pst 0.0.1

data/.gitignore

@@ -0,0 +1,9 @@
+*.gem
+.bundle
+Gemfile.lock
+pkg/*
+.#*
+\#*#
+vendor/jars/*.jar
+test/data/*.pst
+.DS_Store

data/Gemfile

@@ -0,0 +1,6 @@
+source "http://rubygems.org"
+
+# Specify your gem's dependencies in pst.gemspec
+gemspec
+
+gem 'rspec'

data/README.mkd

@@ -0,0 +1,44 @@
+# pst.rb
+
+## Description
+
+JRuby gem for working with Outlook PST files. Syntactic sugar over
+`java-libpst`
+
+## Example
+
+    pstfile = Pff::PSTFile.new("foo.pst")
+
+    # iterate over all (nested) folders
+    folders = pstfile.root.sub_folders.inject({}){|acc,f|
+      acc[f.name] = f
+      acc
+    }
+
+    folder = folders["Inbox"] 
+    email = folder.children.first
+
+    pp email.subject        # -> "Re: obama's new debt plan"
+    pp email.num_recipients # -> 2
+
+    # handy human ids and hash strings
+    pp email.human_id "nates-mails:/path/to/foo.pst:/Top of Personal Folders/Inbox:Fri Aug 03 01:02:00 PDT 2011:<foo@bar.com>:Re: obama's new debt plan")
+    pp email.hash_string    # -> "f161dd2a45952784c440bd5879684ae89b8b0523"
+
+    recipient = email.recipients.first
+    pp recipient.name       # -> "Nate Murray"
+    pp recipient.email      # -> "nate@natemurray.com
+
+## References 
+
+The heavy lifting is done by Richard Johnson's java-libpst. Bless
+him for working out the protocol details so I didn't have to.
+
+* http://code.google.com/p/java-libpst/
+* http://www.rjohnson.id.au/wordpress/2010/01/26/java-libpst-pst2gmail/
+
+# Author
+
+Nate Murray <nate@natemurray.com>
+http://www.xcombinator.com
+

data/Rakefile

@@ -0,0 +1 @@
+require 'bundler/gem_tasks'

data/lib/pst.rb

@@ -0,0 +1,22 @@
+here = File.expand_path(File.dirname(__FILE__))
+$LOAD_PATH << "#{here}"
+
+require 'rubygems'
+require 'pp'
+require 'java'
+require 'andand'
+
+module Pst; end
+
+jars_dir = File.dirname(__FILE__) + "/../vendor/jars"
+$LOAD_PATH << jars_dir
+
+Dir.entries(jars_dir).sort.each do |entry|
+  if entry =~ /.jar$/
+    require entry
+  end
+end
+
+require "pst/extensions"
+require "pst/version"
+require "pst/base"

