rawk 0.1.0

data/bin/rawk

@@ -0,0 +1,11 @@
+#!/usr/bin/env ruby
+
+PROJECT_DIR=File.expand_path(File.join(File.dirname(__FILE__), '..'))
+$: << File.join(PROJECT_DIR, 'lib')
+
+require 'rawk/rawk'
+
+DATA = STDIN
+DSL_CODE = ARGV[0]
+
+Rawk::Program.new(DATA).run(DSL_CODE)

data/lib/rawk/rawk.rb

@@ -0,0 +1,94 @@
+require 'set'
+require 'observer'
+
+module Rawk  
+  class Program    
+    attr_reader :fs
+    
+    def initialize(io)
+      @start, @every, @finish = Set.new, Set.new, Set.new
+      @input_stream = InputStream.new(io)
+      @input_stream.add_observer(self)
+      initialize_builtins!
+    end
+    
+    private
+    def initialize_builtins!
+      @fs = " "
+      @nr = 0
+    end
+    
+    public
+    def on_new_line
+      @nr += 1
+    end
+    alias :update :on_new_line
+    
+    def run(code = "", &block)
+      load!(code, &block)
+      execute_code!
+    end
+    
+    # DSL 
+    def start(&block)
+      @start << block
+    end
+    
+    def every(&block)
+      @every << block
+    end
+    
+    def finish(&block)
+      @finish << block
+    end
+    
+    private
+    def load!(code, &block)
+      if code.empty?
+        instance_eval(&block) 
+      else
+        instance_eval(code)
+      end
+    end
+    
+    def execute_code!
+      @start.each {|b| b.call}
+      @input_stream.each_line do |row|
+        @every.each {|b| b.call(Line.new(row, @fs))}
+      end
+      @finish.each {|b| b.call}
+    end
+  end
+  
+  class InputStream
+    include Observable
+    
+    def initialize(io)
+      @io = io
+    end
+    
+    def each_line
+      @io.each_line do |line| 
+        changed
+        notify_observers
+        yield line
+      end
+    end
+  end
+  
+  class Line < String
+    def initialize(str, fs)
+      self.replace(str.chomp)
+      @fs = fs
+    end
+    
+    def cols
+      split(@fs)
+    end
+    alias :c :cols
+    
+    def nf 
+      cols.length
+    end
+  end
+end

data/spec/rawk/line_spec.rb

@@ -0,0 +1,37 @@
+require 'spec_helper'
+
+module Rawk
+  describe Line do
+    before do
+      @fs = " "
+      @data = "a b c\n"
+      @line = Line.new(@data, @fs)
+    end
+    it "is a string" do
+      @line.is_a?(String).should be_true
+    end
+    it "chomps itself on creation" do
+      @line.should == "a b c"
+    end
+    it "finds space-delimited columns by" do
+      @line.cols.should == ["a","b","c"]
+      @line.c.should == @line.cols
+    end
+    it "calculates the number of fields" do
+      @line.nf.should == 3
+    end
+    
+    context "with a field separator ','" do
+      it "splits the line into columns using a comma" do
+        line = Line.new("a,b,c d", ",") 
+        line.c.should == ["a","b","c d"]
+      end
+    end
+    context "with a regular expression field separator" do
+      it "splits the line into columns using the regular expression" do
+        line = Line.new("a,b|c", /[,|]/)
+        line.c.should == ["a","b","c"]
+      end
+    end
+  end
+end

