pg_histogram 0.1 → 0.2

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA1:
-  metadata.gz: f1fd000101bac4dc9af44882ab5dd641b3062430
-  data.tar.gz: dc532f7711f65c11aa57706edb3ac6a17124f2e1
+  metadata.gz: 8fa7ea082187e3b4ded20bbcd8580ad3ff06f7ee
+  data.tar.gz: 510d7260c0cff0c67e2cccacbb4b788859b7a58f
 SHA512:
-  metadata.gz: dd6ef99ccd68ce5bf02bcf75c79a76bc95b61b91c0a44417cc4283ccd659b71af989b80f3e225bfd1b720695c79df0cbb8f00b045134a2bf01e91d7348af3c18
-  data.tar.gz: 096ae0c70711f89ca3a3f9d9a5dc0c7f4585a14248504244859a49e24f6e2ff2edb34017d4f21f4e1853e70d644eb793caa346b0a3a03e8f2875d1120e0e9b33
+  metadata.gz: b27b4ef2219e53ee2b407743419dd5d4756ba7ddc0b21a916e69aa5c05a709ff0bbdf14424f8f6913078bd17fc812ec3af4692cbe635144595fcabe74c9ab4a1
+  data.tar.gz: 41fa94e1a691710b4b88d5f1656aea2fad5fca5e3a05d70caeb54c09fee4d2286c6416d0c61363c97821e7c8a1c0cdad6a57f2f7b3fe8f0c58005dd4babb3b40

data/README.md

@@ -2,9 +2,7 @@
 
 This gem allows for you to efficiently create a histogram from large data sets in your Rails applications.
 
-It uses PostgreSQL's [width_bucket](http://www.postgresql.org/docs/9.3/static/functions-math.html) function to handle the majority of the processing in the database, and only requires 3 database queries.
-
-
+It uses PostgreSQL's [width_bucket](http://www.postgresql.org/docs/9.3/static/functions-math.html) function to handle the majority of the processing in the database, and only requires 3 database queries (and only one query if min and max values are specified).
 
 ## Installation
 
@@ -22,28 +20,31 @@ Or install it yourself as:
 
 ## Usage
 
-Create a Histogram object using the following three parameters:
-
-1. ActiveRecord query to use
-2. Name of column to count frequency of
-3. Bucket size (OPTIONAL - default is 0.5)
-
-<!-- -->
-    histogram = PgHistogram::Histogram.new(Widget.all, 'price', 0.5)
+Create a Histogram object using the following parameters:
 
+1. ActiveRecord Relation (query) to use.
+2. Name of column to count frequency of. Also allows for aliased queries such as `'price*discount as final_price'` to create histograms on expressions.
+3. Options hash (optional). Not all combinations are allowed. For example, if `:buckets` is specified, `:min` and `:max` are required and `:bucket_size` is ignored, and calculated. If `:buckets` is not specified, the number of buckets depends on `:bucket_size`, and `:min` and `:max` are optional.
+    - `:buckets`: number of buckets (integer)
+    - `:min` and `:max`: See [width_bucket](http://www.postgresql.org/docs/9.3/static/functions-math.html)'s docs for exact meaning (defaults to the min and max values of the column).
+    - `:bucket_size`: Width of each bucket (defaults to 1).
 
-Call the results method to retrieve a Hash of bucket minimums and frequency counts
+### Example
+Create sample data:
 
-    # create sample data
     5.times do { Widget.create(price: 1.2) }
     10.times do { Widget.create(price: 2.9 ) }
 
-    # get the results
+Create the histogram object:
+    
+    histogram = PgHistogram::Histogram.new(Widget.all, 'price', 0.5)
+
+Call the results method to retrieve a Hash of bucket minimums and frequency counts:
+
     @histogram_data = histogram.results
      => {1.0=>5, 2.5=>10}
 
-
-The results can be used by your favorite charting libary, such as [Chartkick](https://github.com/ankane/chartkick), to plot the data.
+The results can be used by your favorite charting libary, such as [Chartkick](https://github.com/ankane/chartkick), to plot the data:
 
