ruby_native_statistics 0.9.0 → 0.10.0

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA256:
-  metadata.gz: b99d6ace1e36cdd53315356b8c4de080289a46bc307b41bc1979de4cbaff2e42
-  data.tar.gz: 5b8c3498b2a3014313b0b249eb87350e7f07e5a22237de4ea9e95f67096572c7
+  metadata.gz: 99451de1aa0d6eab93118c4ce68330f21210437115e05ef3bcdff4ae63e780d7
+  data.tar.gz: edf6bdec21f87d70e808eaca249ddf1447b0ca57979a039e2295d679976e73e0
 SHA512:
-  metadata.gz: c837f3f6f7e2a8d47291d1e289d09e48de81c5216bca319e1be24aca7a821f0ba8e1acd3548ccfbe83a50b92752af5dc281a9a69dc052e58cee1f5e49b05e06d
-  data.tar.gz: 3c8e3629ffe57bccdbdeaa8b792d593aa89bdce401e034a2a0d08fdf62fb1ad0841229cc60c222aaa37fe0895c2fdee83109176ec553e0c1529c33814868fa4e
+  metadata.gz: 6e12d43e6f7ac1dd5b92350252e2feb83d69a07c311edd7f97040f2af471dad7763bb3515c769dab7e64b40e0d6ecabd92b3f546dea5178cc16638ce5564fbd2
+  data.tar.gz: cd0ef3c36d85d60505f5826d25a0189e8bcdfa0d7799b574f16d010ea9e4ba1367f66dba522e3287df5248e0dc11d5631252ad726b1050b8e730bc3fc7cbc045

data/.vscode/settings.json

@@ -3,5 +3,9 @@
   "editor.tabSize": 2,
   "search.useGlobalIgnoreFiles": true,
   "editor.wordWrap": "on",
-  "editor.renderWhitespace": "all"
+  "editor.renderWhitespace": "all",
+  "files.associations": {
+    "conversions.h": "c",
+    "algorithm": "c"
+  }
 }

data/Gemfile.lock

@@ -1,7 +1,7 @@
 PATH
   remote: .
   specs:
-    ruby_native_statistics (0.8)
+    ruby_native_statistics (0.10.0)
 
 GEM
   remote: https://rubygems.org/

data/README.md

@@ -8,6 +8,7 @@ This is a native extension to Ruby that adds native (C) statistical functions to
 - [Population Variance](https://en.wikipedia.org/wiki/Variance#Population_variance_and_sample_variance) (varp)
 - [Median](https://en.wikipedia.org/wiki/Median) (median)
 - [Mean](https://en.wikipedia.org/wiki/Arithmetic_mean) (mean)
+- [Percentile](https://en.wikipedia.org/wiki/Quantile) (percentile)
 
 Check the Github Actions build to see the currently supported versions of Ruby. This list will match whatever stable versions are specified at https://www.ruby-lang.org/en/downloads/.
 
@@ -16,6 +17,10 @@ It is much more performant than calculating the standard deviation with pure Rub
     bench_native_dispersion	 0.000425	 0.000341	 0.000420	 0.000324	 0.000319
     bench_ruby_dispersion	 0.002168	 0.002156	 0.002148	 0.002149	 0.002151
 
+## Found a bug? Need a function?
+
+If you found a bug or would need a particulr function, please let me know! I work on this gem in my spare time, mainly for learning purposes. Feel free to open a PR or a Github issue and I'll take a look as soon as possible.
+
 ## Usage
 
     require 'ruby_native_statistics'
@@ -33,6 +38,15 @@ It is much more performant than calculating the standard deviation with pure Rub
     # calculate median
     p r.median
 
+    # calculate percentile
+    p r.percentile(0.3333)
+
+## Implementation notes
+
+### Percentile
+
+Percentile uses the same rounding method as Excel, sometimes called R7.
+
 ## Links
 
 This is the third version of this gem, and it is a total rewrite of a SWIG-based design. Lots of thanks to the following resources:

data/Rakefile

@@ -1,12 +1,8 @@
 require "rake/extensiontask"
 require "rake/testtask"
 
