ruby-nuggets 0.9.6.pre1 → 0.9.6.pre2

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA1:
-  metadata.gz: bfa0c7f59b19e5581563b907497449211abbf7f1
-  data.tar.gz: 15e011ede9ab07f1c0cab0813f34e986e7415c00
+  metadata.gz: 204c7a58b06959ce60a857c5e70c60ca6e278e09
+  data.tar.gz: ecbcbaaa1a3ed0df93694678a4f42811dd406385
 SHA512:
-  metadata.gz: 8a2bc736534dca11da0d83c5eb6000615c0bd425b55cc666436073a39382ba3787ed92f643fcb4b508213f6f3ceac48acec16b0993d9309ace1359c4b987c501
-  data.tar.gz: 6076264bce920e79530637eaab28147813b940dec9221ee83d4e643b84cf3269cbf5e529d21fb7672e6a33367d38df53d2fcee202cda5e0b9517c902290ebffe
+  metadata.gz: 070ece04d3311e4eb79cf033e89f404ce79ed19027fadd1d6fd97ae9eeb9b6df1d6aea7dbf545ddd592d5a4950fdde74763a4111bc0bae7496bcc63638f330de
+  data.tar.gz: 18bda153c1274c40e83fb65149fc2306ea9581239ee4d3b7b3ea80d972e8da4cdb5a251471eb2ea76d1bc065a1cb80637fc40787664b4f04defddd42e430fe4b

data/lib/nuggets/array/flush.rb

@@ -0,0 +1,5 @@
+require 'nuggets/array/flush_mixin'
+
+class Array
+  include Nuggets::Array::FlushMixin
+end

data/lib/nuggets/array/flush_mixin.rb

@@ -0,0 +1,47 @@
+#--
+###############################################################################
+#                                                                             #
+# A component of ruby-nuggets, some extensions to the Ruby programming        #
+# language.                                                                   #
+#                                                                             #
+# Copyright (C) 2007-2013 Jens Wille                                          #
+#                                                                             #
+# Authors:                                                                    #
+#     Jens Wille <jens.wille@gmail.com>                                       #
+#                                                                             #
+# ruby-nuggets is free software; you can redistribute it and/or modify it     #
+# under the terms of the GNU Affero General Public License as published by    #
+# the Free Software Foundation; either version 3 of the License, or (at your  #
+# option) any later version.                                                  #
+#                                                                             #
+# ruby-nuggets is distributed in the hope that it will be useful, but WITHOUT #
+# ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or       #
+# FITNESS FOR A PARTICULAR PURPOSE. See the GNU Affero General Public License #
+# for more details.                                                           #
+#                                                                             #
+# You should have received a copy of the GNU Affero General Public License    #
+# along with ruby-nuggets. If not, see <http://www.gnu.org/licenses/>.        #
+#                                                                             #
+###############################################################################
+#++
+
+module Nuggets
+  class Array
+    module FlushMixin
+
+  # call-seq:
+  #   array.flush => anArray
+  #   array.flush { |array| ... } => anObject
+  #
+  # If a block is given, yields _array_ to the block and
+  # clears it afterwards; returns the return value of the
+  # block. Otherwise clears _array_ and returns its contents.
+  def flush
+    ret = block_given? ? yield(self) : dup
+    clear
+    ret
+  end
+
+    end
+  end
+end

