsMath load script; if provided,
+  #     it will be linked to in the header of standalone HTML documents.
+  # <tt>:gladtex</tt>::
+  #     Enclose TeX math in +<eq>+ tags in HTML output.
+  #     These can then be processed by gladTeX to produce links
+  #     to images of the typeset formulas.
+  # <tt>:mimetex</tt>::
+  #     Render TeX math using the mimeTeX CGI script.  If URL is not specified,
+  #     it is assumed that the script is at <b>/cgi-bin/mimetex.cgi</b>.
+  # <tt>:no_wrap</tt>:: Disable text wrapping in output.
+  #                     (Default is to wrap text.)
+  # <tt>:sanitize_html</tt>::
+  #     Sanitizes HTML using a whitelist.  Unsafe tags are replaced by
+  #     HTML comments; unsafe attributes are omitted.  URIs in links and
+  #     images are also checked against a whitelist of URI schemes.
+  # <tt>:email_obfuscation</tt> <i>(<tt>nil</tt>|<tt>:javascript</tt>|<tt>:references</tt>)</i>::
+  #     Specify a method for obfuscating <tt>mailto:</tt> links
+  #     in HTML documents.
+  #     * <tt>:nil</tt> leaves <tt>mailto:</tt> links as they are.
+  #     * <tt>:javascript</tt> obfuscates them using javascript.
+  #     * <tt>:references</tt> obfuscates them by printing their letters
+  #       as decimal or hexadecimal character references.
+  #       If <tt>:strict => true</tt> is specified, references is
+  #       used regardless of the presence of this option.
+  # <tt>:toc</tt>:: Include an automatically generated table of contents.
+  # <tt>:css</tt>:: Link to a CSS style sheet.
+  #                 The pathname of the style sheet goes value.
+  # <tt>:include_in_header</tt>::
+  #     Include contents of the given filename at the end of the header.
+  #     Implies <tt>:standalone</tt>.
+  # <tt>:include_before_body</tt>:: Include contents of the given filename
+  #                                 at the beginning of the document body.
+  # <tt>:include_after_body</tt>:: Include contents of the given filename
+  #                                at the end of the document body.
+  # <tt>:custom_header</tt>::
+  #     Use contents of the given filename as the document header.
+  #     Implies <tt>:standalone</tt>.
   class HTML < Pandoku::Format
     include Pandoku::InputFormat, Pandoku::OutputFormat
 

lib/pandoku/formats/markdown.rb

@@ -5,17 +5,17 @@ module Pandoku::Formats
   #
   # == Available Options
   #
-  # * <tt>:strict</tt> - Use strict syntax, with no extensions or variants.
-  # * <tt>:reference_links</tt> - Use reference-style links,
-  #   rather than inline links, in writing.
-  # * <tt>:smart</tt> - Use smart quotes, dashes, and ellipses.
-  # * <tt>:no_wrap</tt> - Disable text wrapping in output.
-  #   (Default is to wrap text.)
-  # * <tt>:sanitize_html</tt> - Sanitizes HTML using a whitelist.
-  #   Unsafe tags are replaced by HTML comments; unsafe attributes are omitted.
-  #   URIs in links and images are also checked against a whitelist of URI
-  #   schemes.
-  # * <tt>:toc</tt> - Include an automatically generated table of contents.
+  # <tt>:strict</tt>:: Use strict syntax, with no extensions or variants.
+  # <tt>:reference_links</tt>:: Use reference-style links,
+  #                             rather than inline links, in writing.
+  # <tt>:smart</tt>:: Use smart quotes, dashes, and ellipses.
+  # <tt>:no_wrap</tt>:: Disable text wrapping in output.
+  #                     (Default is to wrap text.)
+  # <tt>:sanitize_html</tt>::
+  #     Sanitizes HTML using a whitelist.  Unsafe tags are replaced by
+  #     HTML comments; unsafe attributes are omitted.  URIs in links and
+  #     images are also checked against a whitelist of URI schemes.
+  # <tt>:toc</tt>:: Include an automatically generated table of contents.
   class Markdown < Pandoku::Format
     include Pandoku::InputFormat, Pandoku::OutputFormat
 

lib/pandoku/formats/odt.rb

@@ -1,27 +1,53 @@
+require 'tempfile'
 require File.dirname(File.dirname(__FILE__)) + '/format'
 
 module Pandoku::Formats
   # ODT (OpenOffice text document) format.
-  class ODT
-    include OutputFormat
+  #
+  # == Available Options
+  #
+  # <tt>:toc</tt>:: Include an automatically generated table of contents.
+  class ODT < Pandoku::Format
+    include Pandoku::OutputFormat
 
     def self.name
       :odt
     end
 
+    def self.default_options
+      { :toc => false }
+    end
+
     def cliopts
-      @tmpfile = @tmpfile || self._tmpfile
-      "--output=#{@tmpfile} #{super.cliopts}"
+      @tmpfile = @tmpfile || _tmpfile
+      super + ["--output=#{@tmpfile}"]
     end
 
-    def compile(document)
-      @tmpfile = self._tmpfile
+    # Compiles the given +document+ to the format.
+    # If a second argument +file+ is +true+, returns +File+ instead of +String+.
+    #
+    # *Note* that it creates a temporary file for the result.
+    # If +file+ is +false+, the temporary file become deleted.
+    def compile(document, file = false)
+      @tmpfile = _tmpfile
+      ps = IO.popen(document.command_for(self), 'w')
+      ps.print(document.text)
+      ps.close
+      f = File.new(@tmpfile, 'r')
+      return f if file
+      result = f.read
+      f.close
+      File.unlink(@tmpfile)
+      return result
     end
 
     private
 
     def _tmpfile
-      Dir.tmpdir
+      f = Tempfile.new('pandoku')
+      path = f.path
+      f.close(false)
+      path
     end
   end
 end

pandoku.gemspec

@@ -3,7 +3,7 @@ require 'pathname'
 
 Gem::Specification.new do |s|
   s.name = 'pandoku'
-  s.version = '0.1'
+  s.version = '0.2'
   s.homepage = 'http://github.com/dahlia/pandoku'
   s.summary = s.description = 'Loose Ruby interface for Pandoc, ' \
                               'the most powerful markup processor.'