-Rake::ExtensionTask.new "dispersion" do |ext|
-  ext.lib_dir = "lib"
-end
-
-Rake::ExtensionTask.new "mathematics" do |ext|
-  ext.lib_dir = "lib"
+Rake::ExtensionTask.new "ruby_native_statistics" do |ext|
+  ext.lib_dir = "lib/ruby_native_statistics"
 end
 
 Rake::TestTask.new(:test) do |t|

data/ext/ruby_native_statistics/conversions.c

@@ -0,0 +1,45 @@
+#include "conversions.h"
+
+int compare_doubles(const void *a, const void *b)
+{
+  double *dbl_a = (double *)a;
+  double *dbl_b = (double *)b;
+
+  double cmp_a = *dbl_a;
+  double cmp_b = *dbl_b;
+
+  return (cmp_a - cmp_b);
+}
+
+double *sorted_ruby_array(VALUE array, long array_length)
+{
+  long i;
+  double *working_array;
+
+  working_array = malloc(array_length * sizeof(double));
+
+  if (working_array == NULL)
+  {
+    rb_raise(rb_eStandardError, "unknown problem calculating median (possibly array is too large)");
+  }
+
+  for (i = 0; i < array_length; i++)
+  {
+    VALUE item = rb_ary_entry(array, i);
+
+    if (!RB_INTEGER_TYPE_P(item) && !RB_FLOAT_TYPE_P(item))
+    {
+      free(working_array);
+      rb_raise(rb_eTypeError, "element is not a number");
+    }
+
+    working_array[i] = NUM2DBL(item);
+  }
+
+  // Reminder to myself as I'm learning C. Using an array as a function parameter decays that reference
+  // to a pointer to the first element in the array.
+  // https://www.gnu.org/software/gnu-c-manual/gnu-c-manual.html#Function-Parameters
+  qsort(working_array, array_length, sizeof(double), compare_doubles);
+
+  return working_array;
+}

data/ext/ruby_native_statistics/conversions.h

@@ -0,0 +1,5 @@
+#include <stdbool.h>
+#include <ruby.h>
+
+int compare_doubles(const void *a, const void *b);
+double *sorted_ruby_array(VALUE array, long array_length);

data/ext/ruby_native_statistics/dispersion.c

@@ -0,0 +1,102 @@
+#include "dispersion.h"
+
+VALUE rb_sample_standard_deviation(VALUE self)
+{
+  unsigned int array_length;
+
+  Check_Type(self, T_ARRAY);
+
+  array_length = rb_long2int(RARRAY_LEN(self));
+
+  if (array_length <= 1)
+  {
+    rb_raise(rb_eRangeError, "array must have more than one element");
+  }
+
+  return DBL2NUM(sqrt((calculate_total_distance_from_mean(self, array_length) / (array_length - 1))));
+}
+
+VALUE rb_sample_variance(VALUE self)
+{
+  unsigned int array_length;
+
+  Check_Type(self, T_ARRAY);
+
+  array_length = rb_long2int(RARRAY_LEN(self));
+
+  if (array_length <= 1)
+  {
+    rb_raise(rb_eRangeError, "array must have more than one element");
+  }
+
+  return DBL2NUM((calculate_total_distance_from_mean(self, array_length) / (array_length - 1)));
+}
+
+VALUE rb_population_standard_deviation(VALUE self)
+{
+  unsigned int array_length;
+
+  Check_Type(self, T_ARRAY);
+
+  array_length = rb_long2int(RARRAY_LEN(self));
+
+  if (array_length <= 1)
+  {
+    rb_raise(rb_eRangeError, "array must have more than one element");
+  }
+
+  return DBL2NUM(sqrt(calculate_total_distance_from_mean(self, array_length) / array_length));
+}
+
+VALUE rb_population_variance(VALUE self)
+{
+  unsigned int array_length;
+
+  Check_Type(self, T_ARRAY);
+
+  array_length = rb_long2int(RARRAY_LEN(self));
+
+  if (array_length <= 1)
+  {
+    rb_raise(rb_eRangeError, "array must have more than one element");
+  }
+
+  return DBL2NUM(calculate_total_distance_from_mean(self, array_length) / array_length);
+}
+
+VALUE rb_percentile(VALUE self, VALUE r_percentile)
+{
+  double result;
+  Check_Type(self, T_ARRAY);
+
+  long array_length = rb_array_len(self);
+  double percentile = NUM2DBL(r_percentile);
+
+  if (array_length == 0)
+  {
+    rb_raise(rb_eRangeError, "array must have at least one element");
+  }
+
+  if (percentile < 0 || percentile > 1)
+  {
+    rb_raise(rb_eRangeError, "percentile must be between 0 and 1 inclusive");
+  }
+
+  double *sorted_array = sorted_ruby_array(self, array_length);
+
+  double h = (array_length - 1) * percentile + 1;
+
+  if (trunc(h) == h)
+  {
+    result = sorted_array[(long)h - 1];
+  }
+  else
+  {
+    long h_floor = (long)trunc(h);
+    result = (h - h_floor) * (sorted_array[h_floor] - sorted_array[h_floor - 1]) + sorted_array[h_floor - 1];
+  }
+
+  free(sorted_array);
+
+  return DBL2NUM(result);
+}