     <%= column_chart @histogram_data %>
 

data/lib/pg_histogram/histogram.rb

@@ -11,10 +11,23 @@ module PgHistogram
     }
 
     # column_name name must be safe for SQL injection
-    def initialize(query, column_name, bucket_size = 0.5)
+    def initialize(query, column_name, options = {})
       @query = query
       @column = column_name.to_s
-      @bucket_size = bucket_size
+      if options.is_a? Hash
+        if options[:buckets]
+          @min = options[:min] || 0
+          @max = options[:max]
+          @buckets = options[:buckets]
+          @bucket_size = calculate_bucket_size
+        else
+          @min = options[:min]
+          @max = options[:max]
+          @bucket_size = (options[:bucket_size] || 1).to_f
+        end
+      else
+        @bucket_size = options.to_f
+      end
     end
 
     # returns histogram as hash
@@ -23,22 +36,34 @@ module PgHistogram
     def results
       # error handling case
       if max == min
-        { min => query.where("#{column} = ?", min).count }
+        { min => subquery.where("#{pure_column} = ?", min).count }
       else
         labeled_histogram
       end
     end
 
     def min
-      @min ||= round_to_increment(query.minimum(column), :down)
+      @min ||= round_to_increment(source_min, :down)
     end
 
     def max
-      @max ||= round_to_increment(query.maximum(column), :up)
+      @max ||= round_to_increment(source_max, :up)
     end
 
     private
 
+    def source_min
+      @source_min ||= subquery.minimum(pure_column(true))
+    end
+
+    def source_max
+      @source_max ||= subquery.maximum(pure_column(true))
+    end
+
+    def calculate_bucket_size
+      (source_max - source_min).to_f / @buckets
+    end
+
     def num_buckets
       @buckets ||= ((max - min) / bucket_size).to_i
     end
@@ -68,19 +93,44 @@ module PgHistogram
     def query_for_buckets
       ActiveRecord::Base.connection.execute(
         <<-SQL
-          SELECT width_bucket(#{column}, #{min}, #{max}, #{num_buckets}) as #{BUCKET_COL},
+          SELECT width_bucket(#{pure_column}, #{min}, #{max}, #{num_buckets}) as #{BUCKET_COL},
             count(*) as #{FREQUENCY_COL}
-          FROM (#{subquery.to_sql}) as subq_results
+          FROM (#{subquery_sql}) as subq_results
           GROUP BY #{BUCKET_COL}
           ORDER BY #{BUCKET_COL}
         SQL
       )
     end
-
     # use passed AR query as a subquery to not interfere with group clause
     def subquery
       # override default order
       query.select(column).order('1')
     end
+
+    # Use unprepared statement per https://github.com/rails/rails/issues/8743
+    def subquery_sql
+      ActiveRecord::Base.connection.unprepared_statement do 
+        subquery.to_sql
+      end
+    end
+    
+    # In case the column has an alias, the pure column is just the aliased name
+    # If expression is true, only the expression (before the 'AS') is returned
+    def pure_column(expression = false)
+      index = column =~ / as /i
+      # If AS is present, split and keep either side
+      if index
+        if expression
+          # Keep left side
+          column[0..index]
+        else
+          # Keep right side
+          column[index + 4..-1]
+        end
+      else
+        # Column was already good.
+        column
+      end
+    end
   end
-end
+end

data/lib/pg_histogram/version.rb

@@ -1,3 +1,3 @@
 module PgHistogram
-  VERSION = "0.1"
+  VERSION = "0.2"
 end

data/test/histogram_test.rb

@@ -60,7 +60,7 @@ class HistogramTest < Minitest::Test
   end
 
   def test_rounding_to_bucket_size
-    hist = PgHistogram::Histogram.new(nil, nil, 0.25)
+    hist = PgHistogram::Histogram.new(nil, nil, bucket_size: 0.25)
 