data/lib/nuggets/array/mean_mixin.rb

@@ -4,7 +4,7 @@
 # A component of ruby-nuggets, some extensions to the Ruby programming        #
 # language.                                                                   #
 #                                                                             #
-# Copyright (C) 2007-2012 Jens Wille                                          #
+# Copyright (C) 2007-2013 Jens Wille                                          #
 #                                                                             #
 # Authors:                                                                    #
 #     Jens Wille <jens.wille@gmail.com>                                       #
@@ -145,7 +145,12 @@ module Nuggets
   # with the standard deviation, if available; all values are subject to
   # +precision+.
   #
-  # Examples:
+  # If _array_ is a flat array of numeric values, it is treated as a single
+  # "column".
+  #
+  # Returns +nil+ if _array_ is empty.
+  #
+  # Examples (with standard deviation):
   #
   #   [[9.4, 34.75], [9.46, 34.68], [9.51, 34.61]].report_mean
   #   #=> ["9.4567 +/- 0.0450", "34.6800 +/- 0.0572"]
@@ -158,7 +163,21 @@ module Nuggets
   #
   #   CSV.read('csv', headers: true, converters: :numeric).to_a.report_mean
   #   #=> ["a    9.4567 +/- 0.0450", "b    34.6800 +/- 0.0572"]
+  #
+  #   [9.4, 9.46, 9.51].report_mean
+  #   #=> ["9.4567 +/- 0.0450"]
+  #
+  #   [34.75, 34.68, 34.61].report_mean
+  #   #=> ["34.6800 +/- 0.0572"]
+  #
+  #   [].report_mean
+  #   #=> nil
   def report_mean(method = nil, precision = 4)
+    return if empty?
+
+    return clone.replace(self.class.new.push(self).transpose).
+      report_mean(method, precision) unless first.is_a?(self.class)
+
     met, sep = [method ||= :mean, 'mean'], ['', '_']
     lab, std = first.first.is_a?(::String), respond_to?(:std)
 

data/lib/nuggets/env/user_home_mixin.rb

@@ -34,6 +34,12 @@ module Nuggets
   #
   # Returns the user's home directory, or +default+ if it could not be found.
   def user_home(default = ::File::ALT_SEPARATOR ? 'C:/' : '/')
+    begin
+      return ::Dir.home
+    rescue ::ArgumentError
+      # "couldn't find HOME environment -- expanding `~'"
+    end if ::Dir.respond_to?(:home)
+
     %w[HOME HOMEDRIVE:HOMEPATH USERPROFILE APPDATA].each { |key|
       home = values_at(*key.split(':')).join
       return home.gsub(/\\/, '/') if home && !home.empty?

data/lib/nuggets/lsi.rb

@@ -25,32 +25,55 @@
 ###############################################################################
 #++
 
+require 'forwardable'
 require 'gsl'
 
 module Nuggets
 
   class LSI
 
-    include Enumerable
+    include ::Enumerable
 
-    DEFAULT_EPSILON = Float::EPSILON * 10
+    extend ::Forwardable
+
+    DEFAULT_EPSILON   = ::Float::EPSILON * 10
+
+    DEFAULT_PRECISION = 2
+
+    DEFAULT_TRANSFORM = :tfidf
+
+    DEFAULT_CUTOFF    = 0.75
+
+    class << self
+
+      def build(items, options = {})
+        lsi = new(items)
+        lsi if lsi.build(options)
+      end
+
+      def each_norm(items, options = {}, build_options = {}, &block)
+        lsi = new(items)
+        lsi.each_norm(nil, options, &block) if lsi.build(build_options)
+      end
 
-    def self.each_norm(items, options = {}, &block)
-      lsi = new(items)
-      lsi.each_norm(nil, options, &block) if lsi.build
     end
 
     def initialize(items = {})
       reset
-      items.each { |k, v| self[k] = v }
+      items.each { |k, v| self[k] = v || k }
     end
 
-    def [](key)
-      @hash[key]
-    end
+    def_delegators :@hash, :[], :each, :include?, :key, :keys, :size
+
+    def_delegator  :@hash, :values,    :docs
+    def_delegator  :@hash, :values_at, :docs_at
+
+    def_delegator  :@list, :keys, :terms
+
+    alias_method :doc, :[]
 
     def []=(key, value)
-      @hash[key] = Doc.new(key, value, @list)
+      @hash[key] = Doc.new(key, value, @list, @freq)
     end
 
     def add(key, value = key)
@@ -62,27 +85,13 @@ module Nuggets
       add(value.object_id, value)
     end
 
-    def size
-      @hash.size
-    end
-
-    def keys
-      @hash.keys
-    end
-
-    def docs
-      @hash.values
-    end
-
-    def each(&block)
-      @hash.each(&block)
-    end
-
     # min:: minimum value to consider
     # abs:: minimum absolute value to consider
     # nul:: exclude null values (true or Float)
     # new:: exclude original terms / only yield new ones
     def each_norm(key = nil, options = {})
+      return enum_for(:each_norm, key, options) unless block_given?
+
       min, abs, nul, new = options.values_at(:min, :abs, :nul, :new)
       nul = DEFAULT_EPSILON if nul == true
 