data/ext/{dispersion → ruby_native_statistics}/dispersion.h

@@ -1,5 +1,10 @@
-VALUE DispersionModule = Qnil;
+#include <ruby.h>
+#include <math.h>
+#include "conversions.h"
+#include "mathematics.h"
+
 VALUE rb_sample_standard_deviation(VALUE self);
 VALUE rb_population_standard_deviation(VALUE self);
 VALUE rb_sample_variance(VALUE self);
 VALUE rb_population_variance(VALUE self);
+VALUE rb_percentile(VALUE self, VALUE percentile);

data/ext/ruby_native_statistics/extconf.rb

@@ -0,0 +1,9 @@
+require "mkmf"
+
+abort "missing pow()" unless have_func "pow"
+abort "missing sqrt()" unless have_func "sqrt"
+abort "missing malloc()" unless have_func "malloc"
+abort "missing free()" unless have_func "free"
+abort "missing trunc()" unless have_func "trunc"
+
+create_makefile "ruby_native_statistics/ruby_native_statistics"

data/ext/{mathematics → ruby_native_statistics}/mathematics.c

@@ -1,13 +1,5 @@
-#include "stdbool.h"
-#include "ruby.h"
-#include "mathematics.h"
 
-void Init_mathematics()
-{
-  MathematicsModule = rb_define_module("Mathematics");
-  rb_define_method(MathematicsModule, "mean", rb_mean, 0);
-  rb_define_method(MathematicsModule, "median", rb_median, 0);
-}
+#include "mathematics.h"
 
 double calculate_mean(VALUE array, unsigned long array_length)
 {
@@ -57,22 +49,10 @@ VALUE rb_mean(VALUE self)
   return DBL2NUM(calculate_mean(self, array_length));
 }
 
-int compare_doubles(const void *a, const void *b)
-{
-  double *dbl_a = (double *)a;
-  double *dbl_b = (double *)b;
-
-  double cmp_a = *dbl_a;
-  double cmp_b = *dbl_b;
-
-  return (cmp_a - cmp_b);
-}
-
 VALUE rb_median(VALUE self)
 {
   unsigned long array_length;
-  unsigned long i;
-  double *working_array;
+
   VALUE result;
 
   Check_Type(self, T_ARRAY);
@@ -87,30 +67,7 @@ VALUE rb_median(VALUE self)
   bool array_even_size = (array_length % 2) == 0;
   unsigned long middle = (long)floor(array_length / 2.0);
 
