rflow 1.0.0a1 → 1.0.0a2

data/spec/rflow/message/data_spec.rb

@@ -0,0 +1,58 @@
+require 'spec_helper'
+require 'rflow/message'
+
+class RFlow
+  class Message
+    describe Data do
+      let(:string) { 'this is a string to be serialized' }
+      let(:invalid_schema) { 'invalid schema' }
+      let(:valid_schema) { '{"type": "string"}' }
+      let(:serialized_string) { encode_avro(valid_schema, string) }
+
+      context "if created without a schema" do
+        it "should throw an exception" do
+          expect { Data.new(nil) }.to raise_error(ArgumentError, /^Invalid schema/)
+        end
+      end
+
+      context "if created with an invalid schema for the serialization" do
+        ['avro', :avro].each do |it|
+          it "should throw an exception for serialization type #{it.inspect}" do
+            expect { Data.new(invalid_schema, it) }.to raise_error(ArgumentError, /^Invalid schema/)
+          end
+        end
+      end
+
+      context "if created with a valid avro schema" do
+        ['avro', :avro].each do |it|
+          it "should instantiate correctly for serialization type #{it.inspect}" do
+            expect { Data.new(valid_schema, it) }.to_not raise_error
+          end
+        end
+
+        context "if created with a non-avro data serialization" do
+          ['unknown', :unknown, 'xml', :xml].each do |it|
+            it "should throw an exception for serialization type #{it.inspect}" do
+              expect { Data.new(valid_schema, it) }.to raise_error(
+                ArgumentError, 'Only Avro serialization_type supported at the moment')
+            end
+          end
+        end
+
+        context "if created with an avro serialization" do
+          ['avro', :avro].each do |it|
+            it "should instantiate correctly for serialization type #{it.inspect}" do
+              expect { Data.new(valid_schema, it) }.to_not raise_error
+            end
+          end
+
+          context "if created with a serialized data object" do
+            it "should instantiate correctly" do
+              expect { Data.new(valid_schema, 'avro', serialized_string )}.to_not raise_error
+            end
+          end
+        end
+      end
+    end
+  end
+end