data/lib/pst/base.rb

@@ -0,0 +1,210 @@
+require 'digest/sha1'
+
+class Java::ComPff::PSTFile
+  attr_accessor :collection
+  attr_accessor :filename
+  alias_method :file, :getFileHandle
+
+  def initialize(name)
+    super(name)
+    @filename = name
+  end
+
+  def name
+    self.getMessageStore.getDisplayName
+  end
+
+  def root
+    f = self.getRootFolder
+    f.file = self
+    f
+  end
+end
+
+class Java::ComPff::PSTFolder
+  attr_accessor :file
+  attr_reader :parent
+  alias_method :subfolder_count, :getSubFolderCount
+  alias_method :email_count, :getContentCount
+
+  def name
+    self.getDisplayName
+  end
+
+  def sub_folders
+    Enumerator.new do |yielder|
+      self.getSubFolders.each do |f|
+        f.parent = self 
+        yielder.yield f
+        f.sub_folders.each do |fc|
+          yielder.yield fc
+        end
+      end
+    end
+  end
+
+  def children
+    # this doesn't work dont use it.  it doesn't work because
+    # Enumerator does some sort of non-deterministic lookaheads
+    # that move the cursor out from underneith the underlying
+    # java-pst library
+    #
+    # Maybe once I understand Enumerator better we can fix this.
+    raise "TODO"
+    Enumerator.new do |yielder|
+      max = self.email_count
+      idx = 0
+      while idx < max 
+        self.moveChildCursorTo(idx)
+        kid = self.getNextChild
+        kid.folder = self
+        yielder.yield kid
+        idx = idx + 1
+      end
+    end
+  end
+
+  def filename 
+    self.file.filename
+  end
+
+  def path
+    levels = [self.name]
+    f = self
+    while p = f.parent
+      levels << p.name
+      f = p
+    end
+    levels.reverse.join("/")
+  end
+
+  def human_id
+    "%s:%s:%s" % [self.file.collection || "no-collection", filename, path]
+  end
+
+  def hash_string
+    Digest::SHA1.hexdigest(human_id)
+  end
+
+  def parent=(the_parent)
+    @parent   = the_parent 
+    self.file = the_parent.file
+  end
+
+  def creation_time
+    t = self.getCreationTime || self.getLastModificationTime
+    t.andand.to_time
+  end
+
+end
+
+class Java::ComPff::PSTMessage 
+  attr_accessor :folder
+
+  alias_method :subject, :getSubject
+  alias_method :display_to, :getDisplayTo
+  alias_method :num_recipients, :getNumberOfRecipients
+  alias_method :num_attachments, :getNumberOfAttachments
+  alias_method :sender_name, :getSenderName
+  alias_method :sender_email, :getSenderEmailAddress
+  alias_method :original_subject, :getOriginalSubject
+  #alias_method :body, :getBody
+  alias_method :html_body, :getBodyHTML
+
+  # things to pay attention to
+  # next.getDescriptorNode().descriptorIdentifier+"";
+  # next.getSentRepresentingName() + " <"+ next.getSentRepresentingEmailAddress() +">";
+  # next.getReceivedByName() + " <"+next.getReceivedByAddress()+">" + 
+  # next.displayTo();
+  # next.getClientSubmitTime();
+
+  def human_id
+    "%s:%s:%s:%s" % [ folder.human_id, self.getClientSubmitTime.to_s, self.getInternetMessageId, self.subject ]
+  end
+
+  def hash_string
+    Digest::SHA1.hexdigest(human_id)
+  end
+
+  def pretty_string
+    "[%s] %s - %s <%s> %s <%s> %s %s a:%s" % [
+          self.getDescriptorNode.descriptorIdentifier,
+          self.getSubject,
+          self.getSentRepresentingName,
+          self.getSentRepresentingEmailAddress,
+          self.getReceivedByName,
+          self.getReceivedByAddress,
+          self.displayTo,
+          self.getClientSubmitTime,
+          self.hasAttachments]
+  end
+
+  def recipients
+    recip = []
+    #Enumerator.new do |yielder|
+      i = 0
+      while i < self.getNumberOfRecipients
+        recipient = self.getRecipient(i)
+        recip << recipient
+        i = i + 1
+      end
+    #end
+    recip
+  end
+
+  def sent_at
+    self.getClientSubmitTime.andand.to_time
+  end
+
+  def contents
+    # this is because [Pff::PSTContact, Pff::PSTTask, Pff::PSTActivity, Pff::PSTRss]
+    # are all PSTMessages but they throw a npe if you call getBody
+    begin
+      return self.getBody
+    rescue
+    end
+    begin
+      return self.toString
+    rescue
+    end
+    raise "no contents found in #{self}"
+  end
+
+  def calculated_recipients_string
+    self.recipients.collect{|r| r.pretty_string}.join(", ")
+  end
+
+  def recipients_string
+    orig = self.getRecipientsString
+    if orig == "No recipients table!"
+      calculated_recipients_string
+    else
+      orig
+    end
+  end
+end
+
+class Java::ComPff::PSTRecipient
+  alias_method :name, :getDisplayName
+  alias_method :email, :getEmailAddress
+  alias_method :smtp, :getSmtpAddress
+
+  def pretty_string
+    "%s <%s>" % [name, email]
+  end
+
+  def human_id
+    pretty_string
+  end
+
+  def hash_string
+    Digest::SHA1.hexdigest(human_id)
+  end
+end
+
+class Java::ComPff::PSTAttachment
+  # todo hash
+  def pretty_string
+    "[%s] %s <%s, %d>" % [self.getContentId, self.getFilename, self.getMimeTag, self.getSize]
+  end
+end

data/lib/pst/extensions.rb

@@ -0,0 +1,9 @@
+class Enumerator 
+  def lazy_select(&block)
+    Enumerator.new do |yielder| 
+      self.each do |val| 
+        yielder.yield(val) if block.call(val) 
+      end
+    end 
+  end
+end

