hydra-file_characterization 0.2.9 → 0.3.0

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA1:
-  metadata.gz: 3359fdef2f631bbd7c3af4ac3c56062c6fbb391a
-  data.tar.gz: daa36c8addb90aec847568e30c3d98512fddb783
+  metadata.gz: e4919a2449be3e3a6127fdc17a5f8552e3b64821
+  data.tar.gz: 1ab2ab15d03a9e4ddfd83b15e2f78a2be5548d86
 SHA512:
-  metadata.gz: 11a66dc69441eb624743bfdd49f851aa88fa15d8463dad586472c5463fc1141a5a23db496b0ce32023f2b658a91fbefe96ac1d30d58e4b5b53dda59e716b8692
-  data.tar.gz: 362f5bfc8cb21590d9290ef18ad6d56cea968d5959bd47e7e1ec2ebaeba0cbc2af7d5e611a8802c50d537f7fca9bfc1e6f3bc1ab7dc2f7265988951b9520e3cb
+  metadata.gz: d5494b689812d74a60ecca77216cc31302b8da2694635ee59be8b5b98c844bb936bd86a7ae3b09901375cc441c795ff7389b9209c36d41daaa5ba039e00cd6dd
+  data.tar.gz: 5e828f94c57282b40fd42a6bdaaf7e30a74e8e3d85871ec3eef8e43f444fefb9eeca6a0946629a061078715f291e5dd22bdb8abd85bfa5f7c9bc900cddd107e8

data/lib/hydra/file_characterization.rb

@@ -39,6 +39,9 @@ module Hydra
     # @example
     #   fits_xml, ffprobe_xml = Hydra::FileCharacterization.characterize(contents_of_a_file, 'file.rb', :fits, :ffprobe)
     #
+    # @example With an open file
+    #   fits_xml, ffprobe_xml = Hydra::FileCharacterization.characterize(File.open('foo.mkv'), :fits, :ffprobe)
+    #
     # @param [String] content - The contents of the original file
     # @param [String] filename - The original file's filename; Some
     #   characterization tools take hints from the file names
@@ -55,16 +58,13 @@ module Hydra
     #    key to the yieldparam with a value, that value will be used as the path
     #
     # @see Hydra::FileCharacterization.configure
-    def self.characterize(content, filename, *tool_names)
-      tool_outputs = []
+    def self.characterize(*args)
+      content, filename, tool_names = extract_arguments(args)
       tool_names = Array(tool_names).flatten.compact
       custom_paths = {}
       yield(custom_paths) if block_given?
-      FileCharacterization::ToTempFile.open(filename, content) do |f|
-        tool_names.each do |tool_name|
-          tool_outputs << FileCharacterization.characterize_with(tool_name, f.path, custom_paths[tool_name])
-        end
-      end
+      
+      tool_outputs = run_characterizers(content, filename, tool_names, custom_paths)
       tool_names.size == 1 ? tool_outputs.first : tool_outputs
     end
 
@@ -73,6 +73,45 @@ module Hydra
       yield(configuration)
     end
 
+    private
+
+      # Break up a list of arguments into two possible lists:
+      #   option1:  [String] content, [String] filename, [Array] tool_names
+      #   option2:  [File] content, [Array] tool_names
+      # In the case of option2, derive the filename from the file's path  
+      # @return [String, File], [String], [Array]
+      def self.extract_arguments(args)
+        content = args.shift
+        filename = if content.is_a? File
+          File.basename(content.path)
+        else
+           args.shift
+        end
+        tool_names = args
+        return content, filename, tool_names
+      end
+
+      # @param [File, String] content Either an open file or a string. If a string is passed
+      #                               a temp file will be created
+      # @param [String] filename Used in creating a temp file name
+      # @param [Array<Symbol>] tool_names A list of symbols referencing the characerization tools to run
+      # @param [Hash] custom_paths The paths to the executables of the tool.
+      def self.run_characterizers(content, filename, tool_names, custom_paths)
+        if content.is_a? File
+          run_characterizers_on_file(content, tool_names, custom_paths)
+        else 
+          FileCharacterization::ToTempFile.open(filename, content) do |f|
+            run_characterizers_on_file(f, tool_names, custom_paths)
+          end
+        end
+      end
+
+      def self.run_characterizers_on_file(f, tool_names, custom_paths)
+        tool_names.map do |tool_name|
+           FileCharacterization.characterize_with(tool_name, f.path, custom_paths[tool_name])
+        end
+      end
+
     class Configuration
       def tool_path(tool_name, tool_path)
         Hydra::FileCharacterization.characterizer(tool_name).tool_path = tool_path
@@ -80,4 +119,4 @@ module Hydra
     end
 
