From 019718fbb3fd04111d9dbd5c6b356504f3b3485d Mon Sep 17 00:00:00 2001
From: Cory Buecker <email@corybuecker.com>
Date: Sat, 2 Mar 2024 20:22:10 -0600
Subject: [PATCH] fix bug with percentile function

Sorting arrays of small doubles was inconsistent with the previous method
---
 Gemfile.lock                             |  2 +-
 changelog.md                             |  4 ++++
 ext/ruby_native_statistics/conversions.c | 15 +++++++++++++--
 lib/ruby_native_statistics/version.rb    |  2 +-
 test/dispersion_test.rb                  | 15 +++++++++++++++
 5 files changed, 34 insertions(+), 4 deletions(-)

diff --git a/Gemfile.lock b/Gemfile.lock
index 2fe9dc7..40777c5 100644
--- a/Gemfile.lock
+++ b/Gemfile.lock
@@ -1,7 +1,7 @@
 PATH
   remote: .
   specs:
-    ruby_native_statistics (1.0.2)
+    ruby_native_statistics (1.1.0)
       rake-compiler (~> 1.2)
 
 GEM
diff --git a/changelog.md b/changelog.md
index 110d417..3cd0ee1 100644
--- a/changelog.md
+++ b/changelog.md
@@ -1,3 +1,7 @@
+# Version 1.1.0
+
+- Fix percentile bug reported by @Gbird22
+
 # Version 1.0.3
 
 - Update all supported Ruby versions
diff --git a/ext/ruby_native_statistics/conversions.c b/ext/ruby_native_statistics/conversions.c
index edf8c3e..f5928b3 100644
--- a/ext/ruby_native_statistics/conversions.c
+++ b/ext/ruby_native_statistics/conversions.c
@@ -1,4 +1,5 @@
 #include "conversions.h"
+#include "float.h"
 
 int compare_doubles(const void *a, const void *b)
 {
@@ -8,7 +9,17 @@ int compare_doubles(const void *a, const void *b)
   double cmp_a = *dbl_a;
   double cmp_b = *dbl_b;
 
-  return (cmp_a - cmp_b);
+  if (fabs(cmp_a - cmp_b) <= (DBL_EPSILON * fabs(cmp_a + cmp_b))) 
+  {
+    return 0;
+  }
+
+  if (cmp_a > cmp_b)
+  {
+    return 1;
+  }
+
+  return -1;
 }
 
 double *sorted_ruby_array(VALUE array, long array_length)
@@ -20,7 +31,7 @@ double *sorted_ruby_array(VALUE array, long array_length)
 
   if (working_array == NULL)
   {
-    rb_raise(rb_eStandardError, "unknown problem calculating median (possibly array is too large)");
+    rb_raise(rb_eStandardError, "unknown problem sorting array (possibly array is too large)");
   }
 
   for (i = 0; i < array_length; i++)
diff --git a/lib/ruby_native_statistics/version.rb b/lib/ruby_native_statistics/version.rb
index a4529dc..75f4d62 100644
--- a/lib/ruby_native_statistics/version.rb
+++ b/lib/ruby_native_statistics/version.rb
@@ -1,3 +1,3 @@
 module RubyNativeStatistics
-  VERSION = "1.0.3"
+  VERSION = "1.1.0"
 end
diff --git a/test/dispersion_test.rb b/test/dispersion_test.rb
index e0123ef..fa27ad4 100644
--- a/test/dispersion_test.rb
+++ b/test/dispersion_test.rb
@@ -129,4 +129,19 @@ def test_percentile_floats
     assert_in_delta 0.73895928, array.percentile(0.333), 0.000001
     assert_in_delta 11.43290852, array.percentile(0.928), 0.000001
   end
+
+  def test_percentile_repeating
+    array = [5.4, 5.3, 5.2, 5.4, 5.2].to_a.shuffle
+
+    assert_in_delta 5.4, array.percentile(0.9), 0.000001
+    assert_in_delta 5.2, array.percentile(0.1), 0.000001
+    assert_in_delta 5.3, array.percentile(0.5), 0.000001
+    assert_in_delta 5.26, array.percentile(0.4), 0.000001
+  end
+
+  def test_percentile_duplicates
+    array = [5.2, 5.2, 5.2, 5.2, 5.2].to_a.shuffle
+
+    assert_in_delta 5.2, array.percentile(0.9), 0.000001
+  end
 end