data/lib/pst/version.rb

@@ -0,0 +1,3 @@
+module Pst
+  VERSION = "0.0.1"
+end

data/pom.xml

@@ -0,0 +1,52 @@
+<?xml version="1.0" encoding="UTF-8"?>
+<project>
+  <modelVersion>4.0.0</modelVersion>
+  <groupId>com.xcombinator</groupId>
+  <artifactId>pst</artifactId>
+  <version>0.0.1</version>
+  <name>pst</name>
+  <description></description>
+  <repositories>
+    <repository>
+      <id>clojars</id>
+      <url>http://clojars.org/repo</url>
+    </repository>
+    <repository>
+      <id>central</id>
+      <url>http://repo1.maven.org/maven2</url>
+    </repository>
+  </repositories>
+  <dependencies>
+    <dependency>
+      <groupId>com.pff</groupId>
+      <artifactId>java-libpst</artifactId>
+      <version>1.0.0</version>
+      <!--
+      <exclusions>
+        <exclusion>
+          <groupId>org.slf4j</groupId>
+          <artifactId>org.slf4j/slf4j-api</artifactId>
+        </exclusion>
+      </exclusions> 
+      -->
+    </dependency>
+  </dependencies>
+  <build>
+    <plugins>
+      <plugin> <!-- just run `mvn process-sources` -->
+        <artifactId>maven-dependency-plugin</artifactId>
+        <executions>
+          <execution>
+            <phase>process-sources</phase>
+            <goals>
+              <goal>copy-dependencies</goal>
+            </goals>
+            <configuration>
+              <outputDirectory>${project.basedir}/vendor/jars</outputDirectory>
+            </configuration>
+          </execution>
+        </executions>
+      </plugin>
+    </plugins>
+  </build>
+</project>

data/pst.gemspec

@@ -0,0 +1,21 @@
+# -*- encoding: utf-8 -*-
+$:.push File.expand_path("../lib", __FILE__)
+require "pst/version"
+
+Gem::Specification.new do |s|
+  s.name        = "pst"
+  s.version     = Pst::VERSION
+  s.authors     = ["Nate Murray"]
+  s.email       = ["nate@natemurray.com"]
+  s.homepage    = "http://www.xcombinator.com/"
+  s.summary     = %q{Syntactic sugar over java-libpst}
+  s.description = %q{Syntactic sugar over java-libpst.}
+
+  s.rubyforge_project = "pst.rb"
+  s.add_dependency("andand")
+
+  s.files         = `git ls-files`.split("\n") + `find vendor/jars -type f -name *.jar`.split("\n")
+  s.test_files    = `git ls-files -- {test,spec,features}/*`.split("\n")
+  s.executables   = `git ls-files -- bin/*`.split("\n").map{ |f| File.basename(f) }
+  s.require_paths = ["lib"]
+end

data/spec/pst_spec.rb

@@ -0,0 +1,139 @@
+require 'spec_helper'
+testdatadir = File.dirname(__FILE__) + "/../test/data"
+
+Pff = Java::ComPff
+
+describe "PST" do
+
+  before(:all) do
+    @filename = testdatadir + "/albert_meyers_000.pst"
+    @pstfile = Pff::PSTFile.new(@filename)
+
+    @folders = @pstfile.root.sub_folders.inject({}){|acc,f|
+      acc[f.name] = f
+      acc
+    }
+    
+  end
+
+  context "PSTFile" do 
+
+    it "should have a name" do
+      @pstfile.name.should eql("albert_meyers_000")
+    end
+
+    it "should have a filename" do
+      @pstfile.filename.should eql(@filename)
+    end
+
+    it "should have a root" do
+      @pstfile.root.should_not be_nil
+    end
+
+    it "should tell root about itself" do
+      @pstfile.root.file.should eql(@pstfile)
+      @pstfile.root.file.name.should eql(@pstfile.name)
+    end
+  end
+
+  context "PSTFolder" do
+
+    before(:all) do
+
+    end
+
+    it "should have subfolder count" do
+       @folders["Top of Personal Folders"].subfolder_count.should eql(5)
+    end
+
+    it "should know the number of immediate emails" do
+      #pp @folders["Deleted Items"].email_count
+    end
+
+    it "should have creation time" do
+      # pp @folders["Inbox"].creation_time
+    end
+
+    it "should have sub folders" do
+      @folders.should have_key("ExMerge - Meyers, Albert")
+      @folders.should have_key("meyers-a")
+    end
+
+    it "should have content counts" do
+      @folders["Deleted Items"].getContentCount.should eql(1130)
+      @folders["Inbox"].getContentCount.should eql(22)
+    end
+
+    it "should have a path" do
+      @folders["Inbox"].path.should eql("/Top of Personal Folders/Inbox")
+    end
+
+    it "should have a hash string" do
+      @folders["Inbox"].human_id.should eql("no-collection:/Users/nmurray/projects/enron/software/pst.rb/spec/../test/data/albert_meyers_000.pst:/Top of Personal Folders/Inbox")
+      @folders["Inbox"].hash_string.should eql("767d47f8134cd5c14786efd0274586b1065278e7")
+    end
+
+  end
+
+  context "PSTMessage" do
+    before(:all) do
+      @folder = @folders["Deleted Items"]
+      @email  = @folder.children.first
+    end
+
+    it "should have basic attributes" do
+      @email.subject.should eql("Re: deal 539246.1 REliant HLP dms 7634/7636") 
+      @email.display_to.should eql("Joy Werner")
+    end
+
+    it "should know about its folder" do
+      @email.folder.should eql(@folder)
+    end
+
+    it "should have an id" do
+      @email.human_id.should eql("no-collection:/Users/nmurray/projects/enron/software/pst.rb/spec/../test/data/albert_meyers_000.pst:/Top of Personal Folders/Deleted Items:Fri Apr 06 01:02:00 PDT 2001:<ML1KCRAP2G52RFDSYPSFSAQ0J30PDFMMB@zlsvr22>:Re: deal 539246.1 REliant HLP dms 7634/7636")
+      @email.hash_string.should eql("c512b175785b28532146be7cdb165a5bbee4d130")
+      # pp @email.pretty_string
+    end
+
+    it "should have the number of recipients" do
+      @email.getNumberOfRecipients.should eql(1)
+    end
+
+    it "should iterate over recipients" do
+      @email.recipients.count.should eql(1)
+      #@email.recipients.each do |r|
+      #  pp r
+      #end
+    end
+
+  end
+
+  describe "PSTRecipient" do
+    before(:all) do
+      @folder = @folders["Deleted Items"]
+      @email = @folder.children.take(5).last
+
+      @recipients = @email.recipients.inject({}){|acc,r|
+        acc[r.name] = r
+        acc
+      }
+    end
+
+    it "should have a name" do
+      @recipients.should have_key("Volume Management")
+      @recipients.should have_key("Williams III")
+      @recipients.should have_key("Bill")
+    end
+
+    it "should have an email field" do
+      @recipients["Williams III"].email.should eql("Williams III")
+      @recipients["Bill"].email.should eql("/O=ENRON/OU=NA/CN=RECIPIENTS/CN=Bwillia5")
+    end
+
+    it "should have a hash string" do
+      @recipients["Bill"].hash_string.should eql("f161dd2a45952784c440bd5879684ae89b8b0523")
+    end
+  end
+
+end