data/spec/rawk/rawk_spec.rb

@@ -0,0 +1,146 @@
+require 'spec_helper'
+require 'stringio'
+ 
+module Rawk
+  describe Program do
+    include TestHelpers
+        
+    before do
+      @data = StringIO.new("a b\nc d\ne f\n")
+      @program = Program.new(@data)
+    end
+    
+    context "when passed code in a string" do      
+      it "runs the string as ruby code" do
+        code = "every {puts 'foo'}"
+        out = capture_stdout { @program.run code }.string
+        out.should == "foo\nfoo\nfoo\n"
+      end
+    end
+    
+    context "when passed code in ruby blocks" do                        
+      it "executes the blocks in order" do
+        start_block, every_block, finish_block = lambda {}, lambda {}, lambda {}
+        start_block.should_receive(:call).once    
+        every_block.should_receive(:call).exactly(3).times        
+        finish_block.should_receive(:call).once
+      
+        @program.run do
+          start  &start_block
+          every  &every_block
+          finish &finish_block
+        end
+      end
+      
+      it "includes the scope of the calling program" do
+        result = []
+        @program.run do 
+          every {|l| result << l}
+        end
+        result.should == ["a b","c d","e f"]
+      end        
+    end
+    
+    describe "condition execution semantics" do
+      describe "start condition" do
+        it "can have many start conditions" do
+          block1, block2 = lambda {}, lambda {}
+          block1.should_receive(:call).once
+          block2.should_receive(:call).once                
+          @program.run do
+            start &block1                    
+            start &block2                    
+          end
+        end
+        it "only calls a duplicate start block once" do
+          block = lambda {}
+          block.should_receive(:call).once
+          @program.run do
+            start &block                  
+            start &block                    
+          end        
+        end
+      end
+  
+      describe "every condition" do 
+        it "runs once for each line of input data" do
+          block = lambda {}
+          block.should_receive(:call).once.with("a b").ordered
+          block.should_receive(:call).once.with("c d").ordered
+          block.should_receive(:call).once.with("e f").ordered
+          @program.run {every &block}
+        end        
+          
+        it "runs multiple every conditions" do
+          block1, block2 = lambda {}, lambda {}
+          block1.should_receive(:call).exactly(3).times
+          block2.should_receive(:call).exactly(3).times
+          @program.run do
+            every &block1
+            every &block2
+          end
+        end
+  
+        it "only runs once when specified many times" do
+          block = lambda {}
+          block.should_receive(:call).exactly(3).times
+          @program.run do
+            every &block
+            every &block
+          end
+        end
+      end
+  
+      describe "finish condition" do  
+        it "can have many finish conditions" do
+          block1, block2 = lambda {}, lambda {}
+          block1.should_receive(:call).once
+          block2.should_receive(:call).once
+          @program.run do
+            finish &block1
+            finish &block2
+          end
+        end
+        it "only runs once when specified many times" do
+          block = lambda {}
+          block.should_receive(:call).once
+          @program.run do
+            finish &block
+            finish &block
+          end
+        end
+      end
+    end
+    
+    describe "support for standard awk built-in variables" do
+      it "calculates the current record num as nr" do
+        record_nums = []
+        @program.run do 
+          start {record_nums << @nr} 
+          every {record_nums << @nr}
+          finish {record_nums << @nr}
+        end
+        record_nums.should == [0,1,2,3,3]
+      end
+      
+      describe "fs" do
+        it "holds the current field separator expression" do
+          @program.fs.should == " "
+        end
+        it "is applied to each line of data" do
+          data = "line"
+          Line.should_receive(:new).with(data, " ")
+          Program.new(data).run {every {|l| nil}}
+        end
+        it "can be changed by the user's program" do
+          data = "line"
+          Line.should_receive(:new).with(data, ",").and_return("dummy")
+          Program.new(data).run do
+            start {@fs = ','}
+            every {|l| nil}
+          end
+        end
+      end
+    end
+  end
+end

data/spec/spec_helper.rb

@@ -0,0 +1,15 @@
+PROJECT_DIR=File.expand_path(File.dirname(__FILE__), '..')
+$: << File.join(PROJECT_DIR, 'lib')
+
+require 'rawk/rawk'
+
+module Rawk::TestHelpers
+  def capture_stdout
+    out = StringIO.new
+    $stdout = out
+    yield
+    return out
+  ensure
+    $stdout = STDOUT
+  end
+end

metadata

@@ -0,0 +1,49 @@
+--- !ruby/object:Gem::Specification
+name: rawk
+version: !ruby/object:Gem::Version
+  version: 0.1.0
+  prerelease: 
+platform: ruby
+authors:
+- Adrian Mowat
+autorequire: rawk/rawk
+bindir: bin
+cert_chain: []
+date: 2011-08-14 00:00:00.000000000Z
+dependencies: []
+description: 
+email: 
+executables: []
+extensions: []
+extra_rdoc_files: []
+files:
+- bin/rawk
+- lib/rawk/rawk.rb
+- spec/rawk/line_spec.rb
+- spec/rawk/rawk_spec.rb
+- spec/spec_helper.rb
+homepage: 
+licenses: []
+post_install_message: 
+rdoc_options: []
+require_paths:
+- - lib
+required_ruby_version: !ruby/object:Gem::Requirement
+  none: false
+  requirements:
+  - - ! '>='
+    - !ruby/object:Gem::Version
+      version: '0'
+required_rubygems_version: !ruby/object:Gem::Requirement
+  none: false
+  requirements:
+  - - ! '>='
+    - !ruby/object:Gem::Version
+      version: '0'
+requirements: []
+rubyforge_project: 
+rubygems_version: 1.8.6
+signing_key: 
+specification_version: 3
+summary: An awk-inspired ruby DSL
+test_files: []