embulk 0.6.27 → 0.7.0

data/lib/embulk/data/bundle/.ruby-version

@@ -0,0 +1 @@
+jruby-9.0.0.0

data/lib/embulk/data/bundle/Gemfile

@@ -1,8 +1,11 @@
 source 'https://rubygems.org/'
+gem 'embulk', '~> 0.7.0'
+
 #
 # Note: After updating this file, run this command to update Gemfile.lock:
 #
-#       embulk bundle <this_dir>
+#       cd this_directory
+#       embulk bundle
 #
 
 #

data/lib/embulk/data/bundle/embulk/input/example.rb

@@ -24,8 +24,8 @@ module Embulk
 
       def self.resume(task, columns, count, &control)
         puts "Example input started."
-        commit_reports = yield(task, columns, count)
-        puts "Example input finished. Commit reports = #{commit_reports.to_json}"
+        task_reports = yield(task, columns, count)
+        puts "Example input finished. Commit reports = #{task_reports.to_json}"
 
         next_config_diff = {}
         return next_config_diff
@@ -45,8 +45,8 @@ module Embulk
         end
         @page_builder.finish  # don't forget to call finish :-)
 
-        commit_report = {}
-        return commit_report
+        task_report = {}
+        return task_report
       end
     end
 

data/lib/embulk/data/bundle/embulk/output/example.rb

@@ -15,8 +15,8 @@ module Embulk
 
       def self.resume(task, schema, count, &control)
         puts "Example output started."
-        commit_reports = yield(task)
-        puts "Example output finished. Commit reports = #{commit_reports.to_json}"
+        task_reports = yield(task)
+        puts "Example output finished. Commit reports = #{task_reports.to_json}"
 
         next_config_diff = {}
         return next_config_diff
@@ -47,10 +47,10 @@ module Embulk
       end
 
       def commit
-        commit_report = {
+        task_report = {
           "records" => @records
         }
-        return commit_report
+        return task_report
       end
     end
 