-  working_array = malloc(array_length * sizeof(double));
-
-  if (working_array == NULL)
-  {
-    rb_raise(rb_eStandardError, "unknown problem calculating median (possibly array is too large)");
-  }
-
-  for (i = 0; i < array_length; i++)
-  {
-    VALUE item = rb_ary_entry(self, i);
-
-    if (!RB_INTEGER_TYPE_P(item) && !RB_FLOAT_TYPE_P(item))
-    {
-      free(working_array);
-      rb_raise(rb_eTypeError, "element is not a number");
-    }
-
-    working_array[i] = NUM2DBL(item);
-  }
-
-  // Reminder to myself as I'm learning C. Using an array as a function parameter decays that reference
-  // to a pointer to the first element in the array.
-  // https://www.gnu.org/software/gnu-c-manual/gnu-c-manual.html#Function-Parameters
-  qsort(working_array, array_length, sizeof(double), compare_doubles);
+  double *working_array = sorted_ruby_array(self, array_length);
 
   if (!array_even_size)
   {

data/ext/{mathematics → ruby_native_statistics}/mathematics.h

@@ -1,5 +1,9 @@
-VALUE MathematicsModule = Qnil;
+#include <stdbool.h>
+#include <ruby.h>
+#include "conversions.h"
+
 VALUE rb_mean(VALUE self);
 VALUE rb_median(VALUE self);
+
 double calculate_mean(VALUE array, unsigned long array_length);
-double calculate_total_distance_from_mean(VALUE array, unsigned long array_length);
+double calculate_total_distance_from_mean(VALUE array, unsigned long array_length);

data/ext/ruby_native_statistics/ruby_native_statistics.c

@@ -0,0 +1,16 @@
+#include "ruby_native_statistics.h"
+
+void Init_ruby_native_statistics()
+{
+  DispersionModule = rb_define_module("Dispersion");
+  rb_define_method(DispersionModule, "stdev", rb_sample_standard_deviation, 0);
+  rb_define_method(DispersionModule, "stdevs", rb_sample_standard_deviation, 0);
+  rb_define_method(DispersionModule, "stdevp", rb_population_standard_deviation, 0);
+  rb_define_method(DispersionModule, "var", rb_sample_variance, 0);
+  rb_define_method(DispersionModule, "varp", rb_population_variance, 0);
+  rb_define_method(DispersionModule, "percentile", rb_percentile, 1);
+
+  MathematicsModule = rb_define_module("Mathematics");
+  rb_define_method(MathematicsModule, "mean", rb_mean, 0);
+  rb_define_method(MathematicsModule, "median", rb_median, 0);
+}

data/ext/ruby_native_statistics/ruby_native_statistics.h

@@ -0,0 +1,5 @@
+#include "dispersion.h"
+#include "mathematics.h"
+
+VALUE MathematicsModule = Qnil;
+VALUE DispersionModule = Qnil;

data/lib/ruby_native_statistics.rb

@@ -1,6 +1,5 @@
 require "ruby_native_statistics/version"
-require "mathematics"
-require "dispersion"
+require "ruby_native_statistics/ruby_native_statistics"
 
 class Array
   include Mathematics

data/lib/ruby_native_statistics/version.rb

@@ -1,3 +1,3 @@
 module RubyNativeStatistics
-  VERSION = "0.9.0"
+  VERSION = "0.10.0"
 end

data/ruby_native_statistics.gemspec

@@ -21,5 +21,5 @@ Gem::Specification.new do |spec|
   spec.files = `git ls-files -z`.split("\x0").reject { |f| f.match(%r{^(test|spec|features)/}) }
   spec.require_paths = ["lib"]
 
-  spec.extensions = %w[ext/mathematics/extconf.rb ext/dispersion/extconf.rb]
+  spec.extensions = %w[ext/ruby_native_statistics/extconf.rb]
 end