@@ -90,14 +99,14 @@ module Nuggets
 
       (key ? [self[key]] : docs).each { |doc|
         if doc && norm = doc.norm
-          i = 0
+          i = -1
 
           norm.each { |v|
+            i += 1
             yield doc, list[i], v unless (min && v < min) ||
                                          (abs && v.abs < abs) ||
                                          (nul && v.abs < nul) ||
                                          (new && doc.include?(i))
-            i += 1
           }
         end
       }
@@ -105,87 +114,160 @@ module Nuggets
 
     def related(key, num = 5)
       if doc = self[key] and norm = doc.norm
-        a = []; norm *= -1
-        each { |k, v| a << [norm * v.norm.col, k] unless k == key }
-        a.sort![0, num].map! { |_, k| k }
+        temp = sort_by { |k, v| -norm * v.norm.col }
+        temp.map! { |k,| k }.delete(key)
+        temp[0, num]
       end
     end
 
-    def build(cutoff = 0.75)
-      build!(docs, @list, cutoff) if size > 1
+    def related_score(key, num = 5)
+      if doc = self[key] and norm = doc.norm
+        temp = map { |k, v| [k, norm * v.norm.col] }.sort_by { |_, i| -i }
+        temp.delete(temp.assoc(key))
+        temp[0, num]
+      end
     end
 
-    def reset
-      @hash, @list, @invlist = {}, Hash.new { |h, k| h[k] = h.size }, {}
+    def build(options = {})
+      build!(docs, @list, options.is_a?(::Hash) ?
+        options : { :cutoff => options }) if size > 1
     end
 
-    private
+    def reset
+      @hash, @list, @freq, @invlist =
+        {}, ::Hash.new { |h, k| h[k] = h.size }, ::Hash.new(0), {}
+    end
 
-    def build!(docs, list, cutoff)
-      u, v, s = GSL::Matrix.alloc(*vectors(docs, list)).trans.SV_decomp
-      reduce(u, v, cutoff(s, cutoff), docs)
-      size
+    def inspect
+      '%s@%d/%d' % [self.class, size, @list.size]
     end
 
-    def vectors(docs, list)
-      @invlist, size = list.invert, list.size
-      docs.map { |doc| transform(doc.raw_vector(size)) }
+    def to_a(norm = true)
+      (norm ? map { |_, doc| doc.norm.to_a } :
+              map { |_, doc| doc.vector.to_a }).transpose
     end
 
-    # FIXME: "first-order association transform" ???
-    def transform(vec, q = 0)
-      return vec #unless (sum = vec.sum) > 1
+    private
+
+    def build!(docs, list, options)
+      Doc.transform = options.fetch(:transform, DEFAULT_TRANSFORM)
+
+      @invlist = list.invert
+
+      # TODO: GSL::ERROR::EUNIMPL: Ruby/GSL error code 24, svd of
+      # MxN matrix, M<N, is not implemented (file svd.c, line 61)
+      u, v, s = matrix(docs, list.size, size = docs.size).SV_decomp
 
-      vec.each { |v| q -= (w = v / sum) * Math.log(w) if v > 0 }
-      vec.map! { |v| Math.log(v + 1) / q }
+      r, i = reduce(s, options.fetch(:cutoff, DEFAULT_CUTOFF)), -1
+      (u * r * v.trans).each_col { |c| docs[i += 1].vector = c.row }
+
+      size
     end
 
-    def cutoff(s, c)
-      w, i = s.sort[-(s.size * c).round], 0
-      s.each { |v| s[i] = 0 if v < w; i += 1 }
-      s
+    def matrix(d = docs, m = @list.size, n = d.size)
+      x = ::GSL::Matrix.alloc(m, n)
+      d.each_with_index { |i, j| x.set_col(j, i.transformed_vector(m, n)) }
+      x
     end
 
-    def reduce(u, v, s, d, i = -1)
-      (u * GSL::Matrix.diagonal(s) * v.trans).each_col { |c|
-        d[i += 1].vector = c.row
-      }
+    # k == nil:: keep all
+    # k >= 1::   keep this many
+    # k < 1::    keep (at most) this proportion
+    def reduce(s, k, m = s.size)
+      if k && k < m
+        if k > 0
+          k = (m * k).floor if k < 1
+          s[k, m - k] = 0
+        else
+          s.set_zero
+        end
+      end
+
+      s.to_m_diagonal
     end
 
     class Doc
 
+      include ::Enumerable
+
+      extend ::Forwardable
+
       TOKEN_RE = %r{\s+}
 
