enumerable-statistics 2.0.1 → 2.0.7

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA256:
-  metadata.gz: e593e551f4f9b857a009db71093abab619e96680c4d15fc0eaa3a96c6e622bd4
-  data.tar.gz: eb449c789843e7e3e2c435890b8a9281d49cb6cdaad4ebbfc979b959413f617f
+  metadata.gz: caf1b8203b177aef325480c7d7d5a4c0ddb7cecdad3d12944d4b2caa71eebe13
+  data.tar.gz: 00d91f25e099eb2577eb1f6078dafb5c884538fe3863eb1b6795de58c0c3b642
 SHA512:
-  metadata.gz: 8866ba9d47c0715b82ffe7370b1fe8f683e1c1a17fd2dfee50db06c536cfa25b469bbf94a0f23c24ce2179c9d81203b761f8d17beadf113752b4f89490a48a2f
-  data.tar.gz: 942e4dacc202cef906e631c38d9450045f9712ffed811b35f7f28fafeecc0e92f70196b0d8d8053423266d60210a1f556ed600d4c2c2b0ee5320aa97830cab47
+  metadata.gz: 2c9ab8a29752eb46cd00bf8f3f9787d6f151bc10e06d2179fa5f80e3ebb4625b96144a49fd907eb52c02e8f88767f9da1717761643d87e1489677a9eab227af9
+  data.tar.gz: de2d3d348ead4d6e99cffdb66f95a56ac7cfe603a4450db23deaddb6eeb82e454f79e84c236ea89142ea334999a82c6c307e16ac7085b1afe2895c63a6a8c5d4

data/.github/workflows/ci.yml

@@ -0,0 +1,47 @@
+name: CI
+
+on:
+- push
+
+jobs:
+  cruby:
+    name: Test
+    runs-on: ${{ matrix.os }}
+
+    strategy:
+      fail-fast: false
+      matrix:
+        os:
+        - ubuntu-latest
+        - macos-latest
+        - windows-latest
+        ruby:
+        - 3.0
+        - 2.7
+        - 2.6
+        - 2.5
+        - 2.4
+        - debug
+        exclude:
+        - os: windows-latest
+          ruby: 3.0
+        - os: windows-latest
+          ruby: debug
+
+    steps:
+    - uses: actions/checkout@v2
+
+    - name: Setup Ruby
+      uses: ruby/setup-ruby@v1
+      with:
+        ruby-version: ${{ matrix.ruby }}
+
+    - run: gem install bundler
+    - run: bundle install
+
+    - run: rake --trace compile
+
+    - run: rake build
+    - run: gem install pkg/*gem
+
+    - run: rake

data/CHANGELOG.md

@@ -1,3 +1,29 @@
+# 2.0.7
+
+- Fix the bug of histogram with bin range that is smaller than value range
+
+# 2.0.6
+
+- Add edges parameter in histogram
+- Rename parameter in histogram to fix typo: `weight` to `weights`
+
+# 2.0.5
+
+- Add weighted histogram support
+
+# 2.0.4
+
+- Add `find_min`, `find_max`, `argmin`, `argmax` methods
+- Fix `nbin=:auto` case in `histogram` method
+
+# 2.0.3
+
+- Ractor-safe version
+
+# 2.0.2
+
+- Support Ruby 3.0
+
 # 2.0.1
 
 - Fix a bug of `histogram` (#9)

data/Rakefile

@@ -8,7 +8,14 @@ Rake::ExtensionTask.new('enumerable/statistics/extension')
 
 directory 'lib/enumerable/statistics'
 
+desc "Run test"
+task :test do
+  ruby("test/run-test.rb")
+end
+task :test => :compile
+
 RSpec::Core::RakeTask.new(:spec)
+task :spec => :compile
 
 task :bench do
   puts "# sum\n"
@@ -20,3 +27,5 @@ task :bench do
   puts "# variance\n"
   system('benchmark-driver bench/variance.yml')
 end
+
+task default: [:test, :spec]

data/enumerable-statistics.gemspec

@@ -27,10 +27,11 @@ Gem::Specification.new do |spec|
 
   spec.required_ruby_version = '>= 2.4'
 
