pdftailor 0.0.1

data/bin/pdftailor

@@ -0,0 +1,6 @@
+#!/usr/bin/env ruby
+
+here = File.dirname(__FILE__)
+jar_path = File.expand_path(File.join(here, "..", "jars"))
+jars = Dir.open(jar_path).select{ |name| name =~ /\.jar$/ }.map{ |jar| File.join(jar_path, jar) }.join(":")
+puts `java -verbose:classes -cp #{jars} org.documentcloud.pdftailor.PdfTailor #{ARGV.join(' ')}`

data/java/src/META-INF/MANIFEST.MF

@@ -0,0 +1 @@
+Main-Class: org.documentcloud.pdftailor.PdfTailor

data/java/src/org/documentcloud/pdftailor/PdfTailor.java

@@ -0,0 +1,160 @@
+package org.documentcloud.pdftailor;
+
+import java.io.FileOutputStream;
+import java.io.IOException;
+import java.util.regex.Pattern;
+import java.util.Arrays;
+import java.util.List;
+import java.util.Iterator;
+
+import com.itextpdf.text.Document;
+import com.itextpdf.text.DocumentException;
+import com.itextpdf.text.pdf.PdfCopy;
+import com.itextpdf.text.pdf.PdfReader;
+import com.itextpdf.text.pdf.PdfWriter;
+
+import com.beust.jcommander.Parameter;
+import com.beust.jcommander.JCommander;
+import com.beust.jcommander.MissingCommandException;
+
+public class PdfTailor {
+  
+  public static void main( String[] args ) throws IOException, DocumentException {
+
+    // Initialize the JCommander parsers
+    JCommander cli = new JCommander();
+    StitchCommand   stitch   = new StitchCommand();
+    UnstitchCommand unstitch = new UnstitchCommand();
+
+    // Add our Stitch and Unstitch commands to the parser
+    cli.addCommand("stitch",   stitch);
+    cli.addCommand("unstitch", unstitch);
+
+    // When called with no arguments display the usage information.
+    if (args.length == 0) { 
+      usage(); 
+    } else {
+      try { 
+        // Parse the provided arguments.
+        cli.parse(args);
+        String command = cli.getParsedCommand();
+        if      (command.equals("stitch"))   { stitch(stitch); } 
+        else if (command.equals("unstitch")) { unstitch(unstitch); }
+        else                                 { usage(); }
+        
+      } catch (MissingCommandException unrecognizedCommand) {
+        // and if unrecognized, display the usage information.
+        usage();
+      }
+    }
+  }
+
+  public static String VERSION       = "0.0.1";
+  public static String USAGE_MESSAGE = "pdftailor stitches and unstitches pdfs.\n\n"                        +
+                                       "Version: " + VERSION + "\n\n"                                       +
+                                       "Usage:\n"                                                           +
+                                       "  pdftailor COMMAND [OPTIONS] <pdf(s)>\n"                           +
+                                       "  Main commands:\n"                                                 +
+                                       "    stitch, unstitch\n\n"                                           +
+                                       "Options:\n"                                                         +
+                                       "  -o, --output\n"                                                   +
+                                       "    The file name or file pattern to which output is written.\n"    +
+                                       "    For commands like unstitch which will write multiple files\n"   +
+                                       "    a pattern including \"%d\" can be used to specify a template\n" +
+                                       "    for where files should be written (e.g. ./foo/bar_%d.pdf).\n\n" +
+                                       "Example:\n"                                                         +
+                                       "  pdftailor stitch --output merged.pdf a.pdf b.pdf\n"               +
+                                       "  pdftailor unstitch --output merged_page_%d.pdf merged.pdf\n";
+
+  public static void usage() {
+   System.out.println(USAGE_MESSAGE);
+  }
+
+  // Stitch together an ordered list of pdfs into a single pdf.
+  public static void stitch( StitchCommand cli ) throws IOException, DocumentException {
+    String outputName = cli.output;
+
+    Document document = new Document();
+    PdfCopy writer = new PdfCopy(document, new FileOutputStream(outputName));
+    document.open();
+
+    // Loop over all of the pdfs specified on the commandline
+    // read out their contents and concatenate them
+    // in the order specified
+    PdfReader reader;
+    Iterator<String> pdfPaths = cli.files.iterator();
+    while (pdfPaths.hasNext()) {
+      String path = pdfPaths.next();
+      reader = new PdfReader(path);
+      for ( int pageNumber = 0; pageNumber < reader.getNumberOfPages(); ) {
+        writer.addPage(writer.getImportedPage( reader, ++pageNumber ));
+      }
+      // make sure to free the reader to prevent memory leak
+      // especially for very large PDFs.
+      writer.freeReader(reader);
+    }
+    document.close();
+  }
+
+    // The stitch command parser.
+    static class StitchCommand {
+      @Parameter(description = "The list of files to stitch together.")
+      private List<String> files;
+    
+      @Parameter(names = {"--output", "-o"}, description = "The filename to write to.", required = true)
+      private String output;
+    }
+
+  // unstitch a pdf into its constitutent pages.
+  public static void unstitch( UnstitchCommand cli ) throws IOException, DocumentException {
+    // use JCommander's default file list to get the file to split.
+    String readerPath = cli.files.get(0);
+    PdfReader reader = new PdfReader(readerPath);
+    
+    // Loop over the document's pages by page number.
+    for ( int pageNumber = 0; pageNumber < reader.getNumberOfPages(); ) {
+      pageNumber++;
+
+      Document document = new Document();
+      String outputName;
+      if (cli.output != null && cli.output.length() > 0) { outputName = cli.output; } else { outputName = readerPath; }
+      PdfCopy writer = new PdfCopy(document, new FileOutputStream(outputPath(outputName, pageNumber)));
+
+      document.open();
+      writer.addPage(writer.getImportedPage(reader, pageNumber));
+      document.close();
+      writer.close();
+    }
+  }
+
+    // The unstitch command parser.
+    static class UnstitchCommand {
+      @Parameter(description = "The file to unstitch.")
+      private List<String> files;
+
+      @Parameter(names = {"--output", "-o"}, description = "The filename pattern to write to. (e.g. mydir/file_%d_name.pdf)")
+      private String output;
+    }
+
+  // Return a file path for page when provided with a page number and
+  // either an existing file path, or a path template.
+  protected static String outputPath( String outputName, int pageNumber ) {
+    Pattern templatePattern = Pattern.compile("%d");
+    Pattern filePattern = Pattern.compile("^(.+)\\.pdf$");
+    String path;
+
+    // if outputName contains %d, we assume it's a template.
+    if ( templatePattern.matcher(outputName).find() ){
+      // replace %d with the page number.
+      path = String.format(outputName, pageNumber);
+    } else if ( filePattern.matcher(outputName).find() ) {
+    // if outputName is a pdf, chop off the file ending and insert the page number before reattaching.
+      path = filePattern.matcher(outputName).replaceFirst("$1_" + pageNumber + ".pdf");
+    } else {
+      // otherwise we'll just append the page number and add the pdf file extension.
+      path = outputName + "_" + pageNumber + ".pdf";
+    }
+    return path;
+  }
+
+}