-      def initialize(key, value, list)
-        @key = key
-        @map = !value.is_a?(Hash) ? build_hash(value, list) :
+      class << self
+
+        attr_reader :transform
+
+        def transform=(transform)
+          method = :transformed_vector
+
+          case transform
+            when ::Proc          then define_method(method, &transform)
+            when ::UnboundMethod then define_method(method, transform)
+            else alias_method(method, "#{transform ||= :raw}_vector")
+          end
+
+          @transform = transform
+        end
+
+      end
+
+      def initialize(key, value, list, freq)
+        @key, @list, @freq, @total = key, list, freq, 1
+
+        @map = !value.is_a?(::Hash) ? build_hash(value, list) :
           value.inject({}) { |h, (k, v)| h[list[k]] = v; h }
+
+        @map.each_key { |k| freq[k] += 1 }
+
+        self.vector = raw_vector
       end
 
       attr_reader :key, :vector, :norm
 
-      def raw_vector(size)
-        vec = GSL::Vector.alloc(size)
-        @map.each { |k, v| vec[k] = v }
+      def_delegators :@map, :each, :include?
+
+      def_delegator :raw_vector, :sum, :size
+
+      def raw_vector(size = @list.size, *)
+        vec = ::GSL::Vector.calloc(size)
+        each { |k, v| vec[k] = v }
         vec
       end
 
+      # TODO: "first-order association transform" ???
+      def foat_vector(*args)
+        vec, q = raw_vector(*args), 0
+        return vec unless (s = vec.sum) > 1
+
+        vec.each { |v| q -= (w = v / s) * ::Math.log(w) if v > 0 }
+        vec.map { |v| ::Math.log(v + 1) / q }
+      end
+
+      def tfidf_vector(*args)
+        vec, f, i = raw_vector(*args), @freq, -1
+        s, d = vec.sum, @total = args.fetch(1, @total).to_f
+        vec.map { |v| i += 1; v > 0 ? ::Math.log(d / f[i]) * v / s : v }
+      end
+
+      self.transform = DEFAULT_TRANSFORM
+
       def vector=(vec)
         @vector, @norm = vec, vec.normalize
       end
 
-      def include?(k)
-        @map.include?(k)
+      def inspect
+        '%s@%p/%d' % [self.class, key, size]
       end
 
       private
 
-      def build_hash(value, list, hash = Hash.new(0))
-        build_array(value).each { |i| hash[list[i]] += 1 }
+      def build_hash(value, list, hash = ::Hash.new(0))
+        build_enum(value).each { |i| hash[list[i]] += 1 }
         hash
       end
 
-      def build_array(value, re = TOKEN_RE)
+      def build_enum(value, re = TOKEN_RE)
         value = value.read if value.respond_to?(:read)
         value = value.split(re) if value.respond_to?(:split)
         value

data/lib/nuggets/midos.rb

@@ -221,9 +221,7 @@ module Nuggets
           record.each { |k, v|
             if v
               v = v.is_a?(::Array) ? v.join(vs) : v.to_s
-              v.gsub!("\n", nl)
-
-              target << k << fs << v << le
+              target << k << fs << v.gsub("\n", nl) << le
             end
           }
 

data/lib/nuggets/ruby.rb

@@ -259,15 +259,11 @@ module Nuggets
         opt = "-#{key.to_s[0, 1]}"
 
         if val.is_a?(::Array)
-          val.each { |wal|
-            argv << opt << wal.to_s
-          }
-        else
-          if opt == '-e'
-            argv << opt << val.to_s
-          elsif val != false
-            argv << "#{opt}#{val unless val == true}"
-          end
+          val.each { |wal| argv << opt << wal.to_s }
+        elsif opt == '-e'
+          argv << opt << val.to_s
+        elsif val != false
+          argv << "#{opt}#{val unless val == true}"
         end
       } if args.last.is_a?(::Hash)
 

data/lib/nuggets/uri/redirect_mixin.rb

@@ -64,7 +64,7 @@ module Nuggets
     return nil unless steps > 0
 
     follow_redirect(res['Location'], steps - 1, cache) { |*a| yield(*a) }
-  rescue ::SocketError, ::Errno::EHOSTUNREACH
+  rescue ::SocketError, ::Errno::EHOSTUNREACH, ::Errno::ENOENT
   end
 
