chupa-text-decomposer-pdf 1.0.1 → 1.0.2

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA1:
-  metadata.gz: 6e31d3498186c960a9dae07539e83f77ffa5b995
-  data.tar.gz: 6ecc6033cd1420eb5c7dba3540cf9b4bdca569e9
+  metadata.gz: 71901aa054dc878f955620473635ee11df436a2b
+  data.tar.gz: 68bbc514ce94a38865e48a27e1849c1baec7df9e
 SHA512:
-  metadata.gz: 661698768712b75f81cccef8df76cdd86ea76ef95bb38320d0e713c2a2a0a68cc222c62c64219baad59b72788bcb5ff5e0fee0277ef2f74125c3a6e24d1a6ca6
-  data.tar.gz: 54d47d4bbd031c02f8774200177b29c5c112a574f812bc68ee43d20019a00e461e8ca4aad955f6c50913e28578656ffdaa12fdca33e04d603df32f9c79662c62
+  metadata.gz: 66d265273482895c235c053c5cd69ae92a9510288abcc2a72a8a4f686d9942c15fe6ae4d442d13cf95bb21f5696b81aac68e354c7a023b25da1b292aaece0ade
+  data.tar.gz: 649fa65ae2b74fc051896afc161eb91c8db9ad04626af94d287bf32bf8961a0ce7feb521f635ccf043297172d550ce2bb1c59ea826e211c43eb6a8fa84d1a943

data/chupa-text-decomposer-pdf.gemspec

@@ -22,7 +22,7 @@ end
 
 Gem::Specification.new do |spec|
   spec.name = "chupa-text-decomposer-pdf"
-  spec.version = "1.0.1"
+  spec.version = "1.0.2"
   spec.homepage = "https://github.com/ranguba/chupa-text-decomposer-pdf"
   spec.authors = ["Kouhei Sutou"]
   spec.email = ["kou@clear-code.com"]

data/doc/text/news.md

@@ -1,5 +1,11 @@
 # News
 
+## 1.0.2: 2014-02-18
+
+### Improvements
+
+  * Supported encrypted PDF. Use `:password` option.
+
 ## 1.0.1: 2014-02-16
 
 ### Improvements

data/lib/chupa-text/decomposers/pdf.rb

@@ -1,4 +1,4 @@
-# Copyright (C) 2013  Kouhei Sutou <kou@clear-code.com>
+# Copyright (C) 2013-2014  Kouhei Sutou <kou@clear-code.com>
 #
 # This library is free software; you can redistribute it and/or
 # modify it under the terms of the GNU Lesser General Public
@@ -29,13 +29,13 @@ module ChupaText
       end
 
       def decompose(data)
-        document = Poppler::Document.new(data.body)
+        document = create_document(data)
         text = ""
         document.each do |page|
           page_text = page.get_text
           next if page_text.empty?
-          text << "\n" unless text.empty?
           text << page_text
+          text << "\n" unless page_text.end_with?("\n")
         end
         text_data = TextData.new(text)
         text_data.uri = data.uri
@@ -45,17 +45,52 @@ module ChupaText
         add_attribute(text_data, document, :keywords)
         add_attribute(text_data, document, :creator)
         add_attribute(text_data, document, :producer)
-        add_attribute(text_data, document, :creation_date)
+        add_attribute(text_data, document, :creation_date, :created_time)
         yield(text_data)
       end
 
       private
-      def add_attribute(text_data, document, name)
-        value = document.send(name)
+      def create_document(data)
+        _password = password(data)
+        begin
+          wrap_stderr do
+            Poppler::Document.new(data.body, _password)
+          end
+        rescue GLib::Error => error
+          case error.code
+          when Poppler::Error::ENCRYPTED.to_i
+            raise ChupaText::EncryptedError.new(data)
+          else
+            raise ChupaText::InvalidDataError.new(data, error.message)
+          end
+        end
+      end
+
+      def password(data)
+        password = @options[:password]
+        if password.respond_to?(:call)
+          password = password.call(data)
+        end
+        password
+      end
+
+      def wrap_stderr
+        stderr = $stderr.dup
+        input, output = IO.pipe
+        _ = input # TODO: Report output
+        $stderr.reopen(output)
+        yield
+      ensure
+        $stderr.reopen(stderr)
+      end
+
+      def add_attribute(text_data, document,
+                        pdf_attribute_name, data_attribute_name=nil)
+        value = document.send(pdf_attribute_name)
         return if value.nil?
-        attribute_name = name.to_s.gsub(/_/, "-")
         value = Time.at(value).utc.iso8601 if value.is_a?(Integer)
-        text_data[attribute_name] = value
+        data_attribute_name ||= pdf_attribute_name.to_s.gsub(/_/, "-")
+        text_data[data_attribute_name] = value
       end
     end
   end