data/spec/rflow/message_spec.rb

@@ -0,0 +1,154 @@
+require 'spec_helper'
+require 'digest/md5'
+require 'rflow/message'
+
+class RFlow
+  describe Message do
+    context "if created with an unknown data type" do
+      it "should throw an exception" do
+        expect { Message.new('non_existent_data_type') }.to raise_error(
+          ArgumentError, "Data type 'non_existent_data_type' with serialization_type 'avro' not found")
+      end
+    end
+
+    context "if created with a known data type" do
+      before(:all) do
+        @schema = '{"type": "string"}'
+        Configuration.add_available_data_type(:string_type, 'avro', @schema)
+      end
+
+      it "should instantiate correctly" do
+        expect { Message.new('string_type') }.to_not raise_error
+      end
+
+      context "if created with empty provenance" do
+        context "if created with an unknown data serialization" do
+          ['unknown', :unknown].each do |it|
+            it "should throw an exception for #{it.inspect}" do
+              expect { Message.new('string_type', [], it) }.to raise_error(
+                ArgumentError, "Data type 'string_type' with serialization_type 'unknown' not found")
+            end
+          end
+        end
+
+        context "if created with a known data serialization" do
+          ['avro', :avro].each do |it|
+            it "should instantiate correctly for #{it.inspect}" do
+              expect { Message.new('string_type', [], it) }.to_not raise_error
+            end
+          end
+
+          context "if created with a mismatched schema" do
+            it
+          end
+          context "if created with a matched schema" do
+            it
+          end
+
+          context "if created with a nil schema" do
+            context "if created with a serialized data object" do
+              let(:serialized_string) { encode_avro(@schema, 'this is a string to be serialized') }
+
+              it "should instantiate correctly" do
+                expect { Message.new('string_type', [], 'avro', nil, serialized_string) }.to_not raise_error
+              end
+            end
+          end
+        end
+      end
+
+      context "if created with invalid provenance" do
+        let(:invalid_processing_event_hash) { {'started_at' => 'bad time string'} }
+        let(:invalid_provenance) { [invalid_processing_event_hash] }
+
+        it "should throw an exception" do
+          expect { Message.new('string_type', invalid_provenance) }.to raise_error(
+            ArgumentError, 'invalid date: "bad time string"')
+        end
+      end
+
+      context "if created with valid provenance" do
+        let(:valid_xmlschema_time) { '2001-01-01T01:01:01.000001Z' }
+        let(:valid_processing_event_hash) { {'component_instance_uuid' => 'uuid', 'started_at' => valid_xmlschema_time } }
+        let(:valid_processing_event) { Message::ProcessingEvent.new('uuid', valid_xmlschema_time, valid_xmlschema_time, 'context') }
+        let(:valid_provenance) do
+          [Message::ProcessingEvent.new('uuid'),
+           valid_processing_event_hash,
+           valid_processing_event]
+        end
+
+        it "should instantiate correctly" do
+          expect { Message.new('string_type', valid_provenance) }.to_not raise_error
+        end
+
+        it "should correctly set the provenance processing events" do
+          Message.new('string_type', valid_provenance).provenance[1].tap do |p|
+            p.component_instance_uuid.should == 'uuid'
+            p.started_at.should == Time.xmlschema(valid_xmlschema_time)
+            p.completed_at.should be_nil
+            p.context.should be_nil
+          end
+        end
+
+        it "should to_hash its provenance correctly" do
+          Message.new('string_type', valid_provenance).provenance.map(&:to_hash).should == [
+            {"component_instance_uuid" => "uuid", "started_at" => nil, "completed_at" => nil, "context" => nil},
+            {"component_instance_uuid" => "uuid", "started_at" => valid_xmlschema_time, "completed_at" => nil, "context" => nil},
+            {"component_instance_uuid" => "uuid", "started_at" => valid_xmlschema_time, "completed_at" => valid_xmlschema_time, "context" => "context"}]
+        end
+      end
+
+      context "if correctly created" do
+        it "should serialize and deserialize correctly to/from avro" do
+          message = Message.new('string_type').tap do |m|
+            m.provenance << Message::ProcessingEvent.new('UUID')
+            m.data.data_object = 'teh awesome'
+          end
+
+          Message.from_avro(message.to_avro).tap do |processed|
+            processed.data.to_avro.should == message.data.to_avro
+            processed.data.data_object.should == message.data.data_object
+          end
+        end
+      end
+
+      context "if data extensions exist" do
+        it "should extend the data element with the extension" do
+          module ExtensionModule; def ext_method; end; end
+
+          message = Message.new('string_type')
+          message.data.methods.should_not include(:ext_method)
+
+          Configuration.add_available_data_extension('string_type', ExtensionModule)
+          message = Message.new('string_type')
+          message.data.methods.should include(:ext_method)
+        end
+      end
+    end
+
+    it "should correctly handle large raw types" do
+      message = Message.new('RFlow::Message::Data::Raw').tap do |m|
+        m.data.raw = Array.new(101) { rand(256) }.pack('c*')
+      end
+
+      message_avro = message.to_avro.force_encoding('BINARY')
+
+      processed_message = Message.from_avro(message_avro)
+      processed_message_avro = processed_message.to_avro.force_encoding('BINARY')
+
+      @raw_schema = Configuration.available_data_types['RFlow::Message::Data::Raw']['avro']
+
+      encode_avro(@raw_schema, message.data.data_object).should == message.data.to_avro
+      decode_avro(@raw_schema, message.data.to_avro).should == message.data.data_object
+
+      message_data_avro = message.data.to_avro.force_encoding('BINARY')
+      processed_message_data_avro = processed_message.data.to_avro.force_encoding('BINARY')
+
+      Digest::MD5.hexdigest(message_avro).should == Digest::MD5.hexdigest(processed_message_avro)
+
+      message_data_avro.should == processed_message_data_avro
+      Digest::MD5.hexdigest(message_data_avro).should == Digest::MD5.hexdigest(processed_message_data_avro)
+      Digest::MD5.hexdigest(message.data.raw).should == Digest::MD5.hexdigest(processed_message.data.raw)
+    end
+  end
+end