-  spec.add_development_dependency "bundler", ">= 1.17.2"
+  spec.add_development_dependency "bundler"
   spec.add_development_dependency "rake"
   spec.add_development_dependency "rake-compiler", ">= 0.9.8"
   spec.add_development_dependency "rspec", ">= 3.4"
+  spec.add_development_dependency "test-unit"
   spec.add_development_dependency "fuubar"
   spec.add_development_dependency "yard"
   spec.add_development_dependency "benchmark-driver"

data/ext/enumerable/statistics/extension/array_ext.c

@@ -0,0 +1,60 @@
+#include <ruby/ruby.h>
+
+static VALUE
+ary_find_max(VALUE ary)
+{
+  const long n = RARRAY_LEN(ary);
+  if (n == 0) {
+    return Qnil;
+  }
+
+  long imax = 0;
+  VALUE max = RARRAY_AREF(ary, imax);
+
+  long i;
+  for (i = 1; i < n; ++i) {
+    VALUE v = RARRAY_AREF(ary, i);
+    if (RTEST(rb_funcall(v, '>', 1, max))) {
+      imax = i;
+      max = v;
+    }
+  }
+
+  return rb_assoc_new(max, LONG2NUM(imax));
+}
+
+static VALUE
+ary_find_min(VALUE ary)
+{
+  const long n = RARRAY_LEN(ary);
+  if (n == 0) {
+    return Qnil;
+  }
+
+  long imin = 0;
+  VALUE min = RARRAY_AREF(ary, imin);
+
+  long i;
+  for (i = 1; i < n; ++i) {
+    VALUE v = RARRAY_AREF(ary, i);
+    if (RTEST(rb_funcall(v, '<', 1, min))) {
+      imin = i;
+      min = v;
+    }
+  }
+
+  return rb_assoc_new(min, LONG2NUM(imin));
+}
+
+void
+Init_array_extension(void)
+{
+  VALUE mEnumerableStatistics = rb_const_get_at(rb_cObject, rb_intern("EnumerableStatistics"));
+  VALUE mArrayExtension = rb_const_get_at(mEnumerableStatistics, rb_intern("ArrayExtension"));
+
+  rb_undef_method(mArrayExtension, "find_max");
+  rb_define_method(mArrayExtension, "find_max", ary_find_max, 0);
+
+  rb_undef_method(mArrayExtension, "find_min");
+  rb_define_method(mArrayExtension, "find_min", ary_find_min, 0);
+}

data/ext/enumerable/statistics/extension/statistics.c

@@ -44,7 +44,7 @@ struct RRational {
 #endif
 
 #ifndef RRATIONAL
-# define RRATIONAL(obj) (R_CAST(RRational)(obj))
+# define RRATIONAL(obj) ((struct RRational *)(obj))
 #endif
 
 #ifndef RRATIONAL_SET_NUM
@@ -64,7 +64,7 @@ struct RComplex {
 #endif
 
 #ifndef RCOMPLEX
-# define RCOMPLEX(obj) (R_CAST(RComplex)(obj))
+# define RCOMPLEX(obj) ((struct RComplex *)(obj))
 #endif
 
 #ifndef RCOMPLEX_SET_REAL
@@ -97,7 +97,7 @@ static ID idPow, idPLUS, idMINUS, idSTAR, idDIV, idGE;
 static ID id_eqeq_p, id_idiv, id_negate, id_to_f, id_cmp, id_nan_p;
 static ID id_each, id_real_p, id_sum, id_population, id_closed, id_edge;
 
-static VALUE sym_left, sym_right;
+static VALUE sym_auto, sym_left, sym_right;
 
 static VALUE cHistogram;
 
@@ -2102,50 +2102,67 @@ histogram_edge_bin_index(VALUE edge, VALUE rb_x, int left_p)
 }
 
 static void
-histogram_weights_push_values(VALUE weights, VALUE edge, VALUE values, int left_p)
+histogram_weights_push_values(VALUE bin_weights, VALUE edge, VALUE values, VALUE weight_array, int left_p)
 {
+  const VALUE one = INT2FIX(1);
+  long bi, i, n, n_bins, weighted = 0;
   VALUE x, cur;
-  long i, n, bi;
+
+  assert(RB_TYPE_P(edge, T_ARRAY));
+  assert(RB_TYPE_P(values, T_ARRAY));
 
   n = RARRAY_LEN(values);
+  n_bins = RARRAY_LEN(edge) - 1;
+
+  if (! NIL_P(weight_array)) {
+    assert(RB_TYPE_P(weight_array, T_ARRAY));
+    assert(RARRAY_LEN(weight_array) == n);
+    weighted = 1;
+  }
+
   for (i = 0; i < n; ++i) {
     x = RARRAY_AREF(values, i);
 
