extractor 0.1

data/README

@@ -0,0 +1,10 @@
+This is a simple binding for libextractor http://gnunet.org/libextractor/. 
+It allows you to read metadata from a wide number of file types.
+
+It can be used this way
+  require 'extractor'
+  metadata = Extractor.extract('happy.mp3')
+  puts metadata['mimetype']
+  puts metadata['artist']
+Sometimes there will be more than one entry for a metadata type. For example, an mp3 may have two artists. If this is the case then you might have
+  metadata['artist'] == ['Puff Daddy', 'The Family']

data/Rakefile

@@ -0,0 +1,61 @@
+require 'rake'
+require 'rake/testtask'
+require 'rake/clean'
+require 'rake/gempackagetask'
+require 'rake/rdoctask'
+require 'fileutils'
+include FileUtils
+
+CLEAN.add %w{ extractor.o }
+CLOBBER.add %w{ Makefile mkmf.log extractor.bundle }
+
+PKGFILES = FileList.new
+PKGFILES.add %w{ README Rakefile extconf.rb extractor.c test/*}
+
+file "Makefile" do |t|
+  ruby "extconf.rb"
+end
+
+task :default => [:compile, :test]
+
+desc "Compile the code"
+task :compile => ["Makefile"] do |t|
+  sh 'make'
+end
+
+Rake::RDocTask.new do |rd|
+  rd.main = "README"
+  rd.rdoc_files = Rake::FileList['README', 'extractor.c']
+end
+
+Rake::TestTask.new do |t|
+  t.test_files = ['test/test.rb']
+  t.verbose = true
+end
+
+spec = Gem::Specification.new do |s|
+  s.summary = "A binding for libextractor; allows you to read metadata from a number of file types."
+  s.name = 'extractor'
+  s.author = 'Ry Dahl'
+  s.email = 'ry@tinyclouds.org'
+  s.version = '0.1'
+
+  s.rubyforge_project = 'extractor'
+  
+  s.has_rdoc = true
+  s.extra_rdoc_files = ['README', 'extractor.c']
+  
+  s.test_files = ['test/test.rb']
+  
+  s.extensions = ['extconf.rb']
+  
+  #s.requirements << 'extractor'
+  s.require_path = '.'
+  s.autorequire = 'rake'
+  
+  s.files = PKGFILES.to_a
+end
+
+Rake::GemPackageTask.new(spec) do |pkg|
+  pkg.need_zip = true
+end

data/extconf.rb

@@ -0,0 +1,7 @@
+require 'mkmf'
+
+dir_config("libextractor")
+
+if have_library('extractor') and have_header('extractor.h')
+  create_makefile ('extractor')
+end

data/extractor.c

@@ -0,0 +1,46 @@
+#include <ruby.h>
+#include <extractor.h>
+
+EXTRACTOR_ExtractorList *rb_extractor_libraries;
+
+/* Extracts from the filename parameter a hash of metadata. */
+VALUE rb_extractor_extract(VALUE module, VALUE filename)
+{
+  EXTRACTOR_KeywordList *keywords, *k;
+  VALUE new_value, old_value, hash_key, hash = rb_hash_new();
+  
+  
+  keywords = EXTRACTOR_getKeywords(rb_extractor_libraries, STR2CSTR(filename));
+  
+  for(k = keywords; k; k = k->next) {
+    hash_key = rb_str_new2(EXTRACTOR_getKeywordTypeAsString(k->keywordType));  
+    old_value = rb_hash_aref(hash, hash_key);
+    new_value = rb_str_new2(k->keyword);
+    
+    // if there is no value yet, then set it    
+    if(TYPE(old_value) == T_NIL) { 
+      rb_hash_aset(hash, hash_key, new_value);
+    
+    // if we have a string, then create an array with both values  
+    } else if(TYPE(old_value) == T_STRING) { 
+      rb_hash_aset(hash, hash_key,
+        rb_ary_new3(2, old_value, new_value));
+    
+    // if we have an array already, add the new value    
+    } else { 
+      rb_ary_push(old_value, new_value);
+    }  
+  }
+  
+  EXTRACTOR_freeKeywords(keywords);
+  
+  return hash;
+}
+
+void Init_extractor ()
+{
+  VALUE mExtractor;
+  rb_extractor_libraries = EXTRACTOR_loadDefaultLibraries();
+  mExtractor = rb_define_module("Extractor");
+  rb_define_module_function (mExtractor, "extract", rb_extractor_extract, 1);
+}

data/test/test.rb

@@ -0,0 +1,31 @@
+require 'test/unit'
+require File.dirname(__FILE__) + '/../extractor'
+
+class BasicTest < Test::Unit::TestCase
+  
+  def setup
+    @mp3_filename = File.dirname(__FILE__) + '/things_fall_apart.mp3'
+  end
+  
+  def test_extract_types
+    metadata = Extractor.extract(@mp3_filename)
+    
+    metadata.each do |type, keyword|
+      assert_kind_of( String, type )
+      if keyword.kind_of?(Array)
+        keyword.each { |k| assert_kind_of( String, k ) }
+      else
+        assert_kind_of( String, keyword )
+      end
+    end
+    
+  end
+  
+  
+  def test_extract
+    metadata = Extractor.extract(@mp3_filename)
+    assert_equal(metadata['mimetype'], 'audio/mpeg')
+    assert_equal(metadata['artist'], 'The Roots')
+  end
+  
+end

metadata

@@ -0,0 +1,52 @@
+--- !ruby/object:Gem::Specification 
+rubygems_version: 0.9.1
+specification_version: 1
+name: extractor
+version: !ruby/object:Gem::Version 
+  version: "0.1"
+date: 2007-02-26 00:00:00 +01:00
+summary: A binding for libextractor; allows you to read metadata from a number of file types.
+require_paths: 
+- .
+email: ry@tinyclouds.org
+homepage: 
+rubyforge_project: extractor
+description: 
+autorequire: rake
+default_executable: 
+bindir: bin
+has_rdoc: true
+required_ruby_version: !ruby/object:Gem::Version::Requirement 
+  requirements: 
+  - - ">"
+    - !ruby/object:Gem::Version 
+      version: 0.0.0
+  version: 
+platform: ruby
+signing_key: 
+cert_chain: 
+post_install_message: 
+authors: 
+- Ry Dahl
+files: 
+- README
+- Rakefile
+- extconf.rb
+- extractor.c
+- test/test.rb
+- test/things_fall_apart.mp3
+test_files: 
+- test/test.rb
+rdoc_options: []
+
+extra_rdoc_files: 
+- README
+- extractor.c
+executables: []
+
+extensions: 
+- extconf.rb
+requirements: []
+
+dependencies: []
+