mandy 0.5.3 → 0.5.5

data/VERSION

@@ -1 +1 @@
-0.5.3
+0.5.5

data/bin/mandy-hadoop

@@ -11,6 +11,10 @@ options = OpenStruct.new
 OptionParser.new do |opts|
   opts.banner = "USAGE: mandy-hadoop script input output [options]"
 
+  opts.on("-i", "--input-format class", "Set the hadoop input format type") do |format_class|
+    options.input_format = format_class
+  end
+
   opts.on("-p", "--payload PAYLOAD", "Add a working directory to be sent to the cluster.") do |payload|
     options.payload = payload
   end
@@ -68,6 +72,8 @@ payload = Mandy::Packer.pack(file, options.payload || ARGV[0], gemfile(options.g
 cmdenv = options.cmdenv
 set_env(cmdenv)
 
+inputformat = !options.input_format.nil? ? options.input_format : "TextInputFormat"
+
 at_exit do
   puts
   puts "Cleaning up..."
@@ -93,13 +99,14 @@ begin
     command = %($HADOOP_HOME/bin/hadoop jar $HADOOP_HOME/contrib/streaming/hadoop-*-streaming.jar #{jobconf}\
                     -files "#{payload}","#{bootstrap_file}" \
                     -conf '#{config}' \
+                    -inputformat '#{inputformat}' \
                     #{input} \
                     #{ inputreader.nil? ? '' : "-inputreader \"#{inputreader}\"" } \
                     -mapper "ruby bootstrap.rb #{File.basename(payload)} map #{filename} '#{job.name}'"  \
                     -reducer "ruby bootstrap.rb #{File.basename(payload)} reduce #{filename} '#{job.name}'"  \
                     #{ cmdenv.nil? ? '' : "-cmdenv #{cmdenv}" }\
                     -output "#{output}" 2>&1)
-
+                    
     result = []
     IO.popen(command, 'r') do |subprocess|
       while line = subprocess.gets

data/lib/mandy/packer.rb

@@ -22,8 +22,8 @@ module Mandy
     end
     
     def self.cleanup!(file)
-      tmp_dir = File.dirname(file)
-      FileUtils.rm_r(tmp_dir)
+      return false unless File.extname(file) == '.tar'
+      `rm #{file}`
     end
   end
 end

metadata

@@ -5,8 +5,8 @@ version: !ruby/object:Gem::Version
   segments: 
   - 0
   - 5
-  - 3
-  version: 0.5.3
+  - 5
+  version: 0.5.5
 platform: ruby
 authors: 
 - Andy Kent
@@ -15,7 +15,7 @@ autorequire:
 bindir: bin
 cert_chain: []
 
-date: 2009-10-19 00:00:00 +01:00
+date: 2010-08-20 00:00:00 +01:00
 default_executable: 
 dependencies: 
 - !ruby/object:Gem::Dependency