peregrin 1.1.1

data/test/formats/ochook_test.rb

@@ -0,0 +1,104 @@
+require 'test_helper'
+
+class Peregrin::Tests::OchookTest < Test::Unit::TestCase
+
+  def test_validates
+    # FIXME: tests for:
+    #  - DirectoryNotFound
+    #  - MissingManifest
+    #  - IndexHTMLRootHasNoManifest
+    #
+    # ???
+    assert_nothing_raised {
+      Peregrin::Ochook.validate('test/fixtures/ochooks/basic')
+    }
+  end
+
+
+  def test_read
+    ook = Peregrin::Ochook.read('test/fixtures/ochooks/basic')
+    book = ook.to_book
+    assert_equal(1, book.components.length)
+    assert_equal("index.html", book.components.first.src)
+    assert_equal(['cover.png'], book.resources.collect { |res| res.src })
+    assert_equal("A Basic Ochook", book.property_for('title'))
+    assert_equal(1, book.chapters.size)
+
+    chp = book.chapters.first
+    assert_equal("A Basic Ochook", chp.title)
+    assert_equal("index.html", chp.src)
+    assert_equal("Part One", chp.children[0].title)
+    assert_equal("index.html#part1", chp.children[0].src)
+    assert_equal("Part Two", chp.children[1].title)
+    assert_equal("index.html#part2", chp.children[1].src)
+  end
+
+
+  def test_write_from_epub
+    epub = Peregrin::Epub.read('test/fixtures/epubs/alice.epub')
+    book = epub.to_book
+    ook = Peregrin::Ochook.new(book)
+    ook.write('test/output/alice_ochook')
+    assert_nothing_raised {
+      Peregrin::Ochook.validate('test/output/alice_ochook')
+    }
+  end
+
+
+  def test_to_book_cover_html
+    ook = Peregrin::Ochook.read("test/fixtures/ochooks/illustrated")
+    book = ook.to_book(:componentize => true)
+    cov_html = book.components.detect { |cmpt|
+      cmpt.src == "cover.html"
+    }.contents
+    doc = Nokogiri::HTML::Document.parse(cov_html)
+    assert_equal('cover.png', doc.at_xpath('/html/body/div/img')['src'])
+  end
+
+
+  def test_to_book_toc_html
+    ook = Peregrin::Ochook.read("test/fixtures/ochooks/illustrated")
+    book = ook.to_book(:componentize => true)
+    toc_html = book.components.detect { |cmpt| cmpt.src == "toc.html" }.contents
+    doc = Nokogiri::HTML::Document.parse(toc_html)
+    assert_equal(3, doc.xpath('/html/body/ol/li').size)
+  end
+
+
+  def test_to_book_loi_html
+    ook = Peregrin::Ochook.read("test/fixtures/ochooks/illustrated")
+    book = ook.to_book(:componentize => true)
+    loi_html = book.components.detect { |cmpt| cmpt.src == "loi.html" }.contents
+    doc = Nokogiri::HTML::Document.parse(loi_html)
+    assert_equal(2, doc.xpath('/html/body/ol/li').size)
+  end
+
+
+  def test_to_book_rel_links
+    ook = Peregrin::Ochook.read("test/fixtures/ochooks/illustrated")
+    book = ook.to_book(:componentize => true)
+    cmpt_html = book.components[3].contents
+    doc = Nokogiri::HTML::Document.parse(cmpt_html)
+    assert_equal(
+      "cover.html",
+      doc.at_xpath('/html/head/link[@rel="start"]')['href']
+    )
+    assert_equal(
+      "toc.html",
+      doc.at_xpath('/html/head/link[@rel="contents"]')['href']
+    )
+    assert_equal(
+      "index.html",
+      doc.at_xpath('/html/head/link[@rel="first"]')['href']
+    )
+    assert_equal(
+      "part002.html",
+      doc.at_xpath('/html/head/link[@rel="last"]')['href']
+    )
+    assert_equal(
+      "part001.html",
+      doc.at_xpath('/html/head/link[@rel="prev"]')['href']
+    )
+  end
+
+end