data/spec/rflow_spec.rb

@@ -1,62 +1,51 @@
-require 'spec_helper.rb'
-
+require 'spec_helper'
 require 'open3'
 require 'rflow'
 
 describe RFlow do
-
   before(:all) do
     @extensions_file_name = File.join(File.dirname(__FILE__), 'fixtures', 'extensions_ints.rb')
   end
 
-  context "when executing from the test script" do
+  before(:each) do
+    @original_directory_path = Dir.getwd
+    @run_directory_path = File.join(@temp_directory_path, 'run')
+    @log_directory_path = File.join(@temp_directory_path, 'log')
+    Dir.mkdir @run_directory_path
+    Dir.mkdir @log_directory_path
+    Dir.chdir @temp_directory_path
+  end
 
-    before(:all) do
-      load @extensions_file_name
-    end
+  after(:each) { Dir.chdir @original_directory_path }
 
-    describe '.run' do
-      before(:each) do
-        @original_directory_path = Dir.getwd
-        @run_directory_path = File.join(@temp_directory_path, 'run')
-        @log_directory_path = File.join(@temp_directory_path, 'log')
-        Dir.mkdir @run_directory_path
-        Dir.mkdir @log_directory_path
-      end
-
-      after(:each) do
-        Dir.chdir @original_directory_path
-      end
+  context "when executing from the test script" do
+    before(:all) { load @extensions_file_name }
 
+    describe '.run!' do
       def run_rflow_with_dsl(&block)
         rflow_thread = Thread.new do
           ActiveRecord::Base.establish_connection adapter: "sqlite3", database: ":memory:"
           RFlow::Configuration.migrate_database
-          RFlow::Configuration::RubyDSL.configure do |c|
-            block.call(c)
-          end
-
+          RFlow::Configuration::RubyDSL.configure {|c| block.call(c) }
           RFlow::Configuration.merge_defaults!
-
-          RFlow.run nil, false
+          RFlow.run! nil, false
         end
 
         # TODO: figure out a way to get rid of this sleep, as there
         # should be a better way to figure out when RFlow is done
-        sleep(2)
+        sleep(5)
 
-        # Shut down the reactor and the thread
+        # Shut down the workers, the reactor, and the thread
+        RFlow.master.shutdown! 'SIGQUIT'
         EM.run { EM.stop }
         rflow_thread.join
       end
 
-
       it "should run a non-sharded workflow" do
-
         run_rflow_with_dsl do |c|
-          c.setting('rflow.log_level', 'DEBUG')
-          c.setting('rflow.application_directory_path', @temp_directory_path)
-          c.setting('rflow.application_name', 'nonsharded_test')
+          c.setting 'rflow.log_level', 'FATAL'
+          c.setting 'rflow.application_directory_path', @temp_directory_path
+          c.setting 'rflow.application_name', 'nonsharded_test'
 
           c.component 'generate_ints', 'RFlow::Components::GenerateIntegerSequence', 'start' => 20, 'finish' => 30
           c.component 'output', 'RFlow::Components::FileOutput', 'output_file_path' => 'out'
@@ -75,8 +64,8 @@ describe RFlow do
           c.connect 'generate_ints2#even_odd_out' => 'output_even_odd2#in'
         end
 