data/lib/pdftailor.rb

@@ -0,0 +1,7 @@
+module PDFTailor
+  def self.stitch(pdfs, options)
+  end
+  
+  def self.unstitch(pdf, options)
+  end
+end

data/pdftailor.gemspec

@@ -0,0 +1,20 @@
+Gem::Specification.new do |s|
+  s.name      = 'pdftailor'
+  s.version   = '0.0.1'
+  s.date      = '2013-09-09'
+
+  s.summary     = "Stitching and unstitching for PDFs"
+  s.description = <<-EOS
+    Stitching and unstitching for PDFs.  A java library delivered via ruby out of convenience.
+  EOS
+
+  s.authors           = ['Ted Han']
+  s.email             = 'opensource@documentcloud.org'
+  s.homepage          = 'http://documentcloud.github.io/pdftailor'
+
+  s.require_paths     = ['lib']
+  s.executables       = ['pdftailor']
+
+  s.files = Dir['jars/*', 'lib/**/*', 'bin/*', 'java/**/*', 'pdftailor.gemspec']
+  s.license = "MIT"
+end

metadata

@@ -0,0 +1,58 @@
+--- !ruby/object:Gem::Specification
+name: pdftailor
+version: !ruby/object:Gem::Version
+  prerelease:
+  version: 0.0.1
+platform: ruby
+authors:
+- Ted Han
+autorequire:
+bindir: bin
+cert_chain: []
+date: 2013-09-09 00:00:00.000000000 Z
+dependencies: []
+description: |2
+      Stitching and unstitching for PDFs.  A java library delivered via ruby out of convenience.
+email: opensource@documentcloud.org
+executables:
+- pdftailor
+extensions: []
+extra_rdoc_files: []
+files:
+- jars/itextpdf-5.3.4.jar
+- jars/jcommander-1.31-SNAPSHOT.jar
+- jars/pdftailor.jar
+- lib/pdftailor.rb
+- bin/pdftailor
+- java/src/META-INF/MANIFEST.MF
+- java/src/org/documentcloud/pdftailor/PdfTailor$StitchCommand.class
+- java/src/org/documentcloud/pdftailor/PdfTailor$UnstitchCommand.class
+- java/src/org/documentcloud/pdftailor/PdfTailor.class
+- java/src/org/documentcloud/pdftailor/PdfTailor.java
+- pdftailor.gemspec
+homepage: http://documentcloud.github.io/pdftailor
+licenses:
+- MIT
+post_install_message:
+rdoc_options: []
+require_paths:
+- lib
+required_ruby_version: !ruby/object:Gem::Requirement
+  requirements:
+  - - '>='
+    - !ruby/object:Gem::Version
+      version: '0'
+  none: false
+required_rubygems_version: !ruby/object:Gem::Requirement
+  requirements:
+  - - '>='
+    - !ruby/object:Gem::Version
+      version: '0'
+  none: false
+requirements: []
+rubyforge_project:
+rubygems_version: 1.8.24
+signing_key:
+specification_version: 3
+summary: Stitching and unstitching for PDFs
+test_files: []