data/lib/embulk/data/new/README.md.erb

@@ -106,6 +106,6 @@ $ embulk guess -g <%= name %> config.yml -o guessed.yml
 %when :ruby
 $ rake
 %when :java
-$ ./gradlew gem
+$ ./gradlew gem  # -t to watch change of files and rebuild continuously
 % end
 ```

data/lib/embulk/data/new/java/file_input.java.erb

@@ -4,7 +4,7 @@ import java.util.List;
 import java.util.ArrayList;
 import com.google.common.base.Optional;
 import com.google.common.collect.ImmutableList;
-import org.embulk.config.CommitReport;
+import org.embulk.config.TaskReport;
 import org.embulk.config.Config;
 import org.embulk.config.ConfigDefault;
 import org.embulk.config.ConfigInject;
@@ -107,7 +107,7 @@ public class <%= java_class_name %>
     @Override
     public void cleanup(TaskSource taskSource,
             int taskCount,
-            List<CommitReport> successCommitReports)
+            List<TaskReport> successTaskReports)
     {
     }
 
@@ -129,9 +129,9 @@ public class <%= java_class_name %>
         //    { }
         //
         //    @Override
-        //    public CommitReport commit()
+        //    public TaskReport commit()
         //    {
-        //        return Exec.newCommitReport();
+        //        return Exec.newTaskReport();
         //    }
         //}
     }

data/lib/embulk/data/new/java/file_output.java.erb

@@ -2,7 +2,7 @@ package <%= java_package_name %>;
 
 import java.util.List;
 import com.google.common.base.Optional;
-import org.embulk.config.CommitReport;
+import org.embulk.config.TaskReport;
 import org.embulk.config.Config;
 import org.embulk.config.ConfigDefault;
 import org.embulk.config.ConfigDiff;
@@ -75,7 +75,7 @@ public class <%= java_class_name %>
     @Override
     public void cleanup(TaskSource taskSource,
             int taskCount,
-            List<CommitReport> successCommitReports)
+            List<TaskReport> successTaskReports)
     {
     }
 

data/lib/embulk/data/new/java/gradle/wrapper/gradle-wrapper.properties

@@ -1,6 +1,6 @@
-#Wed Feb 04 13:46:12 PST 2015
+#Tue Aug 11 00:26:20 PDT 2015
 distributionBase=GRADLE_USER_HOME
 distributionPath=wrapper/dists
 zipStoreBase=GRADLE_USER_HOME
 zipStorePath=wrapper/dists
-distributionUrl=https\://services.gradle.org/distributions/gradle-2.2.1-bin.zip
+distributionUrl=https\://services.gradle.org/distributions/gradle-2.6-bin.zip

data/lib/embulk/data/new/java/input.java.erb

@@ -2,7 +2,7 @@ package <%= java_package_name %>;
 
 import java.util.List;
 import com.google.common.base.Optional;
-import org.embulk.config.CommitReport;
+import org.embulk.config.TaskReport;
 import org.embulk.config.Config;
 import org.embulk.config.ConfigDefault;
 import org.embulk.config.ConfigDiff;
@@ -64,12 +64,12 @@ public class <%= java_class_name %>
     @Override
     public void cleanup(TaskSource taskSource,
             Schema schema, int taskCount,
-            List<CommitReport> successCommitReports)
+            List<TaskReport> successTaskReports)
     {
     }
 
     @Override
-    public CommitReport run(TaskSource taskSource,
+    public TaskReport run(TaskSource taskSource,
             Schema schema, int taskIndex,
             PageOutput output)
     {

data/lib/embulk/data/new/java/output.java.erb

@@ -2,7 +2,7 @@ package <%= java_package_name %>;
 
 import java.util.List;
 import com.google.common.base.Optional;
-import org.embulk.config.CommitReport;
+import org.embulk.config.TaskReport;
 import org.embulk.config.Config;
 import org.embulk.config.ConfigDefault;
 import org.embulk.config.ConfigDiff;
@@ -62,7 +62,7 @@ public class <%= java_class_name %>
     @Override
     public void cleanup(TaskSource taskSource,
             Schema schema, int taskCount,
-            List<CommitReport> successCommitReports)
+            List<TaskReport> successTaskReports)
     {
     }
 

data/lib/embulk/data/new/ruby/.ruby-version

@@ -0,0 +1 @@
+jruby-9.0.0.0

data/lib/embulk/data/new/ruby/gemspec.erb

@@ -14,6 +14,7 @@ Gem::Specification.new do |spec|
   spec.require_paths = ["lib"]
 
   #spec.add_dependency 'YOUR_GEM_DEPENDENCY', ['~> YOUR_GEM_DEPENDENCY_VERSION']
+  spec.add_development_dependency 'embulk', ['~> <%= Embulk::VERSION %>']
   spec.add_development_dependency 'bundler', ['~> 1.0']
   spec.add_development_dependency 'rake', ['>= 10.0']
 end

data/lib/embulk/data/new/ruby/input.rb.erb

@@ -22,7 +22,7 @@ module Embulk
       end
 
       def self.resume(task, columns, count, &control)
-        commit_reports = yield(task, columns, count)
+        task_reports = yield(task, columns, count)
 
         next_config_diff = {}
         return next_config_diff
@@ -50,8 +50,8 @@ module Embulk
         page_builder.add(["example-value", 2, 0.2])
         page_builder.finish
 
-        commit_report = {}
-        return commit_report
+        task_report = {}
+        return task_report
       end
     end
 

data/lib/embulk/data/new/ruby/output.rb.erb

@@ -16,13 +16,13 @@ module Embulk
         # resume(task, schema, count, &control)
 
         # non-resumable output:
-        commit_reports = yield(task)
+        task_reports = yield(task)
         next_config_diff = {}