data/spec/spec_helper.rb

@@ -0,0 +1,8 @@
+$:.unshift(File.expand_path(File.dirname(__FILE__))+ "/../lib")
+require 'rubygems'
+require 'pst'
+require 'bundler/setup'
+
+RSpec.configure do |config|
+  # some (optional) config here
+end

data/test/data/README

@@ -0,0 +1,4 @@
+# I haven't put the test PST into git because of the filesize. To download try here:
+# contact me if the link breaks <nate@natemurray.com>
+
+curl -0 http://www.xcombinator.com/wp-content/uploads/2011/05/albert_meyers_000.pst > test/data/albert_meyers_000.pst

metadata

@@ -0,0 +1,75 @@
+--- !ruby/object:Gem::Specification
+name: pst
+version: !ruby/object:Gem::Version
+  version: 0.0.1
+  prerelease: 
+platform: ruby
+authors:
+- Nate Murray
+autorequire: 
+bindir: bin
+cert_chain: []
+date: 2011-08-27 00:00:00.000000000Z
+dependencies:
+- !ruby/object:Gem::Dependency
+  name: andand
+  requirement: &2153586440 !ruby/object:Gem::Requirement
+    none: false
+    requirements:
+    - - ! '>='
+      - !ruby/object:Gem::Version
+        version: '0'
+  type: :runtime
+  prerelease: false
+  version_requirements: *2153586440
+description: Syntactic sugar over java-libpst.
+email:
+- nate@natemurray.com
+executables: []
+extensions: []
+extra_rdoc_files: []
+files:
+- .gitignore
+- Gemfile
+- README.mkd
+- Rakefile
+- lib/pst.rb
+- lib/pst/base.rb
+- lib/pst/extensions.rb
+- lib/pst/version.rb
+- pom.xml
+- pst.gemspec
+- spec/pst_spec.rb
+- spec/spec_helper.rb
+- test/data/.gitkeep
+- test/data/README
+- vendor/jars/java-libpst-1.0.0.jar
+homepage: http://www.xcombinator.com/
+licenses: []
+post_install_message: 
+rdoc_options: []
+require_paths:
+- lib
+required_ruby_version: !ruby/object:Gem::Requirement
+  none: false
+  requirements:
+  - - ! '>='
+    - !ruby/object:Gem::Version
+      version: '0'
+required_rubygems_version: !ruby/object:Gem::Requirement
+  none: false
+  requirements:
+  - - ! '>='
+    - !ruby/object:Gem::Version
+      version: '0'
+requirements: []
+rubyforge_project: pst.rb
+rubygems_version: 1.8.6
+signing_key: 
+specification_version: 3
+summary: Syntactic sugar over java-libpst
+test_files:
+- spec/pst_spec.rb
+- spec/spec_helper.rb
+- test/data/.gitkeep
+- test/data/README