-        RFlow.master.shards.count.should == 1
-        RFlow.master.shards.first.workers.count.should == 1
+        RFlow.master.should have(1).shard
+        RFlow.master.shards.first.should have(1).worker
 
         output_files = {
           'out'           => [20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30],
@@ -88,19 +77,17 @@ describe RFlow do
         }
 
         output_files.each do |file_name, expected_contents|
-          File.exist?(File.join(@temp_directory_path, file_name)).should be_true
+          File.exist?(File.join(@temp_directory_path, file_name)).should be true
           File.readlines(file_name).map(&:to_i).should == expected_contents
         end
       end
 
-
       it "should run a sharded workflow" do
         run_rflow_with_dsl do |c|
-          c.setting('rflow.log_level', 'DEBUG')
-          c.setting('rflow.application_directory_path', @temp_directory_path)
-          c.setting('rflow.application_name', 'sharded_test')
+          c.setting 'rflow.log_level', 'FATAL'
+          c.setting 'rflow.application_directory_path', @temp_directory_path
+          c.setting 'rflow.application_name', 'sharded_test'
 
-          # Instantiate components
           c.shard 's1', :process => 3 do |s|
             s.component 'generate_ints1', 'RFlow::Components::GenerateIntegerSequence', 'start' => 0, 'finish' => 10, 'step' => 3
           end
@@ -119,7 +106,6 @@ describe RFlow do
           c.component 'output3', 'RFlow::Components::FileOutput', 'output_file_path' => 'out3'
           c.component 'output_all', 'RFlow::Components::FileOutput', 'output_file_path' => 'out_all'
 
-          # Hook components together
           c.connect 'generate_ints1#out' => 'output1#in'
           c.connect 'generate_ints2#out' => 'output2#in'
           c.connect 'generate_ints3#out' => 'output3#in'
@@ -128,7 +114,7 @@ describe RFlow do
           c.connect 'generate_ints3#out' => 'output_all#in'
         end
 
-        RFlow.master.shards.count.should == 4
+        RFlow.master.should have(4).shards
         RFlow.master.shards.map(&:count).should == [1, 3, 2, 2]
         RFlow.master.shards.map(&:workers).map(&:count).should == [1, 3, 2, 2]
 
@@ -140,7 +126,7 @@ describe RFlow do
         }
 
         output_files.each do |file_name, expected_contents|
-          File.exist?(File.join(@temp_directory_path, file_name)).should be_true
+          File.exist?(File.join(@temp_directory_path, file_name)).should be true
           File.readlines(file_name).map(&:to_i).sort.should == expected_contents.sort
         end
       end
@@ -148,29 +134,17 @@ describe RFlow do
   end
 
   context "when executing via the rflow binary" do
-    before(:each) do
-      @original_directory_path = Dir.getwd
-      @run_directory_path = File.join(@temp_directory_path, 'run')
-      @log_directory_path = File.join(@temp_directory_path, 'log')
-      Dir.mkdir @run_directory_path
-      Dir.mkdir @log_directory_path
-      Dir.chdir @temp_directory_path
-    end
-
-    after(:each) do
-      Dir.chdir @original_directory_path
-    end
-
-    def execute_rflow(rflow_args)
-      r = {}
-      r[:stdout], r[:stderr], r[:status] = Open3.capture3("bundle exec rflow #{rflow_args}")
-      r
+    def execute_rflow(args)
+      stdout, stderr, status = Open3.capture3("bundle exec rflow #{args}")
+      {:stdout => stdout, :stderr => stderr, :status => status}
     end
 
     context "with a simple ruby DSL config file" do
+      let(:config_file_name) { 'input_config' }
+      let(:db_file_name) { 'outdb' }
+
       before(:each) do
-        @config_file_name = 'input_config'
-        File.open('input_config', 'w+') do |file|
+        File.open(config_file_name, 'w+') do |file|
           file.write <<-EOF
             RFlow::Configuration::RubyDSL.configure do |c|
               c.setting 'mysetting', 'myvalue'
@@ -180,9 +154,7 @@ describe RFlow do
       end
 
       it "should load a ruby dsl file into a sqlite DB" do
-        db_file_name = 'outdb'
-
-        r = execute_rflow("load -d #{db_file_name} -c #{@config_file_name}")
+        r = execute_rflow("load -d #{db_file_name} -c #{config_file_name}")
 