   # call-seq:

data/lib/nuggets/version.rb

@@ -15,7 +15,7 @@ module Nuggets
 
       # Short-cut for version string.
       def to_s
-        to_a.join('.') + '.pre1'
+        to_a.join('.') + '.pre2'
       end
 
     end

data/spec/nuggets/array/flush_spec.rb

@@ -0,0 +1,45 @@
+require 'nuggets/array/flush'
+
+describe Array, 'when extended by', Nuggets::Array::FlushMixin do
+
+  it { Array.ancestors.should include(Nuggets::Array::FlushMixin) }
+
+  example do
+    a = [1, 2, 3, 4]
+    a.flush.should == [1, 2, 3, 4]
+    a.should be_empty
+  end
+
+  example do
+    a = [1, 2, 3, 4]
+    a.flush.push(42).should == [1, 2, 3, 4, 42]
+    a.should be_empty
+  end
+
+  example do
+    a = [1, 2, 3, 4]
+    a.flush { |b| b.should == [1, 2, 3, 4]; 42 }.should == 42
+    a.should be_empty
+  end
+
+  example do
+    a = [1, 2, 3, 4]
+    a.flush { |b| b }.should be_empty
+    a.should be_empty
+  end
+
+  example do
+    a, c = [1, 2, 3, 4], nil
+    a.flush { |b| c = b; c.should == [1, 2, 3, 4] }
+    a.should be_empty
+    c.should be_empty
+  end
+
+  example do
+    a, c = [1, 2, 3, 4], nil
+    a.flush { |b| c = b.dup; c.should == [1, 2, 3, 4] }
+    a.should be_empty
+    c.should == [1, 2, 3, 4]
+  end
+
+end

data/spec/nuggets/array/mean_spec.rb

@@ -1,4 +1,5 @@
 require 'nuggets/array/mean'
+require 'nuggets/array/standard_deviation'
 
 describe Array, 'when extended by', Nuggets::Array::MeanMixin do
 
@@ -77,6 +78,14 @@ describe Array, 'when extended by', Nuggets::Array::MeanMixin do
       [-3, -2, -1].arithmetic_mean.should == -[3, 2, 1].arithmetic_mean
     end
 
+    example do
+      [1e16, 1, 3, -1e16].arithmetic_mean.should == 1.0
+    end
+
+    example do
+      pending { [1e30, 1, 3, -1e30].arithmetic_mean.should == 1.0 }
+    end
+
     context do
 
       before :each do
@@ -124,12 +133,17 @@ describe Array, 'when extended by', Nuggets::Array::MeanMixin do
 
     before :each do
       @ary = [[9.4, 34.75], [9.46, 34.68], [9.51, 34.61]]
+
+      @ary1 = @ary.map(&:first)
+      @ary2 = @ary.map(&:last)
     end
 
     describe 'w/o stddev' do
 
       before :each do
-        class << @ary; undef_method :std; end if @ary.respond_to?(:std)
+        [@ary, @ary1, @ary2].each { |ary|
+          class << ary; undef_method :std; end if ary.respond_to?(:std)
+        }
       end
 
       example do
@@ -144,15 +158,22 @@ describe Array, 'when extended by', Nuggets::Array::MeanMixin do
         @ary.unshift(%w[a b]).report_mean(nil, 2).should == ['a  9.46', 'b  34.68']
       end
 
-    end
+      example do
+        @ary1.report_mean.should == ['9.4567']
+      end
 
-    describe 'w/ stddev' do
+      example do
+        @ary2.report_mean.should == ['34.6800']
+      end
 
-      before :each do
-        require 'nuggets/array/standard_deviation_mixin'
-        @ary.extend(Nuggets::Array::StandardDeviationMixin)
+      example do
+        [].report_mean.should be_nil
       end
 