data/test/formats/zhook_test.rb

@@ -0,0 +1,219 @@
+require 'test_helper'
+
+class Peregrin::Tests::ZhookTest < Test::Unit::TestCase
+
+  def test_validates
+    # File does not exist
+    assert_raise(Peregrin::Zhook::FileNotFound) {
+      Peregrin::Zhook.validate('test/fixtures/zhooks/invalid/missing.zhook')
+    }
+
+    # File extension is not .zhook
+    assert_raise(Peregrin::Zhook::WrongExtension) {
+      Peregrin::Zhook.validate('test/fixtures/zhooks/invalid/wrongext.zip')
+    }
+
+    # File is not a zip archive
+    assert_raise(Peregrin::Zhook::NotAZipArchive) {
+      Peregrin::Zhook.validate('test/fixtures/zhooks/invalid/notazip.zhook')
+    }
+
+    # Archive does not contain index.html
+    assert_raise(Peregrin::Zhook::MissingIndexHTML) {
+      Peregrin::Zhook.validate('test/fixtures/zhooks/invalid/noindex.zhook')
+    }
+
+    # Archive does not contain cover.png
+    assert_raise(Peregrin::Zhook::MissingCoverPNG) {
+      Peregrin::Zhook.validate('test/fixtures/zhooks/invalid/nocover.zhook')
+    }
+
+    # Index file has a HTML element with an id.
+    assert_raise(Peregrin::Zhook::IndexHTMLRootHasId) {
+      Peregrin::Zhook.validate('test/fixtures/zhooks/invalid/rootid.zhook')
+    }
+
+    # An actual valid .zhook
+    assert_nothing_raised {
+      Peregrin::Zhook.validate('test/fixtures/zhooks/flat.zhook')
+    }
+  end
+
+
+  def test_read
+    ook = Peregrin::Zhook.read('test/fixtures/zhooks/2level.zhook')
+    book = ook.to_book
+    assert_equal(1, book.components.length)
+    assert_equal("index.html", book.components.first.src)
+    assert_equal(['cover.png'], book.resources.collect { |res| res.src })
+    assert_equal("A Two-Level Zhook", book.property_for('title'))
+    assert_equal(1, book.chapters.size)
+    chp = book.chapters.first
+    assert_equal("A Two-Level Zhook", chp.title)
+    assert_equal("index.html", chp.src)
+    assert_equal("Part One", chp.children[0].title)
+    assert_equal("index.html#part1", chp.children[0].src)
+    assert_equal("Part Two", chp.children[1].title)
+    assert_equal("index.html#part2", chp.children[1].src)
+  end
+
+
+  def test_to_book_componentization
+    ook = Peregrin::Zhook.read('test/fixtures/zhooks/flat.zhook')
+    book = ook.to_book(:componentize => true)
+    assert_equal(
+      ["cover.html", "index.html", "part001.html", "part002.html", "toc.html"],
+      book.components.collect { |cmpt| cmpt.src }
+    )
+    assert_equal(3, book.chapters.size)
+    assert_equal("A Flat Zhook", book.chapters[0].title)
+    assert_equal("Part One", book.chapters[1].title)
+    assert_equal("Part Two", book.chapters[2].title)
+  end
+
+
+  def test_2_level_componentization
+    ook = Peregrin::Zhook.read('test/fixtures/zhooks/2level.zhook')
+    book = ook.to_book(:componentize => true)
+    assert_equal(1, book.chapters.size)
+    chp = book.chapters.first
+    assert_equal("A Two-Level Zhook", chp.title)
+    assert_equal("index.html", chp.src)
+    assert_equal("Part One", chp.children[0].title)
+    assert_equal("part001.html#part1", chp.children[0].src)
+    assert_equal("Part Two", chp.children[1].title)
+    assert_equal("part002.html#part2", chp.children[1].src)
+  end
+
+
+  def test_3_level_componentization
+    ook = Peregrin::Zhook.read('test/fixtures/zhooks/3level.zhook')
+    book = ook.to_book(:componentize => true)
+    assert_equal(1, book.chapters.size)
+    chp = book.chapters.first
+    assert_equal("A Three-Level Zhook", chp.title)
+    assert_equal("index.html", chp.src)
+    assert_equal("Part One", chp.children[0].title)
+    assert_equal("part001.html#part1", chp.children[0].src)
+    assert_equal("Part Two", chp.children[1].title)
+    assert_equal("part003.html#part2", chp.children[1].src)