   end
-end
+end

data/lib/hydra/file_characterization/characterizer.rb

@@ -27,11 +27,15 @@ module Hydra::FileCharacterization
     end
 
     def tool_path
-      @tool_path || self.class.tool_path || (raise Hydra::FileCharacterization::UnspecifiedToolPathError.new(self.class))
+      @tool_path || self.class.tool_path || convention_based_tool_name
     end
 
     protected
 
+    def convention_based_tool_name
+      self.class.name.split("::").last.downcase
+    end
+
     def internal_call
       stdin, stdout, stderr, wait_thr = popen3(command)
       begin

data/lib/hydra/file_characterization/exceptions.rb

@@ -1,11 +1,5 @@
 module Hydra::FileCharacterization
 
-  class UnspecifiedToolPathError < RuntimeError
-    def initialize(tool_class)
-      super("Unspecified tool path for #{tool_class}")
-    end
-  end
-
   class FileNotFoundError < RuntimeError
   end
 

data/lib/hydra/file_characterization/version.rb

@@ -1,5 +1,5 @@
 module Hydra
   module FileCharacterization
-    VERSION = "0.2.9"
+    VERSION = "0.3.0"
   end
 end

data/spec/lib/hydra/file_characterization/characterizer_spec.rb

@@ -45,11 +45,7 @@ module Hydra::FileCharacterization
       end
 
       context 'without a specified tool_path' do
-        it 'should raise Hydra::FileCharacterization::UnspecifiedToolPathError' do
-          expect {
-            subject.tool_path
-          }.to raise_error(Hydra::FileCharacterization::UnspecifiedToolPathError)
-        end
+        its(:tool_path) { should eq 'characterizer' }
       end
     end
   end

data/spec/lib/hydra/file_characterization_spec.rb

@@ -7,56 +7,67 @@ module Hydra
   describe FileCharacterization do
 
     describe '.characterize' do
-      let(:content) { "class Test; end\n" }
-      let(:filename) { 'test.rb' }
-      subject { Hydra::FileCharacterization.characterize(content, filename, tool_names) }
+      describe "for content in memory" do
+        let(:content) { "class Test; end\n" }
+        let(:filename) { 'test.rb' }
+        subject { Hydra::FileCharacterization.characterize(content, filename, tool_names) }
 
-      describe 'for fits' do
-        let(:tool_names) { [:fits] }
-        it { should match(/#{'<identity format="Plain text" mimetype="text/plain"'}/) }
-      end
+        describe 'for fits' do
+          let(:tool_names) { [:fits] }
+          it { should match(/#{'<identity format="Plain text" mimetype="text/plain"'}/) }
+        end
 
-      describe 'with configured path' do
-        it {
-          response = Hydra::FileCharacterization.characterize(content, filename, :fits) do |config|
-            config[:fits] = `which fits || which fits.sh`.strip
-          end
-          expect(response).to match(/#{'<identity format="Plain text" mimetype="text/plain"'}/)
-        }
-      end
+        describe 'with configured path' do
+          it {
+            response = Hydra::FileCharacterization.characterize(content, filename, :fits) do |config|
+              config[:fits] = `which fits || which fits.sh`.strip
+            end
+            expect(response).to match(/#{'<identity format="Plain text" mimetype="text/plain"'}/)
+          }
+        end
 