-    bi = histogram_edge_bin_index(edge, x, left_p);
-
-    cur = rb_ary_entry(weights, bi);
-    if (NIL_P(cur)) {
-      cur = INT2FIX(1);
+    VALUE w;
+    if (weighted) {
+      w = RARRAY_AREF(weight_array, i);
+      if (RB_TYPE_P(w, T_COMPLEX)) {
+        VALUE imag = RCOMPLEX(w)->imag;
+        if (! RTEST(f_zero_p(imag))) {
+          goto type_error;
+        }
+      }
+      else if (rb_obj_is_kind_of(w, rb_cNumeric)) {
+        if (!RTEST(f_real_p(w))) {
+          goto type_error;
+        }
+      }
+      else {
+        goto type_error;
+      }
     }
     else {
-      cur = rb_funcall(cur, idPLUS, 1, INT2FIX(1));
+      w = one;
     }
 
-    rb_ary_store(weights, bi, cur);
-  }
-}
+    bi = histogram_edge_bin_index(edge, x, left_p);
 
-static int
-opt_closed_left_p(VALUE opts)
-{
-  int left_p = 1;
+    if (0 <= bi && bi < n_bins) {
+      cur = rb_ary_entry(bin_weights, bi);
+      if (NIL_P(cur)) {
+        cur = w;
+      }
+      else {
+        cur = rb_funcall(cur, idPLUS, 1, w);
+      }
 
-  if (!NIL_P(opts)) {
-    VALUE closed;
-#ifdef HAVE_RB_GET_KWARGS
-    ID kwargs = id_closed;
-    rb_get_kwargs(opts, &kwargs, 0, 1, &closed);
-#else
-    closed = rb_hash_lookup2(opts, ID2SYM(id_closed), sym_left);
-#endif
-    left_p = (closed != sym_right);
-    if (left_p && closed != sym_left) {
-      rb_raise(rb_eArgError, "invalid value for :closed keyword "
-               "(%"PRIsVALUE" for :left or :right)", closed);
+      rb_ary_store(bin_weights, bi, cur);
     }
   }
+  return;
 
-  return left_p;
+type_error:
+      rb_raise(rb_eTypeError, "weight array must have only real numbers");
 }
 
 static inline long
@@ -2269,10 +2286,47 @@ ary_histogram_calculate_edge(VALUE ary, const long nbins, const int left_p)
   return edge;
 }
 
+static VALUE
+check_histogram_weight_array(VALUE weight_array, const long ary_len)
+{
+  if (weight_array == Qundef) return Qnil;
+
+  weight_array = rb_check_convert_type(weight_array, T_ARRAY, "Array", "to_ary");
+  if (RARRAY_LEN(weight_array) != ary_len) {
+    rb_raise(rb_eArgError, "weight array must have the same number of items as the receiver array");
+  }
+
+  return weight_array;
+}
+
+static VALUE
+check_histogram_edges(VALUE edges)
+{
+  if (edges == Qundef) return Qnil;
+  return rb_check_convert_type(edges, T_ARRAY, "Array", "to_ary");
+}
+
+static int
+check_histogram_left_p(VALUE closed)
+{
+  int left_p = (closed != sym_right);
+  if (left_p && closed != Qundef && closed != sym_left) {
+    rb_raise(rb_eArgError, "invalid value for :closed keyword "
+        "(%"PRIsVALUE" for :left or :right)", closed);
+  }
+  return left_p;
+}
+
 /* call-seq:
- *    ary.histogram(nbins=:auto, closed: :left)
+ *    ary.histogram(nbins=:auto, weight: nil, closed: :left)
  *
  * @param [Integer] nbins  The approximate number of bins
+ * @params [Array<Numeric>] weights
+ *   An optional weight array, that has the same length as the receiver.
+ *   `weight[i]` means the weight value of the i-th element in the receiver.
+ * @params [Array<Numeric>] edges
+ *   An optional edge array, that specify the bin edges.
+ *   This array must be sorted.
  * @param [:left, :right] closed
  *   If :left (the default), the bin interval are left-closed.
  *   If :right, the bin interval are right-closed.
@@ -2282,30 +2336,52 @@ ary_histogram_calculate_edge(VALUE ary, const long nbins, const int left_p)
 static VALUE
 ary_histogram(int argc, VALUE *argv, VALUE ary)
 {
-  VALUE arg0, opts, edge, weights;
-  int left_p;
-  long nbins, nweights, i;
+  VALUE arg0, kwargs, bin_weights;
+  long nbins, n_bin_weights, i;
 
