active_median 0.2.1 → 0.2.6

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA256:
-  metadata.gz: cfbbd982f15a11c2f820533f977d7d066e1a8f46a397155681ca06673cf44cc2
-  data.tar.gz: d2156aea58744c1d2955c00077f2b4ef064b2d541d8c69f0ca0a923d93ecbd29
+  metadata.gz: 97636a27f16c0024056aa067afe5d54dcb01247be3236573db0d4a5365cf72ee
+  data.tar.gz: 5c7dc9c490730959e05f4a633765ded663612f02543d62bb17ab05dfd1828f3e
 SHA512:
-  metadata.gz: b5ff96c9f885f59dbc83dd1744fe427f67d1e43eefdb527fcb7e8994427452de186e12e8e0849bc1b15d3bcca2f65c0ec2a7403c631d2dba0f170b87bda074df
-  data.tar.gz: 39671ed1587132c3e54cf9bb1e44f3d9999e7435aacdb2438d03989e94e8cb29b3448663c06bb23fb45315609098d28dc557d0726db612dc290c0229b9a2361a
+  metadata.gz: 48a5936c99b5325a8b7ac43035c7207912911a2e5ddd0889ccc6a4478d35360482cb2062f1756b5e0ba499e54ffff7e79e8adfe9fe29851a028d70194a497949
+  data.tar.gz: abb786892e89b6361e7bf107d702ac4be5e9439f1a5364a7395c490c335c286bcdbcb6c106d130a6ef0870d3b2f100d6ded7c5c69db3ab87fbef4380dca8b9ca

data/CHANGELOG.md

@@ -1,9 +1,31 @@
-## 0.2.1
+## 0.2.6 (2020-12-18)
+
+- Fixed error with certain column types for Active Record 6.1
+
+## 0.2.5 (2020-09-07)
+
+- Added warning for non-attribute argument
+
+## 0.2.4 (2020-03-12)
+
+- Added `percentile` method
+
+## 0.2.3 (2019-09-03)
+
+- Added support for Mongoid
+- Dropped support for Active Record 4.2
+
+## 0.2.2 (2018-10-29)
+
+- Added support for MySQL with udf_infusion
+- Added support for SQL Server and Redshift
+
+## 0.2.1 (2018-10-15)
 
 - Added support for arrays and hashes
 - Added compatibility with Groupdate 4
 
-## 0.2.0
+## 0.2.0 (2018-10-15)
 
 - Added support for MariaDB 10.3.3+ and SQLite
 - Use `PERCENTILE_CONT` for 4x performance increase
@@ -12,23 +34,23 @@ Breaking
 
 - Dropped support for Postgres < 9.4
 
-## 0.1.4
+## 0.1.4 (2016-12-03)
 
 - Added `drop_function` method
 
-## 0.1.3
+## 0.1.3 (2016-03-22)
 
-- Added support for ActiveRecord 5.0
+- Added support for Active Record 5.0
 
-## 0.1.2
+## 0.1.2 (2014-12-27)
 
-- Added support for ActiveRecord 4.2
+- Added support for Active Record 4.2
 
-## 0.1.1
+## 0.1.1 (2014-08-12)
 
 - 10x faster median
 - Added tests
 
-## 0.1.0
+## 0.1.0 (2014-03-13)
 
 - First release

data/LICENSE.txt

@@ -1,4 +1,4 @@
-Copyright (c) 2013 Andrew Kane
+Copyright (c) 2013-2020 Andrew Kane
 
 MIT License
 

data/README.md

@@ -1,35 +1,47 @@
 # ActiveMedian
 
-Median for ActiveRecord
+Median and percentile for Active Record, Mongoid, arrays, and hashes
 
-Supports PostgreSQL 9.4+, MariaDB 10.3.3+, and SQLite, plus arrays and hashes
+Supports:
+
+- PostgreSQL 9.4+
+- MariaDB 10.3.3+
+- MySQL and SQL (with extensions)
+- SQL Server 2012+
+- MongoDB 2.1+
 
 :fire: Uses native functions for blazing performance
 
-[![Build Status](https://travis-ci.org/ankane/active_median.svg)](https://travis-ci.org/ankane/active_median)
+[![Build Status](https://github.com/ankane/active_median/workflows/build/badge.svg?branch=master)](https://github.com/ankane/active_median/actions)
+
+## Getting Started
 
-## Usage
+Add this line to your application’s Gemfile:
 
 ```ruby
-Item.median(:price)
+gem 'active_median'
 ```
 
-Works with grouping, too.
+For MySQL and SQLite, also follow [these instructions](#additional-instructions).
+
+## Models
+
+Median
 
 ```ruby
-Order.group(:store_id).median(:total)
+Item.median(:price)
 ```
 
-## Arrays and Hashes
+Percentile
 
 ```ruby
-[1, 2, 3].median
+Request.percentile(:response_time, 0.95)
 ```
 
-You can also pass a block.
+Works with grouping, too
 
 ```ruby
-{a: 1, b: 2, c: 3}.median { |k, v| v }
+Order.group(:store_id).median(:total)
 ```
 
 ## User Input
@@ -45,15 +57,40 @@ raise "Unpermitted column" unless ["column_a", "column_b"].include?(column)
 User.median(column)
 ```
 
-## Installation
+## Arrays and Hashes
 
-Add this line to your application’s Gemfile:
+Median
 
 ```ruby
-gem 'active_median'
+[1, 2, 3].median
 ```
 
-For SQLite, also follow the instructions below.
+Percentile
+
+```ruby
+[1, 2, 3].percentile(0.95)
+```
+
+You can also pass a block
+
+```ruby
+{a: 1, b: 2, c: 3}.median { |k, v| v }
+```
+
+## Additional Instructions
+
+### MySQL
+
+MySQL requires the `PERCENTILE_CONT` function from [udf_infusion](https://github.com/infusion/udf_infusion). To install it, do:
+
+```sh
+git clone https://github.com/infusion/udf_infusion.git
+cd udf_infusion
+./configure --enable-functions="percentile_cont"
+make
+sudo make install
+mysql <options> < load.sql
+```
 
 ### SQLite
 
@@ -76,7 +113,7 @@ db.enable_load_extension(0)
 
 ### 0.2.0
 
-A user-defined function is no longer needed. Create a migration with `ActiveMedian.drop_function` to remove it.
+A user-defined function is no longer needed for Postgres. Create a migration with `ActiveMedian.drop_function` to remove it.
 
 ## Contributing
 
@@ -86,3 +123,13 @@ Everyone is encouraged to help improve this project. Here are a few ways you can
 - Fix bugs and [submit pull requests](https://github.com/ankane/active_median/pulls)
 - Write, clarify, or fix documentation
 - Suggest or add new features
+
+To get started with development and testing:
+
+```sh
+git clone https://github.com/ankane/active_median.git
+cd active_median
+createdb active_median_test
+bundle install
+bundle exec rake test
+```

data/lib/active_median.rb

@@ -1,6 +1,5 @@
 require "active_support"
 
-require "active_median/model"
 require "active_median/enumerable"
 require "active_median/version"
 
@@ -15,5 +14,11 @@ module ActiveMedian
 end
 
 ActiveSupport.on_load(:active_record) do
+  require "active_median/model"
   extend(ActiveMedian::Model)
 end
+
+ActiveSupport.on_load(:mongoid) do
+  require "active_median/mongoid"
+  Mongoid::Document::ClassMethods.include(ActiveMedian::Mongoid)
+end

data/lib/active_median/enumerable.rb

@@ -1,12 +1,40 @@
 module Enumerable
   unless method_defined?(:median)
     def median(*args, &block)
-      if respond_to?(:scoping) && !block
+      if !block && respond_to?(:scoping)
         scoping { @klass.median(*args) }
+      elsif !block && respond_to?(:with_scope)
+        with_scope(self) { klass.median(*args) }
       else
+        raise ArgumentError, "wrong number of arguments (given #{args.size}, expected 0)" if args.any?
+        percentile(0.5, &block)
+      end
+    end
+  end
+
+  unless method_defined?(:percentile)
+    def percentile(*args, &block)
+      if !block && respond_to?(:scoping)
+        scoping { @klass.percentile(*args) }
+      elsif !block && respond_to?(:with_scope)
+        with_scope(self) { klass.percentile(*args) }
+      else
+        raise ArgumentError, "wrong number of arguments (given #{args.size}, expected 1)" if args.size != 1
+
+        percentile = args[0].to_f
+        raise ArgumentError, "percentile is not between 0 and 1" if percentile < 0 || percentile > 1
+
+        # uses C=1 variant, like percentile_cont
+        # https://en.wikipedia.org/wiki/Percentile#The_linear_interpolation_between_closest_ranks_method
         sorted = map(&block).sort
-        len = sorted.length
-        (sorted[(len - 1) / 2] + sorted[len / 2]) / 2.0
+        x = percentile * (sorted.size - 1)
+        r = x % 1
+        i = x.floor
+        if i == sorted.size - 1
+          sorted[-1]
+        else
+          sorted[i] + r * (sorted[i + 1] - sorted[i])
+        end
       end
     end
   end

data/lib/active_median/model.rb

@@ -1,20 +1,69 @@
 module ActiveMedian
   module Model
     def median(column)
+      percentile(column, 0.5)
+    end
+
+    def percentile(column, percentile)
+      percentile = percentile.to_f
+      raise ArgumentError, "percentile is not between 0 and 1" if percentile < 0 || percentile > 1
+
+      # basic version of Active Record disallow_raw_sql!
+      # symbol = column (safe), Arel node = SQL (safe), other = untrusted
+      # matches table.column and column
+      unless column.is_a?(Symbol) || column.is_a?(Arel::Nodes::SqlLiteral) || /\A\w+(\.\w+)?\z/i.match(column.to_s)
+        warn "[active_median] Non-attribute argument: #{column}. Use Arel.sql() for known-safe values. This will raise an error in ActiveMedian 0.3.0"
+      end
+
+      # column resolution
+      node = relation.send(:arel_columns, [column]).first
+      node = Arel::Nodes::SqlLiteral.new(node) if node.is_a?(String)
+      column = relation.connection.visitor.accept(node, Arel::Collectors::SQLString.new).value
+
+      # prevent SQL injection
+      percentile = connection.quote(percentile)
+
       group_values = all.group_values
 
       relation =
         case connection.adapter_name
         when /mysql/i
+          # assume mariadb by default
+          # use send as this method is private in Rails 4.2
+          mariadb = connection.send(:mariadb?) rescue true
+
+          if mariadb
+            if group_values.any?
+              over = "PARTITION BY #{group_values.join(", ")}"
+            end
+
+            select(*group_values, "PERCENTILE_CONT(#{percentile}) WITHIN GROUP (ORDER BY #{column}) OVER (#{over})").unscope(:group)
+          else
+            # if mysql gets native function, check (and memoize) version first
+            select(*group_values, "PERCENTILE_CONT(#{column}, #{percentile})")
+          end
+        when /sqlserver/i
           if group_values.any?
             over = "PARTITION BY #{group_values.join(", ")}"
           end
 
-          select(*group_values, "PERCENTILE_CONT(0.50) WITHIN GROUP (ORDER BY #{column}) OVER (#{over})").unscope(:group)
+          select(*group_values, "PERCENTILE_CONT(#{percentile}) WITHIN GROUP (ORDER BY #{column}) OVER (#{over})").unscope(:group)
         when /sqlite/i
-          select(*group_values, "MEDIAN(#{column})")
+          case percentile.to_f
+          when 0
+            select(*group_values, "MIN(#{column})")
+          when 0.5
+            select(*group_values, "MEDIAN(#{column})")
+          when 1
+            select(*group_values, "MAX(#{column})")
+          else
+            # LOWER_QUARTILE and UPPER_QUARTILE use different calculation than 0.25 and 0.75
+            raise "SQLite only supports 0, 0.5, and 1 percentiles"
+          end
+        when /postg/i, /redshift/i # postgis too
+          select(*group_values, "PERCENTILE_CONT(#{percentile}) WITHIN GROUP (ORDER BY #{column})")
         else
-          select(*group_values, "PERCENTILE_CONT(0.50) WITHIN GROUP (ORDER BY #{column})")
+          raise "Connection adapter not supported: #{connection.adapter_name}"
         end
 
       result = connection.select_all(relation.to_sql)
@@ -22,9 +71,8 @@ module ActiveMedian
       # typecast
       rows = []
       columns = result.columns
-      cast_method = ActiveRecord::VERSION::MAJOR < 5 ? :type_cast : :cast_value
       result.rows.each do |untyped_row|
-        rows << (result.column_types.empty? ? untyped_row : columns.each_with_index.map { |c, i| untyped_row[i] ? result.column_types[c].send(cast_method, untyped_row[i]) : untyped_row[i] })
+        rows << (result.column_types.empty? ? untyped_row : columns.each_with_index.map { |c, i| untyped_row[i] && result.column_types[c] ? result.column_types[c].send(:cast_value, untyped_row[i]) : untyped_row[i] })
       end
 
       result =

data/lib/active_median/mongoid.rb

@@ -0,0 +1,30 @@
+module ActiveMedian
+  module Mongoid
+    def median(column)
+      percentile(column, 0.5)
+    end
+
+    # https://www.compose.com/articles/mongo-metrics-finding-a-happy-median/
+    def percentile(column, percentile)
+      percentile = percentile.to_f
+      raise ArgumentError, "percentile is not between 0 and 1" if percentile < 0 || percentile > 1
+
+      relation =
+        all
+        .asc(column)
+        .group(_id: nil, values: {"$push" => "$#{column}"}, count: {"$sum" => 1})
+        .project(values: 1, count: {"$subtract" => ["$count", 1]})
+        .project(values: 1, count: 1, x: {"$multiply" => ["$count", percentile]})
+        .project(values: 1, count: 1, r: {"$mod" => ["$x", 1]}, i: {"$floor" => "$x"})
+        .project(values: 1, count: 1, r: 1, i: 1, i2: {"$add" => ["$i", 1]})
+        .project(values: 1, count: 1, r: 1, i: 1, i2: {"$min" => ["$i2", "$count"]})
+        .project(r: 1, beginValue: {"$arrayElemAt" => ["$values", "$i"]}, endValue: {"$arrayElemAt" => ["$values", "$i2"]})
+        .project(r: 1, beginValue: 1, result: {"$subtract" => ["$endValue", "$beginValue"]})
+        .project(beginValue: 1, result: {"$multiply" => ["$result", "$r"]})
+        .project(result: {"$add" => ["$beginValue", "$result"]})
+
+      res = collection.aggregate(relation.pipeline).first
+      res ? res["result"] : nil
+    end
+  end
+end

data/lib/active_median/version.rb

@@ -1,3 +1,3 @@
 module ActiveMedian
-  VERSION = "0.2.1"
+  VERSION = "0.2.6"
 end

metadata

@@ -1,29 +1,29 @@
 --- !ruby/object:Gem::Specification
 name: active_median
 version: !ruby/object:Gem::Version
-  version: 0.2.1
+  version: 0.2.6
 platform: ruby
 authors:
 - Andrew Kane
-autorequire: 
+autorequire:
 bindir: bin
 cert_chain: []
-date: 2018-10-15 00:00:00.000000000 Z
+date: 2020-12-19 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
-  name: activerecord
+  name: activesupport
   requirement: !ruby/object:Gem::Requirement
     requirements:
     - - ">="
       - !ruby/object:Gem::Version
-        version: '4.2'
+        version: '5'
   type: :runtime
   prerelease: false
   version_requirements: !ruby/object:Gem::Requirement
     requirements:
     - - ">="
       - !ruby/object:Gem::Version
-        version: '4.2'
+        version: '5'
 - !ruby/object:Gem::Dependency
   name: bundler
   requirement: !ruby/object:Gem::Requirement
@@ -56,16 +56,16 @@ dependencies:
   name: pg
   requirement: !ruby/object:Gem::Requirement
     requirements:
-    - - "<"
+    - - ">="
       - !ruby/object:Gem::Version
-        version: '1'
+        version: '0'
   type: :development
   prerelease: false
   version_requirements: !ruby/object:Gem::Requirement
     requirements:
-    - - "<"
+    - - ">="
       - !ruby/object:Gem::Version
-        version: '1'
+        version: '0'
 - !ruby/object:Gem::Dependency
   name: rake
   requirement: !ruby/object:Gem::Requirement
@@ -80,6 +80,20 @@ dependencies:
     - - ">="
       - !ruby/object:Gem::Version
         version: '0'
+- !ruby/object:Gem::Dependency
+  name: activerecord
+  requirement: !ruby/object:Gem::Requirement
+    requirements:
+    - - ">="
+      - !ruby/object:Gem::Version
+        version: '0'
+  type: :development
+  prerelease: false
+  version_requirements: !ruby/object:Gem::Requirement
+    requirements:
+    - - ">="
+      - !ruby/object:Gem::Version
+        version: '0'
 - !ruby/object:Gem::Dependency
   name: groupdate
   requirement: !ruby/object:Gem::Requirement
@@ -94,7 +108,7 @@ dependencies:
     - - ">="
       - !ruby/object:Gem::Version
         version: '0'
-description: 
+description:
 email: andrew@chartkick.com
 executables: []
 extensions: []
@@ -106,12 +120,13 @@ files:
 - lib/active_median.rb
 - lib/active_median/enumerable.rb
 - lib/active_median/model.rb
+- lib/active_median/mongoid.rb
 - lib/active_median/version.rb
 homepage: https://github.com/ankane/active_median
 licenses:
 - MIT
 metadata: {}
-post_install_message: 
+post_install_message:
 rdoc_options: []
 require_paths:
 - lib
@@ -119,16 +134,15 @@ required_ruby_version: !ruby/object:Gem::Requirement
   requirements:
   - - ">="
     - !ruby/object:Gem::Version
-      version: '2.2'
+      version: '2.4'
 required_rubygems_version: !ruby/object:Gem::Requirement
   requirements:
   - - ">="
     - !ruby/object:Gem::Version
       version: '0'
 requirements: []
-rubyforge_project: 
-rubygems_version: 2.7.7
-signing_key: 
+rubygems_version: 3.1.4
+signing_key:
 specification_version: 4
-summary: Median for ActiveRecord
+summary: Median and percentile for Active Record, Mongoid, arrays, and hashes
 test_files: []