metadata

@@ -1,22 +1,21 @@
 --- !ruby/object:Gem::Specification
 name: ruby_native_statistics
 version: !ruby/object:Gem::Version
-  version: 0.9.0
+  version: 0.10.0
 platform: ruby
 authors:
 - Cory Buecker
 autorequire: 
 bindir: bin
 cert_chain: []
-date: 2020-01-12 00:00:00.000000000 Z
+date: 2020-01-20 00:00:00.000000000 Z
 dependencies: []
 description: 
 email:
 - cory.buecker@gmail.com
 executables: []
 extensions:
-- ext/mathematics/extconf.rb
-- ext/dispersion/extconf.rb
+- ext/ruby_native_statistics/extconf.rb
 extra_rdoc_files: []
 files:
 - ".github/workflows/main.yml"
@@ -29,12 +28,15 @@ files:
 - README.md
 - Rakefile
 - changelog.md
-- ext/dispersion/dispersion.c
-- ext/dispersion/dispersion.h
-- ext/dispersion/extconf.rb
-- ext/mathematics/extconf.rb
-- ext/mathematics/mathematics.c
-- ext/mathematics/mathematics.h
+- ext/ruby_native_statistics/conversions.c
+- ext/ruby_native_statistics/conversions.h
+- ext/ruby_native_statistics/dispersion.c
+- ext/ruby_native_statistics/dispersion.h
+- ext/ruby_native_statistics/extconf.rb
+- ext/ruby_native_statistics/mathematics.c
+- ext/ruby_native_statistics/mathematics.h
+- ext/ruby_native_statistics/ruby_native_statistics.c
+- ext/ruby_native_statistics/ruby_native_statistics.h
 - lib/ruby_native_statistics.rb
 - lib/ruby_native_statistics/version.rb
 - ruby_native_statistics.gemspec

data/ext/dispersion/dispersion.c

@@ -1,68 +0,0 @@
-#include "ruby.h"
-#include "dispersion.h"
-#include "../mathematics/mathematics.h"
-
-void Init_dispersion() {
-  DispersionModule = rb_define_module("Dispersion");
-  rb_define_method(DispersionModule, "stdev", rb_sample_standard_deviation, 0);
-  rb_define_method(DispersionModule, "stdevs", rb_sample_standard_deviation, 0);
-  rb_define_method(DispersionModule, "stdevp", rb_population_standard_deviation, 0);
-  rb_define_method(DispersionModule, "var", rb_sample_variance, 0);
-  rb_define_method(DispersionModule, "varp", rb_population_variance, 0);
-}
-
-VALUE rb_sample_standard_deviation(VALUE self) {
-  unsigned int array_length;
-
-  Check_Type(self, T_ARRAY);
-
-  array_length = rb_long2int(RARRAY_LEN(self));
-
-  if (array_length <= 1) {
-    rb_raise(rb_eRangeError, "array must have more than one element");
-  }
-
-  return DBL2NUM(sqrt((calculate_total_distance_from_mean(self, array_length)/(array_length - 1))));
-}
-
-VALUE rb_sample_variance(VALUE self) {
-  unsigned int array_length;
-
-  Check_Type(self, T_ARRAY);
-
-  array_length = rb_long2int(RARRAY_LEN(self));
-
-  if (array_length <= 1) {
-    rb_raise(rb_eRangeError, "array must have more than one element");
-  }
-
-  return DBL2NUM((calculate_total_distance_from_mean(self, array_length)/(array_length - 1)));
-}
-
-VALUE rb_population_standard_deviation(VALUE self) {
-  unsigned int array_length;
-
-  Check_Type(self, T_ARRAY);
-
-  array_length = rb_long2int(RARRAY_LEN(self));
-
-  if (array_length <= 1) {
-    rb_raise(rb_eRangeError, "array must have more than one element");
-  }
-
-  return DBL2NUM(sqrt(calculate_total_distance_from_mean(self, array_length) / array_length));
-}
-
-VALUE rb_population_variance(VALUE self) {
-  unsigned int array_length;
-
-  Check_Type(self, T_ARRAY);
-
-  array_length = rb_long2int(RARRAY_LEN(self));
-
-  if (array_length <= 1) {
-    rb_raise(rb_eRangeError, "array must have more than one element");
-  }
-
-  return DBL2NUM(calculate_total_distance_from_mean(self, array_length) / array_length);
-}

data/ext/dispersion/extconf.rb

@@ -1,3 +0,0 @@
-require "mkmf"
-
-create_makefile "dispersion"

data/ext/mathematics/extconf.rb

@@ -1,5 +0,0 @@
-require "mkmf"
-
-abort "missing pow()" unless have_func "pow"
-
-create_makefile "mathematics"