data_kit 0.0.5 → 0.0.6

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA1:
-  metadata.gz: e93d042b5d65309bf8d99ff311b0a9d9dd3b2899
-  data.tar.gz: 7ddf9b7d5a11ffc59ef0c093dcd2378270099410
+  metadata.gz: 11a12efb158e60cac6e71ac09e37799f811a65e4
+  data.tar.gz: 66ed506814c7273bdc980ada09049ebaabaa3a52
 SHA512:
-  metadata.gz: f4231fed6d650d0b5ba39771469c014ccfe70f1f360c9a299604ee5a4e616566b5bc7cf8feab9cbb5f89e3ae4bfe5387cf528dfa664a9f0d25151a6bade4f34b
-  data.tar.gz: 5caa3516fcb3b8aa5eeef7ef57e9a8c27e9108a580ebd77917502e2c4f9bd1d2f2c90459a85663028e081684a8b3dab7845cdda8879f267fe9a2a966e0fe975e
+  metadata.gz: f2f3eb974bde2f7bdbdcd34fdc4ea57f4ee64c4876768f436a2a5e54f79eba3254f26247b786bbacee869dea17b91de716950fab97ac8ceeb3843f02d51f8a52
+  data.tar.gz: e1f92f4116af3087938dfcc715605c7d805dd87b408f0dfa4c9e9494587c925451e0d749c528df5eaac400b063d54efbbf97fdb02e3f783981f02d09e33cbc3d

data/Gemfile.lock

@@ -1,7 +1,7 @@
 PATH
   remote: .
   specs:
-    data_kit (0.0.4)
+    data_kit (0.0.6)
       rcsv
       timeliness
 

data/lib/data_kit/csv/schema_analysis.rb

@@ -15,10 +15,10 @@ module DataKit
 
         @type_hints = {}
 
-        if options[:use_type_hints].nil? || options[:use_type_hints] == false
-          @use_type_hints = false
-        else
+        if options[:use_type_hints].nil? || options[:use_type_hints] == true
           @use_type_hints = true
+        else
+          @use_type_hints = false
         end
 
         fields.each do |field_name|

data/lib/data_kit/csv/schema_analyzer.rb

@@ -4,16 +4,23 @@ module DataKit
       attr_accessor :csv
       attr_accessor :keys
       attr_accessor :sampling_rate
-      
+      attr_accessor :use_type_hints
+
       def initialize(csv, options = {})
         @csv = csv
         @keys = options[:keys] || []
         @sampling_rate = options[:sampling_rate] || 0.1
+
+        if options[:use_type_hints].nil? || options[:use_type_hints] == true
+          @use_type_hints = true
+        else
+          @use_type_hints = false
+        end
       end
 
       def execute
         random = Random.new
-        analysis = SchemaAnalysis.new(csv.headers)
+        analysis = SchemaAnalysis.new(csv.headers, :use_type_hints => use_type_hints)
 
         csv.each_row do |row|
           analysis.increment_total

data/lib/data_kit/version.rb

@@ -1,3 +1,3 @@
 module DataKit
-  VERSION = "0.0.5"
+  VERSION = "0.0.6"
 end

data/spec/csv/schema_analyzer_spec.rb

@@ -19,6 +19,12 @@ describe DataKit::CSV::SchemaAnalyzer do
     analyzer.csv.should == csv
     analyzer.keys.should == []
     analyzer.sampling_rate.should == 0.1
+    analyzer.use_type_hints.should == true
+  end
+
+  it "should initialize with type hints turned off" do
+    analyzer = DataKit::CSV::SchemaAnalyzer.new(csv, :use_type_hints => false)
+    analyzer.use_type_hints.should == false
   end
 
   it "should initialize schema with an IO" do
@@ -29,7 +35,7 @@ describe DataKit::CSV::SchemaAnalyzer do
     analyzer.sampling_rate.should == 0.1
   end
 
-  it "should execute an analysis" do
+  it "should execute an analysis with type hints" do
     analysis = DataKit::CSV::SchemaAnalyzer.new(csv, :sampling_rate => 0.5).execute
 
     analysis.type?('id').should == :integer
@@ -45,6 +51,29 @@ describe DataKit::CSV::SchemaAnalyzer do
 
     analysis.row_count.should == 10
     analysis.sample_count.should be < 10
+    analysis.use_type_hints.should == true
+
+    puts analysis.type_hints.inspect
+
+  end
+
+  it "should execute an analysis without type hints" do
+    analysis = DataKit::CSV::SchemaAnalyzer.new(csv, :sampling_rate => 0.5, :use_type_hints => false).execute
+
+    analysis.type?('id').should == :integer
+    analysis.type?('first_name').should == :string
+    analysis.type?('last_name').should == :string
+    analysis.type?('email').should == :string
+    analysis.type?('country').should == :string
+    analysis.type?('ip_address').should == :string
+    analysis.type?('amount').should == :number
+    analysis.type?('active').should == :boolean
+    analysis.type?('activated_at').should == :datetime
+    analysis.type?('address').should == :string
+
+    analysis.row_count.should == 10
+    analysis.sample_count.should be < 10
+    analysis.use_type_hints.should == false
   end
 
   it "should calculate a sampling_rate" do

metadata

@@ -1,7 +1,7 @@
 --- !ruby/object:Gem::Specification
 name: data_kit
 version: !ruby/object:Gem::Version
-  version: 0.0.5
+  version: 0.0.6
 platform: ruby
 authors:
 - Mode Analytics
@@ -198,4 +198,3 @@ test_files:
 - spec/fixtures/standard.csv
 - spec/fixtures/utf8.csv
 - spec/spec_helper.rb
-has_rdoc: