mandy 0.3.6 → 0.3.7
This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
- data/lib/job.rb +42 -12
- data/lib/mappers/base_mapper.rb +5 -1
- data/lib/reducers/base_reducer.rb +5 -1
- metadata +1 -1
data/lib/job.rb
CHANGED
@@ -17,8 +17,6 @@ module Mandy
|
|
17
17
|
@name = name
|
18
18
|
@settings = {}
|
19
19
|
@modules = []
|
20
|
-
@mapper_class = Mandy::Mappers::PassThroughMapper
|
21
|
-
@reducer_class = Mandy::Reducers::PassThroughReducer
|
22
20
|
set('mapred.job.name', name)
|
23
21
|
instance_eval(&blk) if blk
|
24
22
|
end
|
@@ -57,35 +55,67 @@ module Mandy
|
|
57
55
|
end
|
58
56
|
end
|
59
57
|
|
58
|
+
def setup(&blk)
|
59
|
+
@setup = blk
|
60
|
+
end
|
61
|
+
|
62
|
+
def teardown(&blk)
|
63
|
+
@teardown = blk
|
64
|
+
end
|
65
|
+
|
60
66
|
def map(klass=nil, &blk)
|
61
|
-
@
|
62
|
-
@modules.each {|m| @mapper_class.send(:include, m) }
|
63
|
-
@mapper_class
|
67
|
+
@map = klass || blk
|
64
68
|
end
|
65
69
|
|
66
70
|
def reduce(klass=nil, &blk)
|
67
|
-
@
|
68
|
-
@modules.each {|m| @reducer_class.send(:include, m) }
|
69
|
-
@reducer_class
|
71
|
+
@reduce = klass || blk
|
70
72
|
end
|
71
73
|
|
72
74
|
def run_map(input=STDIN, output=STDOUT, &blk)
|
73
|
-
|
74
|
-
mapper =
|
75
|
+
mapper_class.send(:include, Mandy::IO::OutputFormatting) unless reducer_defined?
|
76
|
+
mapper = mapper_class.new(input, output, @input_format, @output_format)
|
75
77
|
yield(mapper) if blk
|
76
78
|
mapper.execute
|
77
79
|
end
|
78
80
|
|
79
81
|
def run_reduce(input=STDIN, output=STDOUT, &blk)
|
80
|
-
reducer =
|
82
|
+
reducer = reducer_class.new(input, output, @input_format, @output_format)
|
81
83
|
yield(reducer) if blk
|
82
84
|
reducer.execute
|
83
85
|
end
|
84
86
|
|
85
87
|
private
|
88
|
+
|
89
|
+
def mapper_class
|
90
|
+
return Mandy::Mappers::PassThroughMapper unless @map
|
91
|
+
@mapper_class ||= compile_map
|
92
|
+
end
|
86
93
|
|
94
|
+
def reducer_class
|
95
|
+
return Mandy::Reducers::PassThroughReducer unless @reduce
|
96
|
+
@reducer_class ||= compile_reduce
|
97
|
+
end
|
98
|
+
|
99
|
+
def compile_map
|
100
|
+
args = {}
|
101
|
+
args[:setup] = @setup if @setup
|
102
|
+
args[:teardown] = @teardown if @teardown
|
103
|
+
@mapper_class = @map.is_a?(Proc) ? Mandy::Mappers::Base.compile(args, &@map) : @map
|
104
|
+
@modules.each {|m| @mapper_class.send(:include, m) }
|
105
|
+
@mapper_class
|
106
|
+
end
|
107
|
+
|
108
|
+
def compile_reduce
|
109
|
+
args = {}
|
110
|
+
args[:setup] = @setup if @setup
|
111
|
+
args[:teardown] = @teardown if @teardown
|
112
|
+
@reducer_class = @reduce.is_a?(Proc) ? Mandy::Reducers::Base.compile(args, &@reduce) : @reduce
|
113
|
+
@modules.each {|m| @reducer_class.send(:include, m) }
|
114
|
+
@reducer_class
|
115
|
+
end
|
116
|
+
|
87
117
|
def reducer_defined?
|
88
|
-
|
118
|
+
reducer_class != Mandy::Reducers::PassThroughReducer
|
89
119
|
end
|
90
120
|
|
91
121
|
end
|
data/lib/mappers/base_mapper.rb
CHANGED
@@ -3,15 +3,18 @@ module Mandy
|
|
3
3
|
class Base < Mandy::Task
|
4
4
|
include Mandy::IO::InputFormatting
|
5
5
|
|
6
|
-
def self.compile(&blk)
|
6
|
+
def self.compile(opts={}, &blk)
|
7
7
|
Class.new(Mandy::Mappers::Base) do
|
8
8
|
self.class_eval do
|
9
9
|
define_method(:mapper, blk) if blk
|
10
|
+
define_method(:setup, opts[:setup]) if opts[:setup]
|
11
|
+
define_method(:teardown, opts[:teardown]) if opts[:teardown]
|
10
12
|
end
|
11
13
|
end
|
12
14
|
end
|
13
15
|
|
14
16
|
def execute
|
17
|
+
setup if self.respond_to?(:setup)
|
15
18
|
@input.each_line do |line|
|
16
19
|
key, value = line.split(KEY_VALUE_SEPERATOR, 2)
|
17
20
|
key, value = nil, key if value.nil?
|
@@ -19,6 +22,7 @@ module Mandy
|
|
19
22
|
args = [input_deserialize_key(key), input_deserialize_value(value)].compact
|
20
23
|
mapper(*args)
|
21
24
|
end
|
25
|
+
teardown if self.respond_to?(:teardown)
|
22
26
|
end
|
23
27
|
|
24
28
|
private
|
@@ -3,15 +3,18 @@ module Mandy
|
|
3
3
|
class Base < Mandy::Task
|
4
4
|
include Mandy::IO::OutputFormatting
|
5
5
|
|
6
|
-
def self.compile(&blk)
|
6
|
+
def self.compile(opts={}, &blk)
|
7
7
|
Class.new(Mandy::Reducers::Base) do
|
8
8
|
self.class_eval do
|
9
9
|
define_method(:reducer, blk) if blk
|
10
|
+
define_method(:setup, opts[:setup]) if opts[:setup]
|
11
|
+
define_method(:teardown, opts[:teardown]) if opts[:teardown]
|
10
12
|
end
|
11
13
|
end
|
12
14
|
end
|
13
15
|
|
14
16
|
def execute
|
17
|
+
setup if self.respond_to?(:setup)
|
15
18
|
last_key, values = nil, []
|
16
19
|
@input.each_line do |line|
|
17
20
|
key, value = line.split(KEY_VALUE_SEPERATOR)
|
@@ -24,6 +27,7 @@ module Mandy
|
|
24
27
|
values << value
|
25
28
|
end
|
26
29
|
reducer(deserialize_key(last_key), values.map {|v| deserialize_value(v) })
|
30
|
+
teardown if self.respond_to?(:teardown)
|
27
31
|
end
|
28
32
|
|
29
33
|
private
|