+    end
+
+    describe 'w/ stddev' do
+
       example do
         @ary.report_mean.should == ['9.4567 +/- 0.0450', '34.6800 +/- 0.0572']
       end
@@ -165,6 +186,18 @@ describe Array, 'when extended by', Nuggets::Array::MeanMixin do
         @ary.unshift(%w[a b]).report_mean(nil, 2).should == ['a  9.46 +/- 0.04', 'b  34.68 +/- 0.06']
       end
 
+      example do
+        @ary1.report_mean.should == ['9.4567 +/- 0.0450']
+      end
+
+      example do
+        @ary2.report_mean.should == ['34.6800 +/- 0.0572']
+      end
+
+      example do
+        [].report_mean.should be_nil
+      end
+
     end
 
   end

data/spec/nuggets/array/variance_spec.rb

@@ -32,6 +32,22 @@ describe Array, 'when extended by', Nuggets::Array::VarianceMixin do
     [1, -2, 1, 2, 3, -4, 0, 3, 1, 2, 1, 0, 24].variance.should equal_float(42.0946745562130218)
   end
 
+  example do
+    [1, 2, 4, 5, 8].variance.should == 6.0
+  end
+
+  example do
+    [1, 2, 4, 5, 8].variance { |i| i + 1e7 }.should == 6.0
+  end
+
+  example do
+    pending { [1, 2, 4, 5, 8].variance { |i| i + 1e12 }.should == 6.0 }
+  end
+
+  example do
+    ([1, 2, 4, 5, 8] * 100).variance.should == 6.0
+  end
+
   example do
     [].covariance.should == 0.0
   end

metadata

@@ -1,14 +1,14 @@
 --- !ruby/object:Gem::Specification
 name: ruby-nuggets
 version: !ruby/object:Gem::Version
-  version: 0.9.6.pre1
+  version: 0.9.6.pre2
 platform: ruby
 authors:
 - Jens Wille
 autorequire: 
 bindir: bin
 cert_chain: []
-date: 2013-07-24 00:00:00.000000000 Z
+date: 2013-11-25 00:00:00.000000000 Z
 dependencies: []
 description: Some extensions to the Ruby programming language.
 email: jens.wille@gmail.com
@@ -31,6 +31,8 @@ files:
 - lib/nuggets/array/correlation.rb
 - lib/nuggets/array/correlation_mixin.rb
 - lib/nuggets/array/flatten_once.rb
+- lib/nuggets/array/flush.rb
+- lib/nuggets/array/flush_mixin.rb
 - lib/nuggets/array/format.rb
 - lib/nuggets/array/histogram.rb
 - lib/nuggets/array/histogram_mixin.rb
@@ -182,6 +184,7 @@ files:
 - Rakefile
 - spec/nuggets/array/boost_spec.rb
 - spec/nuggets/array/correlation_spec.rb
+- spec/nuggets/array/flush_spec.rb
 - spec/nuggets/array/histogram_spec.rb
 - spec/nuggets/array/limit_spec.rb
 - spec/nuggets/array/mean_spec.rb
@@ -218,36 +221,36 @@ files:
 - spec/nuggets/uri/content_type_spec.rb
 - spec/nuggets/uri/exist_spec.rb
 - spec/spec_helper.rb
-- .rspec
+- ".rspec"
 homepage: http://github.com/blackwinter/ruby-nuggets
 licenses:
 - AGPL
 metadata: {}
 post_install_message: 
 rdoc_options:
-- --charset
+- "--charset"
 - UTF-8
-- --line-numbers
-- --all
-- --title
-- ruby-nuggets Application documentation (v0.9.6.pre1)
-- --main
+- "--line-numbers"
+- "--all"
+- "--title"
+- ruby-nuggets Application documentation (v0.9.6.pre2)
+- "--main"
 - README
 require_paths:
 - lib
 required_ruby_version: !ruby/object:Gem::Requirement
   requirements:
-  - - '>='
+  - - ">="
     - !ruby/object:Gem::Version
       version: '0'
 required_rubygems_version: !ruby/object:Gem::Requirement
   requirements:
-  - - '>'
+  - - ">"
     - !ruby/object:Gem::Version
       version: 1.3.1
 requirements: []
 rubyforge_project: 
-rubygems_version: 2.0.5
+rubygems_version: 2.1.11
 signing_key: 
 specification_version: 4
 summary: Some extensions to the Ruby programming language.