         return next_config_diff
       end
 
       #def self.resume(task, schema, count, &control)
-      #  commit_reports = yield(task)
+      #  task_reports = yield(task)
       #
       #  next_config_diff = {}
       #  return next_config_diff
@@ -52,8 +52,8 @@ module Embulk
       end
 
       def commit
-        commit_report = {}
-        return commit_report
+        task_report = {}
+        return task_report
       end
     end
 

data/lib/embulk/data/package_data.rb

@@ -1,16 +1,6 @@
 module Embulk
 
   class PackageData
-    if __FILE__ =~ /^classpath:/ || __FILE__.include?('!/')
-      # data is in embulk-core jar
-      resource_class = org.embulk.command.Runner.java_class
-      JAVA_RESOURCE = true
-      RESOURCE_URL = resource_class.resource("/embulk/data")
-    else
-      JAVA_RESOURCE = false
-      FILE_BASE_PATH = File.join(Embulk.home('lib'), 'embulk', 'data')
-    end
-
     def initialize(base_name, dest_dir, erb_binding=nil)
       require 'fileutils'
       @base_name = base_name
@@ -19,11 +9,7 @@ module Embulk
     end
 
     def path(src)
-      if JAVA_RESOURCE
-        "#{RESOURCE_URL}/#{@base_name}/#{src}"
-      else
-        File.join(FILE_BASE_PATH, @base_name, src)
-      end
+      Embulk.lib_path("embulk/data/#{@base_name}/#{src}")
     end
 
     def content(src)

data/lib/embulk/guess/charset.rb

@@ -16,7 +16,16 @@ module Embulk
 
       def guess(config, sample_buffer)
         # ICU4J
-        detector = com.ibm.icu.text.CharsetDetector.new
+        begin
+          detector_class = com.ibm.icu.text.CharsetDetector
+        rescue NameError
+          # icu4j is removed from embulk.gem package explicitly at embulk.gemspec
+          # if gem is packaged for JRuby to reduce binary size. Instead, if it's
+          # packaged for JRuby, embulk.gemspec adds rjack-icu to its dependency.
+          require 'rjack-icu'
+          detector_class = com.ibm.icu.text.CharsetDetector
+        end
+        detector = detector_class.new
         detector.setText(sample_buffer.to_java_bytes)
         best_match = detector.detect
         if best_match.getConfidence < 50

data/lib/embulk/input_plugin.rb

@@ -13,7 +13,7 @@ module Embulk
       raise NotImplementedError, "#{self}.resume(task, columns, count, &control) is not implemented. This plugin is not resumable"
     end
 
-    def self.cleanup(task, schema, count, commit_reports)
+    def self.cleanup(task, schema, count, task_reports)
       # do nothing by default
     end
 
@@ -55,9 +55,9 @@ module Embulk
           config_diff_hash = @ruby_class.transaction(config) do |task_source_hash,columns,task_count|
             java_task_source = DataSource.from_ruby_hash(task_source_hash).to_java
             java_schema = Schema.new(columns).to_java
-            java_commit_reports = java_control.run(java_task_source, java_schema, task_count)
-            java_commit_reports.map {|java_commit_report|
-              DataSource.from_java(java_commit_report)
+            java_task_reports = java_control.run(java_task_source, java_schema, task_count)
+            java_task_reports.map {|java_task_report|
+              DataSource.from_java(java_task_report)
             }
           end
           # TODO check return type of #transaction
@@ -70,20 +70,20 @@ module Embulk
           config_diff_hash = @ruby_class.resume(task_source, schema, task_count) do |task_source_hash,columns,task_count|
             java_task_source = DataSource.from_ruby_hash(task_source_hash).to_java
             java_schema = Schema.new(columns).to_java
-            java_commit_reports = java_control.run(java_task_source, java_schema, task_count)
-            java_commit_reports.map {|java_commit_report|
-              DataSource.from_java(java_commit_report)
+            java_task_reports = java_control.run(java_task_source, java_schema, task_count)
+            java_task_reports.map {|java_task_report|
+              DataSource.from_java(java_task_report)
             }
           end
           # TODO check return type of #resume
           return DataSource.from_ruby_hash(config_diff_hash).to_java
         end
 
-        def cleanup(java_task_source, java_schema, task_count, java_commit_reports)
+        def cleanup(java_task_source, java_schema, task_count, java_task_reports)
           task_source = DataSource.from_java(java_task_source)
           schema = Schema.from_java(java_schema)
-          commit_reports = java_commit_reports.map {|c| DataSource.from_java(c) }
-          @ruby_class.cleanup(task_source, schema, task_count, commit_reports)
+          task_reports = java_task_reports.map {|c| DataSource.from_java(c) }
+          @ruby_class.cleanup(task_source, schema, task_count, task_reports)
           return nil
         end
 
@@ -98,8 +98,8 @@ module Embulk
           schema = Schema.from_java(java_schema)
           page_builder = PageBuilder.new(schema, java_output)
           begin
-            commit_report_hash = @ruby_class.new(task_source, schema, processor_index, page_builder).run
-            return DataSource.from_ruby_hash(commit_report_hash).to_java