         # Make sure that the process execution worked
         r[:status].exitstatus.should == 0
@@ -195,31 +167,30 @@ describe RFlow do
       end
 
       it "should not load a database if the database file already exists" do
-        db_file_name = 'outdb'
-        File.open(db_file_name, 'w') { |file| file.write 'boom' }
+        File.open(db_file_name, 'w') {|file| file.write 'boom' }
 
-        r = execute_rflow("load -d #{db_file_name} -c #{@config_file_name}")
+        r = execute_rflow("load -d #{db_file_name} -c #{config_file_name}")
 
         # Make sure that the process execution worked
         r[:status].exitstatus.should == 1
         r[:stderr].should == ''
         r[:stdout].should match /Config database.*#{db_file_name}.*exists/
       end
-
     end
 
     context "with a complex, sharded ruby DSL config file" do
+      let(:config_file_name) { 'input_config' }
+      let(:db_file_name) { 'config_db' }
+      let(:app_name) { 'sharded_bin_test' }
+
       before(:each) do
-        @config_file_name = 'input_config'
-        @db_file_name = 'config_db'
-        @app_name = 'sharded_bin_test'
-        File.open(@config_file_name, 'w+') do |file|
+        File.open(config_file_name, 'w+') do |file|
           file.write <<-EOF
             RFlow::Configuration::RubyDSL.configure do |c|
               c.setting('rflow.log_level', 'INFO')
               c.setting('rflow.application_directory_path', '#{@temp_directory_path}')
-              c.setting('rflow.application_name', '#{@app_name}')
-              # Instantiate components
+              c.setting('rflow.application_name', '#{app_name}')
+
               c.shard 's1', :process => 3 do |s|
                 s.component 'generate_ints1', 'RFlow::Components::GenerateIntegerSequence', 'start' => 0, 'finish' => 10, 'step' => 3
               end
@@ -233,7 +204,7 @@ describe RFlow do
               end
               c.component 'output3', 'RFlow::Components::FileOutput', 'output_file_path' => 'out3'
               c.component 'output_all', 'RFlow::Components::FileOutput', 'output_file_path' => 'out_all'
-              # Hook components together
+
               c.connect 'generate_ints1#out' => 'output1#in'
               c.connect 'generate_ints2#out' => 'output2#in'
               c.connect 'generate_ints3#out' => 'output3#in'
@@ -243,14 +214,14 @@ describe RFlow do
             end
           EOF
         end
-        r = execute_rflow("load -d #{@db_file_name} -c #{@config_file_name}")
+        r = execute_rflow("load -d #{db_file_name} -c #{config_file_name}")
         r[:status].exitstatus.should == 0
         r[:stderr].should == ''
-        r[:stdout].should match /Successfully initialized database.*#{@db_file_name}/
+        r[:stdout].should match /Successfully initialized database.*#{db_file_name}/
       end
 
       it "should not start if the components aren't loaded" do
-        r = execute_rflow("start -d #{@db_file_name} -f")
+        r = execute_rflow("start -d #{db_file_name} -f")
 
         r[:status].exitstatus.should == 1
         r[:stderr].should == ''
@@ -258,64 +229,63 @@ describe RFlow do
       end
 
       it "should daemonize and run in the background" do
-        r = execute_rflow("start -d #{@db_file_name} -e #{@extensions_file_name}")
+        begin
+          r = execute_rflow("start -d #{db_file_name} -e #{@extensions_file_name}")
 
-        r[:status].exitstatus.should == 0
-        r[:stderr].should == ''
-        r[:stdout].should_not match /error/i
+          r[:status].exitstatus.should == 0
+          r[:stderr].should == ''
+          r[:stdout].should_not match /error/i
 
-        sleep 1 # give the daemon a chance to finish
+          sleep 2 # give the daemon a chance to finish
 
