masticate 0.4.2 → 0.5.0

data/lib/masticate.rb

@@ -1,23 +1,9 @@
 require "open-uri"
 require "csv"
 
-require_relative "masticate/version"
-require_relative "masticate/base"
-require_relative "masticate/myoptparse"
-
-require_relative "masticate/sniffer"
-require_relative "masticate/mender"
-require_relative "masticate/csvify"
-require_relative "masticate/plucker"
-require_relative "masticate/datify"
-require_relative "masticate/gsubber"
-require_relative "masticate/max_rows"
-require_relative "masticate/concat"
-require_relative "masticate/relabel"
-require_relative "masticate/exclude"
-require_relative "masticate/transform"
-
-require_relative "masticate/cook"
+%w{version base myoptparse sniffer mender csvify plucker datify gsubber max_rows concat relabel exclude transform include cook}.each do |f|
+  require_relative "masticate/#{f}"
+end
 
 module Masticate
   def self.sniff(filename, opts = {})
@@ -60,6 +46,10 @@ module Masticate
     Exclude.new(filename).exclude(opts)
   end
 
+  def self.include(filename, opts)
+    Include.new(filename).exclude(opts)
+  end
+
   def self.cook(filename, opts)
     Cook.new(filename).cook(opts)
   end

data/lib/masticate/include.rb

@@ -0,0 +1,43 @@
+# exclude rows based on field = value
+
+class Masticate::Include < Masticate::Base
+  def configure(opts)
+    standard_options(opts)
+
+    @field = opts[:field] or raise "missing field to include"
+    @value = opts[:value] or raise "missing value to include"
+
+    # row-loading automatically strips leading & trailing whitespace and converts blanks to nils,
+    # so when looking for blanks need to compare to nil instead of ''
+    @value = nil if @value.empty?
+  end
+
+  def exclude(opts)
+    execute(opts)
+  end
+
+  def crunch(row)
+    if !@headers
+      @headers = row
+      f = @field
+      @index =
+        case f
+        when Fixnum, /^\d+$/
+          f = f.to_i
+          if f > row.count
+            raise "Cannot pluck column #{f}, there are only #{row.count} fields"
+          else
+            f-1
+          end
+        else
+          row.index(f) or raise "Unable to find column '#{f}' in headers"
+        end
+      row
+    elsif row
+      if row[@index] == @value
+        # include only these matches
+        row
+      end
+    end
+  end
+end

data/lib/masticate/myoptparse.rb

@@ -165,6 +165,10 @@ EOT
       results = Masticate.exclude(filename, options)
       logmessage(command, options, results)
 
+    when 'include'
+      results = Masticate.include(filename, options)
+      logmessage(command, options, results)
+
     when 'transform'
       results = Masticate.transform(filename, options)
       logmessage(command, options, results)

data/lib/masticate/version.rb

@@ -1,3 +1,3 @@
 module Masticate
-  VERSION = "0.4.2"
+  VERSION = "0.5.0"
 end

data/spec/data/include_input.csv

@@ -0,0 +1,6 @@
+ID,DATE,CRITERION,VALUE
+1,201201060826,yes,594823
+2,201201060521,no,594790
+7,201201060429,yes,594780
+11,201201060446,no,594786
+8,201201061011,no,594857

data/spec/data/include_results.csv

@@ -0,0 +1,3 @@
+ID,DATE,CRITERION,VALUE
+1,201201060826,yes,594823
+7,201201060429,yes,594780

data/spec/lib/include_spec.rb

@@ -0,0 +1,15 @@
+# spec for row inclusion
+
+require "spec_helper"
+
+describe "include" do
+  it "should be pick only rows that match criterion" do
+    filename = File.dirname(__FILE__) + "/../data/include_input.csv"
+    tmp = Tempfile.new('include')
+    results = Masticate.include(filename, :output => tmp, :field => 'CRITERION', :value => 'yes')
+    output = File.read(tmp)
+    correct_output = File.read(File.dirname(__FILE__) + "/../data/include_results.csv")
+
+    output.should == correct_output
+  end
+end

metadata

@@ -1,7 +1,7 @@
 --- !ruby/object:Gem::Specification
 name: masticate
 version: !ruby/object:Gem::Version
-  version: 0.4.2
+  version: 0.5.0
   prerelease: 
 platform: ruby
 authors:
@@ -9,7 +9,7 @@ authors:
 autorequire: 
 bindir: bin
 cert_chain: []
-date: 2012-09-27 00:00:00.000000000 Z
+date: 2012-10-03 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: rake
@@ -99,6 +99,7 @@ files:
 - lib/masticate/datify.rb
 - lib/masticate/exclude.rb
 - lib/masticate/gsubber.rb
+- lib/masticate/include.rb
 - lib/masticate/max_rows.rb
 - lib/masticate/mender.rb
 - lib/masticate/myoptparse.rb
@@ -125,6 +126,8 @@ files:
 - spec/data/events_reduced.csv
 - spec/data/exclude_input.csv
 - spec/data/exclude_results.csv
+- spec/data/include_input.csv
+- spec/data/include_results.csv
 - spec/data/inlined_headers.csv
 - spec/data/inlined_headers.csv.output
 - spec/data/junk_header.csv
@@ -147,6 +150,7 @@ files:
 - spec/lib/datify_spec.rb
 - spec/lib/exclude_spec.rb
 - spec/lib/gsub_spec.rb
+- spec/lib/include_spec.rb
 - spec/lib/maxrow_spec.rb
 - spec/lib/mender_spec.rb
 - spec/lib/plucker_spec.rb
@@ -168,7 +172,7 @@ required_ruby_version: !ruby/object:Gem::Requirement
       version: '0'
       segments:
       - 0
-      hash: -88830056329849301
+      hash: -1109871130254575222
 required_rubygems_version: !ruby/object:Gem::Requirement
   none: false
   requirements:
@@ -177,7 +181,7 @@ required_rubygems_version: !ruby/object:Gem::Requirement
       version: '0'
       segments:
       - 0
-      hash: -88830056329849301
+      hash: -1109871130254575222
 requirements: []
 rubyforge_project: masticate
 rubygems_version: 1.8.24
@@ -202,6 +206,8 @@ test_files:
 - spec/data/events_reduced.csv
 - spec/data/exclude_input.csv
 - spec/data/exclude_results.csv
+- spec/data/include_input.csv
+- spec/data/include_results.csv
 - spec/data/inlined_headers.csv
 - spec/data/inlined_headers.csv.output
 - spec/data/junk_header.csv
@@ -224,6 +230,7 @@ test_files:
 - spec/lib/datify_spec.rb
 - spec/lib/exclude_spec.rb
 - spec/lib/gsub_spec.rb
+- spec/lib/include_spec.rb
 - spec/lib/maxrow_spec.rb
 - spec/lib/mender_spec.rb
 - spec/lib/plucker_spec.rb