Add timeout option to JODConverter.

Original work by documentcloud#34 with
modification to not use a default timeout (causing no change from
existing functionality).

Author: Doug Mayer

index.html

@@ -233,6 +233,9 @@ <h2 id="usage">Usage</h2>
       <b>xls</b> and so on, as well as <b>html</b>, <b>odf</b>, <b>rtf</b>, <b>swf</b>, <b>svg</b>, and <b>wpd</b>.
       The first time that you convert a new file type, LibreOffice will lazy-load
       the code that processes it &mdash; subsequent conversions will be much faster.
+      <br />
+      The <tt>--timeout</tt> flag may be used to set the number of seconds before
+      JODConverter's "error on timeout" exception. Useful when you convert very large documents.
     </p>
 <pre>
 docsplit pdf documentation/*.html</pre>

lib/docsplit/command_line.rb

@@ -101,6 +101,9 @@ def parse_options
         opts.on('-r', '--rolling', 'generate images from each previous image') do |r|
           @options[:rolling] = true
         end
+        opts.on('-t', '--timeout', 'Timeout for PDF extraction from OpenOffice-supported document format') do |t|
+          @options[:timeout] = t
+        end
         opts.on_tail('-v', '--version', 'display docsplit version') do
           puts "Docsplit version #{Docsplit::VERSION}"
           exit
@@ -120,4 +123,4 @@ def parse_options
 
   end
 
-end
+end

lib/docsplit/pdf_extractor.rb

@@ -112,6 +112,7 @@ def office_path
     # Convert documents to PDF.
     def extract(docs, opts)
       out = opts[:output] || '.'
+      timeout = opts[:timeout]
       FileUtils.mkdir_p out unless File.exists?(out)
       [docs].flatten.each do |doc|
         ext = File.extname(doc)
@@ -132,6 +133,7 @@ def extract(docs, opts)
             true
           else # open office presumably, rely on JODConverter to figure it out.
             options = "-jar #{ESCAPED_ROOT}/vendor/jodconverter/jodconverter-core-3.0-beta-4.jar -r #{ESCAPED_ROOT}/vendor/conf/document-formats.js"
+            options += "-t #{timeout}" if timeout
             run_jod "#{options} #{escaped_doc} #{escaped_out}/#{escaped_basename}.pdf", [], {}
           end
         end