-  rb_scan_args(argc, argv, "01:", &arg0, &opts);
-  if (NIL_P(arg0)) {
+  VALUE weight_array = Qnil;
+  VALUE edges = Qnil;
+  int left_p = 1;
+
+  rb_scan_args(argc, argv, "01:", &arg0, &kwargs);
+  if (NIL_P(arg0) || arg0 == sym_auto) {
     nbins = sturges(RARRAY_LEN(ary));
   }
   else {
     nbins = NUM2LONG(arg0);
   }
-  left_p = opt_closed_left_p(opts);
 
-  edge = ary_histogram_calculate_edge(ary, nbins, left_p);
+  if (!NIL_P(kwargs)) {
+    enum { kw_weights, kw_edges, kw_closed };
+    static ID kwarg_keys[3];
+    VALUE kwarg_vals[3];
+
+    if (!kwarg_keys[0]) {
+      kwarg_keys[kw_weights] = rb_intern("weights");
+      kwarg_keys[kw_edges]  = rb_intern("edges");
+      kwarg_keys[kw_closed] = rb_intern("closed");
+    }
+
+    rb_get_kwargs(kwargs, kwarg_keys, 0, 3, kwarg_vals);
+
+    weight_array = check_histogram_weight_array(kwarg_vals[kw_weights], RARRAY_LEN(ary));
+    edges = check_histogram_edges(kwarg_vals[kw_edges]);
+    left_p = check_histogram_left_p(kwarg_vals[kw_closed]);
+  }
+
+  if (NIL_P(edges)) {
+    edges = ary_histogram_calculate_edge(ary, nbins, left_p);
+  }
 
-  nweights = RARRAY_LEN(edge) - 1;
-  weights = rb_ary_new_capa(nweights);
-  for (i = 0; i < nweights; ++i) {
-    rb_ary_store(weights, i, INT2FIX(0));
+  n_bin_weights = RARRAY_LEN(edges) - 1;
+  bin_weights = rb_ary_new_capa(n_bin_weights);
+  for (i = 0; i < n_bin_weights; ++i) {
+    rb_ary_store(bin_weights, i, INT2FIX(0));
   }
 
-  histogram_weights_push_values(weights, edge, ary, left_p);
+  histogram_weights_push_values(bin_weights, edges, ary, weight_array, left_p);
 
-  return rb_struct_new(cHistogram, edge, weights,
+  return rb_struct_new(cHistogram, edges, bin_weights,
                        left_p ? sym_left : sym_right,
                        Qfalse);
 }
@@ -2315,6 +2391,10 @@ Init_extension(void)
 {
   VALUE mEnumerableStatistics;
 
+#ifdef HAVE_RB_EXT_RACTOR_SAFE
+  rb_ext_ractor_safe(true);
+#endif
+
 #ifndef HAVE_ENUM_SUM
   rb_define_method(rb_mEnumerable, "sum", enum_sum, -1);
 #endif
@@ -2348,6 +2428,9 @@ Init_extension(void)
 
   rb_define_method(rb_cArray, "histogram", ary_histogram, -1);
 
+  void Init_array_extension(void);
+  Init_array_extension();
+
   idPLUS = '+';
   idMINUS = '-';
   idSTAR = '*';
@@ -2367,6 +2450,7 @@ Init_extension(void)
   id_closed = rb_intern("closed");
   id_edge = rb_intern("edge");
 
+  sym_auto = ID2SYM(rb_intern("auto"));
   sym_left = ID2SYM(rb_intern("left"));
   sym_right = ID2SYM(rb_intern("right"));
 }

data/lib/enumerable_statistics.rb

@@ -1,2 +1,3 @@
 require_relative "enumerable_statistics/version"
+require_relative "enumerable_statistics/array_ext"
 require_relative "enumerable_statistics/histogram"

data/lib/enumerable_statistics/array_ext.rb

@@ -0,0 +1,37 @@
+module EnumerableStatistics
+  module ArrayExtension
+    def find_max
+      n = size
+      return nil if n == 0
+
+      imax, i = 0, 1
+      while i < n
+        imax = i if self[i] > self[imax]
+        i += 1
+      end
+      [self[imax], imax]
+    end
+
+    def argmax
+      find_max[1]
+    end
+
+    def find_min
+      n = size
+      return nil if n == 0
+
+      imin, i = 0, 1
+      while i < n
+        imin = i if self[i] < self[imax]
+        i += 1
+      end
+      [self[imin], imin]
+    end
+
+    def argmin
+      find_min[1]
+    end
+  end
+
+  Array.include ArrayExtension
+end