     assert_equal 0.5, hist.send(:round_to_increment, 0.478), '0.478 rounded to 0.25 interval'
     assert_equal 1.0, hist.send(:round_to_increment, 1.1), '1.1 rounded to 0.25 interval'
@@ -72,4 +72,53 @@ class HistogramTest < Minitest::Test
     assert_equal 1.25, hist.send(:round_to_increment, 1.1, :up), '1.1 rounded up to 0.25 interval'
     assert_equal 0.5, hist.send(:round_to_increment, 0.5, :up), '0.5 rounded up to 0.25 interval'
   end
+
+  def test_with_buckets_option
+    # Specify number of buckets
+    hist = PgHistogram::Histogram.new(Widget.all, 'price', {buckets: 5, min: 0, max: 10})
+
+    10.times { Widget.create!(price: 3.0) }
+    8.times { Widget.create!(price: 5.76) }
+    results = hist.results
+
+    assert_equal 0, hist.min, 'Histogram minimum price'
+    assert_equal 10, hist.max, 'Histogram maximum price'
+    assert_equal 5, hist.send(:num_buckets), 'Histogram buckets'
+    assert_equal 2, results.size, 'Histogram buckets with results'
+    assert_equal 10, results[2.0], 'Frequency of 2.0 bucket'
+    assert_equal 8, results[4.0], 'Frequency of 4.0 bucket'
+  end
+
+  def test_with_min_max_options
+    # Specify number of buckets
+    hist = PgHistogram::Histogram.new(Widget.all, 'price', {min: 0, max: 10})
+
+    10.times { Widget.create!(price: 3.0) }
+    8.times { Widget.create!(price: 5.76) }
+    min_price = Widget.create!(price: 0.98).price
+    max_price = Widget.create!(price: 6.0).price
+    results = hist.results
+
+    assert_equal 0, hist.min, 'Histogram minimum price'
+    assert_equal 10, hist.max, 'Histogram maximum price'
+    assert_equal 10, hist.send(:num_buckets), 'Histogram buckets'
+    assert_equal 4, results.size, 'Histogram buckets with results'
+    assert_equal 10, results[3.0], 'Frequency of 3 bucket'
+    assert_equal nil, results[4.0], 'Frequency of 4.0 bucket'
+  end
+  def test_with_aliased_expression
+    # Specify number of buckets
+    hist = PgHistogram::Histogram.new(Widget.all, 'price*2 as double_price')
+
+    5.times { Widget.create!(price: 3.0) }
+    6.times { Widget.create!(price: 6.0) }
+    results = hist.results
+
+    assert_equal 6, hist.min, 'Histogram minimum price'
+    assert_equal 12, hist.max, 'Histogram maximum price'
+    assert_equal 6, hist.send(:num_buckets), 'Histogram buckets'
+    assert_equal 2, results.size, 'Histogram buckets with results'
+    assert_equal 5, results[6.0], 'Frequency of 6.0 bucket'
+    assert_equal 6, results[12.0], 'Frequency of 12.0 bucket'
+  end
 end

metadata

@@ -1,14 +1,14 @@
 --- !ruby/object:Gem::Specification
 name: pg_histogram
 version: !ruby/object:Gem::Version
-  version: '0.1'
+  version: '0.2'
 platform: ruby
 authors:
 - David Roberts
 autorequire: 
 bindir: bin
 cert_chain: []
-date: 2014-10-09 00:00:00.000000000 Z
+date: 2016-02-12 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: activerecord
@@ -83,7 +83,6 @@ files:
 - lib/pg_histogram/histogram.rb
 - lib/pg_histogram/version.rb
 - pg_histogram.gemspec
-- test/database.yml
 - test/histogram_test.rb
 - test/test_helper.rb
 homepage: https://github.com/eLocal/pg_histogram
@@ -111,6 +110,5 @@ signing_key:
 specification_version: 4
 summary: Histograms using PostgreSQL and ActiveRecord
 test_files:
-- test/database.yml
 - test/histogram_test.rb
 - test/test_helper.rb

data/test/database.yml

@@ -1,6 +0,0 @@
-test:
-  adapter: postgresql
-  database: histogram_test
-  username: histogram
-  password: 5q3KwfCR
-  host: localhost