-        log_contents = File.read("log/#{@app_name}.log").chomp
-        log_lines = log_contents.split("\n")
+          log_contents = File.read("log/#{app_name}.log").chomp
+          log_lines = log_contents.split("\n")
 
-        puts '++++++++++++++++++++'
-        puts log_contents
-        puts '++++++++++++++++++++'
+          log_lines.each {|line| line.should_not match /^ERROR/ }
+          log_lines.each {|line| line.should_not match /^DEBUG/ }
 
-        # Log file testing
-        log_lines.each { |line| line.should_not match /^ERROR/ }
-        log_lines.each { |line| line.should_not match /^DEBUG/ }
+          # Grab all the pids from the log, which seems to be the only
+          # reliable way to get them
+          log_pids = log_lines.map {|line| /\((\d+)\)/.match(line)[1].to_i }.uniq
 
-        # Grab all the pids from the log, which seems to be the only
-        # reliable way to get them
-        log_pids = log_lines.map { |line| /\((\d+)\)/.match(line)[1].to_i }.uniq
+          initial_pid = r[:status].pid
+          master_pid = File.read("run/#{app_name}.pid").chomp.to_i
+          worker_pids = log_pids - [initial_pid, master_pid]
 
-        initial_pid = r[:status].pid
-        master_pid = File.read("run/#{@app_name}.pid").chomp.to_i
-        worker_pids = log_pids - [initial_pid, master_pid]
+          log_pids.should include initial_pid
+          log_pids.should include master_pid
 
-        log_pids.should include initial_pid
-        log_pids.should include master_pid
+          worker_pids.should have(8).pids
+          worker_pids.should_not include 0
 
-        worker_pids.size.should == 8
-        worker_pids.should_not include 0
+          expect { Process.kill(0, initial_pid) }.to raise_error(Errno::ESRCH)
+          ([master_pid] + worker_pids).each do |pid|
+            Process.kill(0, pid).should == 1
+          end
 
-        # Process checks
-        expect { Process.kill(0, initial_pid) }.to raise_error(Errno::ESRCH)
-        ([master_pid] + worker_pids).each do |pid|
-          Process.kill(0, pid).should == 1
-        end
+          output_files = {
+            'out1'    => [0, 3, 6, 9] * 3,
+            'out2'    => (20..30).to_a * 2,
+            'out3'    => (100..105).to_a,
+            'out_all' => [0, 3, 6, 9] * 3 + (20..30).to_a * 2 + (100..105).to_a
+          }
 
-        # Output checks
-        output_files = {
-          'out1'    => [0, 3, 6, 9] * 3,
-          'out2'    => (20..30).to_a * 2,
-          'out3'    => (100..105).to_a,
-          'out_all' => [0, 3, 6, 9] * 3 + (20..30).to_a * 2 + (100..105).to_a
-        }
-
-        output_files.each do |file_name, expected_contents|
-          File.exist?(File.join(@temp_directory_path, file_name)).should be_true
-          File.readlines(file_name).map(&:to_i).sort.should == expected_contents.sort
-        end
+          output_files.each do |file_name, expected_contents|
+            File.exist?(File.join(@temp_directory_path, file_name)).should be true
+            File.readlines(file_name).map(&:to_i).sort.should == expected_contents.sort
+          end
 
-        # Terminate the master
-        Process.kill("TERM", master_pid).should == 1
+          # Terminate the master
+          Process.kill("TERM", master_pid).should == 1
 
-        # Make sure everything is dead
-        ([master_pid] + worker_pids).each do |pid|
-          expect { Process.kill(0, pid) }.to raise_error(Errno::ESRCH)
+          # Make sure everything is dead after a second
+          sleep 1
+          ([master_pid] + worker_pids).each do |pid|
+            expect { Process.kill(0, pid) }.to raise_error(Errno::ESRCH)
+          end
+        rescue Exception => e
+          Process.kill("TERM", master_pid) if master_pid
+          raise
         end
       end
     end