mrtoolkit 0.1.4 → 0.1.5

data/VERSION.yml

@@ -1,5 +1,5 @@
 --- 
 :major: 0
 :minor: 1
-:patch: 4
+:patch: 5
 :build: 

data/lib/mrtoolkit.rb

@@ -779,10 +779,10 @@ class JobBase
   # This gathers the declarations and stores in a stage record.
   def add_stage
     case
-    when @map_class.nil?: raise "Map class not specified"
-    when @reduce_class.nil?: raise "Reduce class not specified"
-    when @in_dirs.empty?: raise "Indir not speficied"
-    when @out_dir.nil?: raise "Outdir not specified"
+    when @map_class.nil?    then raise "Map class not specified"
+    when @reduce_class.nil? then raise "Reduce class not specified"
+    when @in_dirs.empty?    then raise "Indir not speficied"
+    when @out_dir.nil?      then raise "Outdir not specified"
     end       
     @stages << [@map_class, @map_args, @map_opts,
       @reduce_class, @reduce_args, @reduce_opts, 

data/test/utest.rb

@@ -15,7 +15,7 @@ class LogMap < MapBase
     field :date
     field :time
     field :url
-  
+
     emit :date_time
     emit :url
   end
@@ -25,7 +25,7 @@ class LogMap < MapBase
     output.url = input.url
     output
   end
-  
+
 end
 
 class LogReduce < ReduceBase
@@ -58,7 +58,7 @@ class TestMRToolkit < Test::Unit::TestCase
   def test_log
     LogJob.run_command
     out = File.read("test-out")
-    expected = "2008-10-01T10:30:00\t1.2.3.4\tx\n" + 
+    expected = "2008-10-01T10:30:00\t1.2.3.4\tx\n" +
                "2008-10-02T11:30:00\t1.2.3.5\tx\n"
     assert_equal(expected, out)
   end
@@ -72,7 +72,7 @@ end
 class SumMap < MapBase
   def declare
     field :value
-  
+
     emit :count
     emit :total
     emit :sum_of_squares
@@ -85,7 +85,7 @@ class SumMap < MapBase
     output.sum_of_squares = v * v
     output
   end
-  
+
 end
 
 # This could be done with canned reducer
@@ -144,26 +144,25 @@ end
 # Grops times into one-minute buckets
 # Calculates counts for each bucket
 
-require 'parsedate'
+require 'time'
 
 class MinMap < MapBase
   def declare
     field :dt
     field :tm
-  
+
     emit :minute
     emit :count
   end
 
   def process(input, output)
-    res = ParseDate.parsedate(input.dt + " " + input.tm)
-    t = Time.local(*res)
+    t = Time.parse(input.dt + " " + input.tm)
     min = t.min + 60 * (t.hour + 24 * t.wday)
     output.count = 1
     output.minute = min
     output
   end
-  
+
 end
 
 class MyMinReduce < ReduceBase
@@ -211,7 +210,7 @@ class TestMRToolkit < Test::Unit::TestCase
 end
 
 #################################
-# 
+#
 # This is the previous one, but with a standard reducer.
 
 class CollectJob < JobBase
@@ -237,7 +236,7 @@ class TestMRToolkit < Test::Unit::TestCase
 end
 
 #################################
-# 
+#
 # This is the previous one, but with adifferent
 # standard reducer.  This produces the same output
 # as the custom reducer.

metadata

@@ -2,7 +2,7 @@
 name: mrtoolkit
 version: !ruby/object:Gem::Version 
   prerelease: 
-  version: 0.1.4
+  version: 0.1.5
 platform: ruby
 authors: 
 - cchayden
@@ -12,8 +12,7 @@ autorequire:
 bindir: bin
 cert_chain: []
 
-date: 2011-05-16 00:00:00 -07:00
-default_executable: 
+date: 2011-07-06 00:00:00 Z
 dependencies: []
 
 description: 
@@ -57,7 +56,6 @@ files:
 - test/test-in/test8-in
 - test/test-in/test9-in
 - test/utest.rb
-has_rdoc: true
 homepage: http://github.com/jashmenn/mrtoolkit
 licenses: []
 
@@ -81,9 +79,17 @@ required_rubygems_version: !ruby/object:Gem::Requirement
 requirements: []
 
 rubyforge_project: 
-rubygems_version: 1.6.2
+rubygems_version: 1.8.5
 signing_key: 
 specification_version: 3
 summary: Simplify the creation of Hadoop Map/Reduce jobs
-test_files: []
-
+test_files: 
+- examples/hour.rb
+- examples/import.rb
+- examples/ip-result.rb
+- examples/ip-size.rb
+- examples/ip-ua.rb
+- examples/ip.rb
+- examples/section.rb
+- examples/top-file.rb
+- test/utest.rb