+    assert_equal("Sub-part One Dot Two", chp.children[0].children[0].title)
+    assert_equal("part002.html", chp.children[0].children[0].src)
+    assert_equal("Sub-part Two Dot Two", chp.children[1].children[0].title)
+    assert_equal("part004.html", chp.children[1].children[0].src)
+  end
+
+
+  def test_stitching_components
+    book = Peregrin::Book.new
+    book.add_component(
+      "index.html",
+      %Q`
+        <html><head><title>Index</title>
+        <meta http-equiv="Content-Type" content="text/html;charset=US-ASCII">
+        </head><body>
+        <p>A para</p></body></html>
+      `
+    )
+    book.add_component(
+      "foo.html",
+      %Q`
+        <html><head><title>Foo</title>
+        <link rel="stylesheet" href="main.css" />
+        </head><body>
+        <hgroup><h1>Part Foo</h1><h2>Peregrin Took</h2></hgroup>
+        <cite>A cite tag</cite></body></html>
+      `
+    )
+    book.add_component("garply.html", %Q`<p>A floating para.</p>`)
+    ook = Peregrin::Zhook.new(book)
+    assert_equal(
+      whitewash(
+        %Q`
+        <!DOCTYPE html>
+        <html><head>
+          <title>Index</title>
+          <meta http-equiv="Content-Type" content="text/html;charset=US-ASCII">
+          <link rel="stylesheet" href="main.css">
+        </head><body>
+        <article>
+          <p>A para</p>
+        </article>
+        <article>
+          <hgroup><h1>Part Foo</h1><h2>Peregrin Took</h2></hgroup>
+          <cite>A cite tag</cite>
+        </article>
+        <article>
+          <p>A floating para.</p>
+        </article>
+        </body></html>`
+      ),
+      whitewash(ook.to_book.components.first.contents)
+    )
+  end
+
+
+  def test_consolidating_metadata
+    book = Peregrin::Book.new
+    book.add_component(
+      "index.html",
+      "<html><head><title>Foo</title>" +
+      "<meta http-equiv=\"Content-Type\" content=\"text/html;charset=US-ASCII\">" +
+      "</head><body><p>Foo</p></body></html>"
+    )
+    book.add_property("title", "Foo")
+    book.add_property("creator", "Peregrin Took")
+    ook = Peregrin::Zhook.new(book)
+    assert_equal(
+      whitewash(%Q`
+        <!DOCTYPE html>
+        <html><head><title>Foo</title>
+        <meta http-equiv="Content-Type" content="text/html;charset=US-ASCII">
+        <meta name="title" content="Foo">
+        <meta name="creator" content="Peregrin Took">
+        </head><body><p>Foo</p></body></html>
+      `),
+      whitewash(ook.to_book.components.first.contents)
+    )
+  end
+
+
+  def test_write_from_epub
+    epub = Peregrin::Epub.read('test/fixtures/epubs/alice.epub')
+    book = epub.to_book
+    ook = Peregrin::Zhook.new(book)
+    ook.write('test/output/alice.zhook')
+    assert_nothing_raised {
+      Peregrin::Zhook.validate('test/output/alice.zhook')
+    }
+  end
+
+
+  def test_convert_jpg_cover_on_write
+    # Create an epub object, convert it to a book, and verify that the cover
+    # is a JPEG.
+    epub = Peregrin::Epub.read('test/fixtures/epubs/alice.epub')
+    book = epub.to_book
+    assert_equal(
+      "www.gutenberg.org@files@19033@19033-h@images@cover_th.jpg",
+      book.cover.src
+    )
+
+    # Write the book to file as a Zhook, which should convert the cover to PNG.
+    ook = Peregrin::Zhook.new(book)
+    ook.write('test/output/alice.zhook')
+
+    # Load the Zhook from file, and check that it has a cover.png.
+    ook2 = Peregrin::Zhook.read('test/output/alice.zhook')
+    book2 = ook2.to_book
+    assert_equal("cover.png", book2.cover.src)
+
+    # Validate the cover.png using ImageMagick's identify
+    IO.popen("identify -", "r+") { |io|
+      io.write(book2.read_resource(book2.cover))
+      io.close_write
+      assert_match(/^[^\s]+ PNG /, io.read)
+    }
+  end
+
+end