data/test/run-test.rb

@@ -18,6 +18,8 @@
 
 $VERBOSE = true
 
+ENV["TZ"] = "JST"
+
 require "bundler/setup"
 
 require "test-unit"

data/test/test-pdf.rb

@@ -1,4 +1,4 @@
-# Copyright (C) 2013  Kouhei Sutou <kou@clear-code.com>
+# Copyright (C) 2013-2014  Kouhei Sutou <kou@clear-code.com>
 #
 # This library is free software; you can redistribute it and/or
 # modify it under the terms of the GNU Lesser General Public
@@ -18,10 +18,14 @@ require "pathname"
 
 class TestPDF < Test::Unit::TestCase
   def setup
-    @decomposer = ChupaText::Decomposers::PDF.new({})
+    @options = {}
   end
 
   private
+  def decomposer
+    ChupaText::Decomposers::PDF.new(@options)
+  end
+
   def fixture_path(*components)
     base_path = Pathname(__FILE__).dirname + "fixture"
     base_path.join(*components)
@@ -37,11 +41,11 @@ class TestPDF < Test::Unit::TestCase
       end
 
       def test_pdf
-        assert_true(@decomposer.target?(create_data("index.pdf")))
+        assert_true(decomposer.target?(create_data("index.pdf")))
       end
 
       def test_html
-        assert_false(@decomposer.target?(create_data("index.html")))
+        assert_false(decomposer.target?(create_data("index.html")))
       end
     end
 
@@ -53,11 +57,11 @@ class TestPDF < Test::Unit::TestCase
       end
 
       def test_pdf
-        assert_true(@decomposer.target?(create_data("application/pdf")))
+        assert_true(decomposer.target?(create_data("application/pdf")))
       end
 
       def test_html
-        assert_false(@decomposer.target?(create_data("text/html")))
+        assert_false(decomposer.target?(create_data("text/html")))
       end
     end
   end
@@ -69,7 +73,7 @@ class TestPDF < Test::Unit::TestCase
       data.mime_type = "text/pdf"
 
       decomposed = []
-      @decomposer.decompose(data) do |decomposed_data|
+      decomposer.decompose(data) do |decomposed_data|
         decomposed << decomposed_data
       end
       decomposed
@@ -100,8 +104,9 @@ class TestPDF < Test::Unit::TestCase
         assert_equal(["LibreOffice 4.1"], decompose("producer"))
       end
 
-      def test_creation_date
-        assert_equal([nil], decompose("creation_date"))
+      def test_created_time
+        assert_equal([Time.parse("2014-01-06T00:52:45+09:00")],
+                     decompose("created_time"))
       end
 
       private
@@ -114,7 +119,7 @@ class TestPDF < Test::Unit::TestCase
 
     sub_test_case("one page") do
       def test_body
-        assert_equal(["Page1"], decompose.collect(&:body))
+        assert_equal(["Page1\n"], decompose.collect(&:body))
       end
 
       private
@@ -125,7 +130,7 @@ class TestPDF < Test::Unit::TestCase
 
     sub_test_case("multi pages") do
       def test_body
-        assert_equal(["Page1\nPage2"], decompose.collect(&:body))
+        assert_equal(["Page1\nPage2\n"], decompose.collect(&:body))
       end
 
       private
@@ -133,5 +138,30 @@ class TestPDF < Test::Unit::TestCase
         super(fixture_path("multi-pages.pdf"))
       end
     end
+
+    sub_test_case("encrypted") do
+      def test_with_password
+        @options = {:password => "encrypted"}
+        assert_equal(["Password is 'encrypted'.\n"],
+                     decompose.collect(&:body))
+      end
+
+      def test_with_password_block
+        @options = {:password => lambda {|data| "encrypted"}}
+        assert_equal(["Password is 'encrypted'.\n"],
+                     decompose.collect(&:body))
+      end
+
+      def test_without_password
+        assert_raise(ChupaText::EncryptedError) do
+          decompose
+        end
+      end
+
+      private
+      def decompose
+        super(fixture_path("encrypted.pdf"))
+      end
+    end
   end
 end

metadata

@@ -1,14 +1,14 @@
 --- !ruby/object:Gem::Specification
 name: chupa-text-decomposer-pdf
 version: !ruby/object:Gem::Version
-  version: 1.0.1
+  version: 1.0.2
 platform: ruby
 authors:
 - Kouhei Sutou
 autorequire: 
 bindir: bin
 cert_chain: []
-date: 2014-02-16 00:00:00.000000000 Z
+date: 2014-02-17 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: chupa-text
@@ -130,7 +130,9 @@ files:
 - test/test-pdf.rb
 - test/fixture/one-page.pdf
 - test/fixture/multi-pages.pdf
+- test/fixture/encrypted.pdf
 - test/fixture/attributes.pdf
+- test/fixture/encrypted.odt
 - test/fixture/one-page.odt
 - test/fixture/multi-pages.odt
 - test/fixture/attributes.odt