+            task_report_hash = @ruby_class.new(task_source, schema, processor_index, page_builder).run
+            return DataSource.from_ruby_hash(task_report_hash).to_java
           ensure
             page_builder.close
           end

data/lib/embulk/java/bootstrap.rb

@@ -1,11 +1,7 @@
 module Embulk
   module Java
-    #require 'embulk/gems'
-    #Embulk.add_embedded_gem_path
-
     require 'embulk/java/imports'
     require 'embulk/java/time_helper'
-    require 'embulk/java/liquid_helper'
 
     module Injected
       # Following constats are set by org.embulk.jruby.JRubyScriptingModule:
@@ -17,5 +13,12 @@ module Embulk
     def self.injector
       Injected::Injector
     end
+
+    require 'embulk'
+    require 'embulk/error'
+    require 'embulk/buffer'
+    require 'embulk/data_source'
+    require 'embulk/plugin'
+    require 'embulk/runner'
   end
 end

data/lib/embulk/java/imports.rb

@@ -5,8 +5,41 @@ require 'java'
 #
 
 module Embulk::Java
-  java_import 'org.embulk.spi.Buffer'
+  module Config
+    include_package 'org.embulk.config'
+  end
+
+  module Exec
+    include_package 'org.embulk.exec'
+  end
+
+  module Plugin
+    include_package 'org.embulk.plugin'
+  end
+
+  module SPI
+    module Time
+      include_package 'org.embulk.time'
+    end
+
+    module Type
+      include_package 'org.embulk.type'
+    end
 
+    module Unit
+      include_package 'org.embulk.unit'
+    end
+
+    module Util
+      include_package 'org.embulk.util'
+    end
+
+    include_package 'org.embulk.spi'
+  end
+
+  include_package 'org.embulk'
+
+  java_import 'org.embulk.spi.Buffer'
   java_import 'org.embulk.config.DataSourceImpl'
   java_import 'org.embulk.spi.time.Timestamp'
   java_import 'org.embulk.spi.time.TimestampParseException'
@@ -30,7 +63,6 @@ module Embulk::Java
   java_import 'org.embulk.spi.type.Types'
   java_import 'org.embulk.spi.FileInputRunner'
   java_import 'org.embulk.spi.FileOutputRunner'
-  java_import 'org.embulk.spi.Exec'
   java_import 'org.embulk.plugin.PluginClassLoaderFactory'
 
   # TODO

data/lib/embulk/logger.rb

@@ -1,5 +1,7 @@
 
 module Embulk
+  # this file is required before loading embulk-core.jar
+
   require 'logger'
 
   class Logger

data/lib/embulk/output_plugin.rb

@@ -14,7 +14,7 @@ module Embulk
       raise NotImplementedError, "#{self}.resume(task, schema, count, &control) is not implemented. This plugin is not resumable"
     end
 
-    def self.cleanup(task, schema, count, commit_reports)
+    def self.cleanup(task, schema, count, task_reports)
       # do nothing by default
     end
 
@@ -64,9 +64,9 @@ module Embulk
           schema = Schema.from_java(java_schema)
           config_diff_hash = @ruby_class.transaction(config, schema, task_count) do |task_source_hash|
             java_task_source = DataSource.from_ruby_hash(task_source_hash).to_java
-            java_commit_reports = java_control.run(java_task_source)
-            java_commit_reports.map {|java_commit_report|
-              DataSource.from_java(java_commit_report)
+            java_task_reports = java_control.run(java_task_source)
+            java_task_reports.map {|java_task_report|
+              DataSource.from_java(java_task_report)
             }
           end
           # TODO check return type of #transaction
@@ -78,20 +78,20 @@ module Embulk
           schema = Schema.from_java(java_schema)
           config_diff_hash = @ruby_class.resume(task_source, schema, task_count) do |task_source_hash,columns,task_count|
             java_task_source = DataSource.from_ruby_hash(task_source_hash).to_java
-            java_commit_reports = java_control.run(java_task_source)
-            java_commit_reports.map {|java_commit_report|
-              DataSource.from_java(java_commit_report)
+            java_task_reports = java_control.run(java_task_source)
+            java_task_reports.map {|java_task_report|
+              DataSource.from_java(java_task_report)
             }
           end
           # TODO check return type of #resume
           return DataSource.from_ruby_hash(config_diff_hash).to_java
         end
 
-        def cleanup(java_task_source, java_schema, task_count, java_commit_reports)
+        def cleanup(java_task_source, java_schema, task_count, java_task_reports)
           task_source = DataSource.from_java(java_task_source)
           schema = Schema.from_java(java_schema)
-          commit_reports = java_commit_reports.map {|c| DataSource.from_java(c) }
-          @ruby_class.cleanup(task_source, schema, task_count, commit_reports)
+          task_reports = java_task_reports.map {|c| DataSource.from_java(c) }
+          @ruby_class.cleanup(task_source, schema, task_count, task_reports)
           return nil
         end
 
@@ -128,8 +128,8 @@ module Embulk
           end
 
           def commit
-            commit_report_hash = @ruby_object.commit
-            return DataSource.from_ruby_hash(commit_report_hash).to_java
+            task_report_hash = @ruby_object.commit
+            return DataSource.from_ruby_hash(task_report_hash).to_java
           end
         end
       end