-      describe 'with multiple runs' do
-        it {
-          response_1, response_2, response_3 = Hydra::FileCharacterization.characterize(content, filename, :fits, :fits)
-          expect(response_1).to match(/#{'<identity format="Plain text" mimetype="text/plain"'}/)
-          expect(response_2).to match(/#{'<identity format="Plain text" mimetype="text/plain"'}/)
-          expect(response_3).to be_nil
-        }
-      end
+        describe 'with multiple runs' do
+          it {
+            response_1, response_2, response_3 = Hydra::FileCharacterization.characterize(content, filename, :fits, :fits)
+            expect(response_1).to match(/#{'<identity format="Plain text" mimetype="text/plain"'}/)
+            expect(response_2).to match(/#{'<identity format="Plain text" mimetype="text/plain"'}/)
+            expect(response_3).to be_nil
+          }
+        end
 
-      describe 'for a bogus tool' do
-        let(:tool_names) { [:cookie_monster] }
-        it {
-          expect {
-            subject
-          }.to raise_error(Hydra::FileCharacterization::ToolNotFoundError)
-        }
-      end
+        describe 'for a bogus tool' do
+          let(:tool_names) { [:cookie_monster] }
+          it {
+            expect {
+              subject
+            }.to raise_error(Hydra::FileCharacterization::ToolNotFoundError)
+          }
+        end
 
-      describe 'for a mix of bogus and valid tools' do
-        let(:tool_names) { [:fits, :cookie_monster] }
-        it {
-          expect {
-            subject
-          }.to raise_error(Hydra::FileCharacterization::ToolNotFoundError)
-        }
-      end
+        describe 'for a mix of bogus and valid tools' do
+          let(:tool_names) { [:fits, :cookie_monster] }
+          it {
+            expect {
+              subject
+            }.to raise_error(Hydra::FileCharacterization::ToolNotFoundError)
+          }
+        end
 
-      describe 'for no tools' do
-        let(:tool_names) { nil }
-        it { should eq [] }
+        describe 'for no tools' do
+          let(:tool_names) { nil }
+          it { should eq [] }
+        end
       end
 
+      describe "for a file on disk" do
+        let(:file) { File.open(fixture_file('brendan_behan.jpeg')) }
+        subject { Hydra::FileCharacterization.characterize(file, tool_names) }
+
+        describe 'for fits' do
+          let(:tool_names) { [:fits] }
+          it { should match(/#{'<identity format="JPEG File Interchange Format" mimetype="image/jpeg"'}/) }
+        end
+      end
     end
     describe '.configure' do
       let(:content) { "class Test; end\n" }
@@ -71,12 +82,11 @@ module Hydra
         Hydra::FileCharacterization.configure do |config|
           config.tool_path(:fits, nil)
         end
+        response = Hydra::FileCharacterization.characterize(content, filename, :fits)
+        expect(response).to match(/#{'<identity format="Plain text" mimetype="text/plain"'}/)
 
-        expect {
-          Hydra::FileCharacterization.characterize(content, filename, :fits)
-        }.to raise_error(Hydra::FileCharacterization::UnspecifiedToolPathError)
       end
     end
 
   end
-end
+end

metadata

@@ -1,7 +1,7 @@
 --- !ruby/object:Gem::Specification
 name: hydra-file_characterization
 version: !ruby/object:Gem::Version
-  version: 0.2.9
+  version: 0.3.0
 platform: ruby
 authors:
 - James Treacy
@@ -11,7 +11,7 @@ authors:
 autorequire: 
 bindir: bin
 cert_chain: []
-date: 2013-10-12 00:00:00.000000000 Z
+date: 2013-10-24 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: activesupport