data/test/test_helper.rb

@@ -0,0 +1,16 @@
+require 'test/unit'
+require 'peregrin'
+
+module Peregrin::Tests
+
+end
+
+
+
+class Test::Unit::TestCase
+
+  def whitewash(str)
+    str.gsub(/\s+/, '')
+  end
+
+end

data/test/utils/componentizer_test.rb

@@ -0,0 +1,78 @@
+require 'test_helper'
+
+class Peregrin::Tests::ComponentizerTest < Test::Unit::TestCase
+
+  def test_processing
+    cz = process_fixture("components1.html")
+    assert_equal(
+      [
+        '/html/body',
+        '/html/body/article[1]',
+        '/html/body/article[1]/article[1]',
+        '/html/body/article[1]/article[2]',
+        '/html/body/article[2]'
+      ],
+      cz.component_xpaths
+    )
+  end
+
+
+  def test_processing_where_body_should_be_empty
+    cz = process_fixture("components2.html")
+    assert_equal(
+      [
+        '/html/body/article[1]',
+        '/html/body/article[1]/article[1]',
+        '/html/body/article[1]/article[2]',
+        '/html/body/article[2]',
+        '/html/body/article[2]/article'
+      ],
+      cz.component_xpaths
+    )
+  end
+
+
+  def test_generate_component
+    cz = process_fixture("components1.html")
+    assert_equal(
+      whitewash(
+        "<!DOCTYPE html>" +
+        "<html><head><title>Components test 1</title></head><body>" +
+        "<article><h2>B</h2></article>" +
+        "</body></html>"
+      ),
+      whitewash(cz.generate_component('/html/body/article[1]').to_html)
+    )
+  end
+
+
+  def test_write_component
+    cz = process_fixture("components1.html")
+    tmp_path = "test/fixtures/componentizer/tmp.html"
+    cz.write_component("/html/body", tmp_path) { |doc| doc.to_xhtml }
+    assert_equal(
+      whitewash(
+        "<!DOCTYPE html>" +
+        '<html xmlns="http://www.w3.org/1999/xhtml"><head>' +
+        '<meta http-equiv="Content-Type" content="text/html;charset=UTF-8" />' +
+        "<title>Components test 1</title></head>" +
+        "<body><h1>A</h1></body></html>"
+      ),
+      whitewash(IO.read(tmp_path))
+    )
+  ensure
+    File.unlink(tmp_path)  if File.exists?(tmp_path)
+  end
+
+
+  protected
+
+    def process_fixture(filename)
+      fx = IO.read("test/fixtures/componentizer/#{filename}")
+      doc = Nokogiri::HTML::Document.parse(fx)
+      cz = Peregrin::Componentizer.new(doc)
+      cz.process(doc.at_xpath('/html/body'))
+      cz
+    end
+
+end

data/test/utils/outliner_test.rb

@@ -0,0 +1,49 @@
+require 'test_helper'
+
+class Peregrin::Tests::OutlinerTest < Test::Unit::TestCase
+
+  def test_spec_1
+    load_spec_and_compare_out('spec1')
+  end
+
+
+  def test_spec_2
+    load_spec_and_compare_out('spec2') { |section, below|
+      if section.heading_text
+        section.heading_text
+      elsif section.node
+        "<i>Untitled #{section.node.name.upcase}</i>"
+      end
+    }
+  end
+
+
+  def test_spec_3a
+    load_spec_and_compare_out('spec3a')
+  end
+
+
+  def test_spec_3b
+    load_spec_and_compare_out('spec3b')
+  end
+
+
+  def test_spec_4
+    load_spec_and_compare_out('spec4')
+  end
+
+
+  protected
+
+    def load_spec_and_compare_out(spec_name, &blk)
+      src_file = File.new("test/fixtures/outliner/#{spec_name}.doc.html")
+      cmp_file = File.new("test/fixtures/outliner/#{spec_name}.out.html")
+      doc = Nokogiri::HTML::Document.parse(src_file)
+      outliner = Peregrin::Outliner.new(doc.root)
+      outliner.process(doc.root)
+      out = outliner.to_html(&blk)
+      cmp = cmp_file.read
+      assert_equal(whitewash(cmp), whitewash(out))
+    end
+
+end