data/lib/enumerable_statistics/histogram.rb

@@ -1,5 +1,6 @@
 module EnumerableStatistics
-  class Histogram < Struct.new(:edge, :weights, :closed, :isdensity)
+  class Histogram < Struct.new(:edges, :weights, :closed, :isdensity)
+    alias edge edges
     alias density? isdensity
   end
 end

data/lib/enumerable_statistics/version.rb

@@ -1,5 +1,5 @@
 module EnumerableStatistics
-  VERSION = '2.0.1'
+  VERSION = '2.0.7'
 
   module Version
     numbers, TAG = VERSION.split('-', 2)

metadata

@@ -1,14 +1,14 @@
 --- !ruby/object:Gem::Specification
 name: enumerable-statistics
 version: !ruby/object:Gem::Version
-  version: 2.0.1
+  version: 2.0.7
 platform: ruby
 authors:
 - Kenta Murata
 autorequire: 
 bindir: exe
 cert_chain: []
-date: 2020-01-11 00:00:00.000000000 Z
+date: 2021-06-24 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: bundler
@@ -16,14 +16,14 @@ dependencies:
     requirements:
     - - ">="
       - !ruby/object:Gem::Version
-        version: 1.17.2
+        version: '0'
   type: :development
   prerelease: false
   version_requirements: !ruby/object:Gem::Requirement
     requirements:
     - - ">="
       - !ruby/object:Gem::Version
-        version: 1.17.2
+        version: '0'
 - !ruby/object:Gem::Dependency
   name: rake
   requirement: !ruby/object:Gem::Requirement
@@ -66,6 +66,20 @@ dependencies:
     - - ">="
       - !ruby/object:Gem::Version
         version: '3.4'
+- !ruby/object:Gem::Dependency
+  name: test-unit
+  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: fuubar
   requirement: !ruby/object:Gem::Requirement
@@ -114,12 +128,11 @@ email:
 executables: []
 extensions:
 - ext/enumerable/statistics/extension/extconf.rb
-- ext/-bench-/extconf.rb
 extra_rdoc_files: []
 files:
+- ".github/workflows/ci.yml"
 - ".gitignore"
 - ".rspec"
-- ".travis.yml"
 - ".yardopts"
 - CHANGELOG.md
 - Gemfile
@@ -136,12 +149,13 @@ files:
 - bin/rspec
 - bin/setup
 - enumerable-statistics.gemspec
-- ext/-bench-/extconf.rb
+- ext/enumerable/statistics/extension/array_ext.c
 - ext/enumerable/statistics/extension/extconf.rb
 - ext/enumerable/statistics/extension/statistics.c
 - images/benchmark.png
 - lib/enumerable/statistics.rb
 - lib/enumerable_statistics.rb
+- lib/enumerable_statistics/array_ext.rb
 - lib/enumerable_statistics/histogram.rb
 - lib/enumerable_statistics/version.rb
 - templates/default/layout/html/headers.erb
@@ -275,7 +289,7 @@ required_rubygems_version: !ruby/object:Gem::Requirement
     - !ruby/object:Gem::Version
       version: '0'
 requirements: []
-rubygems_version: 3.1.2
+rubygems_version: 3.2.21
 signing_key: 
 specification_version: 4
 summary: Statistics features for Enumerable

data/.travis.yml

@@ -1,32 +0,0 @@
----
-notification:
-  email:
-  - mrkn@ruby-lang.org
-
-language: ruby
-
-before_install:
-  - yes | gem update --system
-  - gem install bundler
-
-install:
-  - bundle install
-
-script:
-  - bundle exec rake --trace clobber compile
-  - bundle exec rake spec
-
-matrix:
-  include:
-    - name: "2.3"
-      rvm: 2.3
-    - name: "2.4"
-      rvm: 2.4.5
-    - name: "2.5"
-      rvm: 2.5.2
-    - name: "2.6"
-      rvm: 2.6
-    - name: "trunk"
-      rvm: ruby-head
-  allow_failures:
-    - rvm: 2.3

data/ext/-bench-/extconf.rb

@@ -1,3 +0,0 @@
-require 'mkmf'
-
-create_makefile('-bench-')