metadata

@@ -0,0 +1,135 @@
+--- !ruby/object:Gem::Specification 
+name: peregrin
+version: !ruby/object:Gem::Version 
+  prerelease: false
+  segments: 
+  - 1
+  - 1
+  - 1
+  version: 1.1.1
+platform: ruby
+authors: 
+- Joseph Pearson
+autorequire: 
+bindir: bin
+cert_chain: []
+
+date: 2011-06-28 00:00:00 +10:00
+default_executable: 
+dependencies: 
+- !ruby/object:Gem::Dependency 
+  name: nokogiri
+  prerelease: false
+  requirement: &id001 !ruby/object:Gem::Requirement 
+    requirements: 
+    - - ">="
+      - !ruby/object:Gem::Version 
+        segments: 
+        - 0
+        version: "0"
+  type: :runtime
+  version_requirements: *id001
+- !ruby/object:Gem::Dependency 
+  name: zipruby
+  prerelease: false
+  requirement: &id002 !ruby/object:Gem::Requirement 
+    requirements: 
+    - - ">="
+      - !ruby/object:Gem::Version 
+        segments: 
+        - 0
+        version: "0"
+  type: :runtime
+  version_requirements: *id002
+- !ruby/object:Gem::Dependency 
+  name: mime-types
+  prerelease: false
+  requirement: &id003 !ruby/object:Gem::Requirement 
+    requirements: 
+    - - ">="
+      - !ruby/object:Gem::Version 
+        segments: 
+        - 0
+        version: "0"
+  type: :runtime
+  version_requirements: *id003
+- !ruby/object:Gem::Dependency 
+  name: rake
+  prerelease: false
+  requirement: &id004 !ruby/object:Gem::Requirement 
+    requirements: 
+    - - ">="
+      - !ruby/object:Gem::Version 
+        segments: 
+        - 0
+        version: "0"
+  type: :development
+  version_requirements: *id004
+description: Peregrin converts EPUBs, Zhooks and Ochooks.
+email: joseph@inventivelabs.com.au
+executables: 
+- peregrin
+extensions: []
+
+extra_rdoc_files: 
+- README.md
+- MIT-LICENSE
+files: 
+- bin/peregrin
+- lib/formats/epub.rb
+- lib/formats/ochook.rb
+- lib/formats/zhook.rb
+- lib/peregrin/book.rb
+- lib/peregrin/chapter.rb
+- lib/peregrin/component.rb
+- lib/peregrin/componentizer.rb
+- lib/peregrin/outliner.rb
+- lib/peregrin/property.rb
+- lib/peregrin/resource.rb
+- lib/peregrin/version.rb
+- lib/peregrin/zip_patch.rb
+- lib/peregrin.rb
+- test/conversion_test.rb
+- test/formats/epub_test.rb
+- test/formats/ochook_test.rb
+- test/formats/zhook_test.rb
+- test/test_helper.rb
+- test/utils/componentizer_test.rb
+- test/utils/outliner_test.rb
+- README.md
+- MIT-LICENSE
+has_rdoc: true
+homepage: http://ochook.org/peregrin
+licenses: []
+
+post_install_message: 
+rdoc_options: 
+- --title
+- Peregrin
+- --main
+- README.md
+require_paths: 
+- lib
+required_ruby_version: !ruby/object:Gem::Requirement 
+  requirements: 
+  - - ">="
+    - !ruby/object:Gem::Version 
+      segments: 
+      - 0
+      version: "0"
+required_rubygems_version: !ruby/object:Gem::Requirement 
+  requirements: 
+  - - ">="
+    - !ruby/object:Gem::Version 
+      segments: 
+      - 0
+      version: "0"
+requirements: []
+
+rubyforge_project: nowarning
+rubygems_version: 1.3.6
+signing_key: 
+specification_version: 3
+summary: Peregrin - ebook conversion
+test_files: []
+