pearson-hashing 0.0.3 → 0.0.4

data/benchmark/collisions.rb

@@ -2,7 +2,7 @@
 
 require File.expand_path('../../lib/pearson-hashing', __FILE__)
 
-SIZE = 100_000
+SIZE = 512**2
 
 def benchmark(name, &block)
   puts "#{name}.."
@@ -21,11 +21,15 @@ end
 benchmark 'PearsonHashing#digest8' do |str|
   PearsonHashing.digest8 str
 end
-  
+
 benchmark 'PearsonHashing#digest16' do |str|
   PearsonHashing.digest16 str
 end
 
+benchmark 'PearsonHashing#digest32' do |str|
+  PearsonHashing.digest32 str
+end
+
 benchmark 'String#hash' do |str|
   str.hash
 end

data/benchmark/speed.rb

@@ -3,21 +3,29 @@
 require 'benchmark'
 require File.expand_path('../../lib/pearson-hashing', __FILE__)
 
+SIZE = 512**2
+
 Benchmark.bm do |b|
   b.report 'PearsonHashing#digest8' do
-    100_000.times do |i|
+    SIZE.times do |i|
       PearsonHashing.digest8 i.to_s
     end
   end
   
   b.report 'PearsonHashing#digest16' do
-    100_000.times do |i|
+    SIZE.times do |i|
       PearsonHashing.digest16 i.to_s
     end
   end
   
+  b.report 'PearsonHashing#digest32' do
+    SIZE.times do |i|
+      PearsonHashing.digest32 i.to_s
+    end
+  end
+  
   b.report 'String#hash' do
-    100_000.times do |i|
+    SIZE.times do |i|
       i.to_s.hash
     end
   end

data/lib/pearson-hashing.rb

@@ -1,6 +1,5 @@
 require File.join(File.dirname(__FILE__), 'pearson-hashing/version')
 
-
 module PearsonHashing
   TABLE = [49, 118,  63, 252,  13, 155, 114, 130, 137,  40, 210,  62, 219, 246, 136, 221,
            174, 106,  37, 227, 166,  25, 139,  19, 204, 212,  64, 176,  70,  11, 170,  58,
@@ -36,6 +35,8 @@ module PearsonHashing
     end
     hash
   end
+  
+  # make #digest8 the default
   class << self
     alias :digest :digest8
   end
@@ -47,9 +48,48 @@ module PearsonHashing
   # @param [String] string
   # @return [Fixnum] hash
   def self.digest16(string)
-    h1 = PearsonHashing.digest(string)
-    string2 = [((string.bytes.first+1)%256)].pack('U*') + string[1,string.size]
-    h2 = PearsonHashing.digest(string2)
-    ("%03d" % h1 + "%03d" % h2).to_i
+    h1 = PearsonHashing.digest8 string
+    h2 = PearsonHashing.digest8 shift string
+    format = '%03d'
+    (format % h1 + format % h2).to_i
+  end
+  
+  # 32 bit hash
+  # @param [String] string
+  # @param [Fixnum] hashvalue
+  def self.digest32(string)
+    h1 = PearsonHashing.digest8 string
+    h2 = PearsonHashing.digest8 shift string
+    h3 = PearsonHashing.digest8 shift shift string
+    format = '%03d'
+    (format % h1 + format % h2 + format % h3).to_i
+  end
+  
+  # this is a dynamic implementation just to enable longer hashes
+  # but known to be slower than the hardcoded methods above
+  #
+  # 1 byte = 8 bit
+  # 2 bytes = 16 bit
+  # ...
+  #
+  # @param [String] string
+  # @param [Fixnum] no of bytes
+  # @return [Fixnum] hashvalue
+  def self.digest_with_x_bytes(string, no_of_bytes)
+    tmp_str = string.dup
+    result = ''
+    no_of_bytes.times do
+      result = result + "%03d" % PearsonHashing.digest8(tmp_str)
+      tmp_str = shift(tmp_str)
+    end
+    result.to_i
+  end
+  
+private
+  
+  # @param [String] str
+  # @return [String] str with first byte moved up
+  def self.shift(str)
+    [(str.bytes.first+1)%256].pack('U*') + str[1,str.size]
   end
 end

data/lib/pearson-hashing/version.rb

@@ -1,3 +1,3 @@
 module PearsonHashing
-  VERSION = "0.0.3"
+  VERSION = "0.0.4"
 end

data/spec/lib/pearson_hashing_spec.rb

@@ -1,6 +1,12 @@
 require 'spec_helper'
 
 describe PearsonHashing do
+  describe '#digest' do
+    it 'should return the same as #digest8' do
+      PearsonHashing.digest('foo').should == PearsonHashing.digest8('foo')
+    end
+  end
+  
   describe '#digest8' do
     it 'should return 96 for "foo"' do
       PearsonHashing.digest('foo').should == 96
@@ -13,9 +19,25 @@ describe PearsonHashing do
     end
   end
   
-  describe '#digest' do
-    it 'should return the same as #digest8' do
-      PearsonHashing.digest('foo').should == PearsonHashing.digest8('foo')
+  describe '#digest32' do
+    it 'should return 96 for "foo"' do
+      PearsonHashing.digest32('foo').should == 96099141
+    end
+  end
+  
+  describe '#digest_with_x_bytes' do
+    it 'should equal #digest8 with 1 byte' do
+      PearsonHashing.digest8('foo').should == \
+        PearsonHashing.digest_with_x_bytes('foo', 1)
+    end
+    it 'should equal #digest16 with 2 byte' do
+      PearsonHashing.digest16('foo').should == \
+        PearsonHashing.digest_with_x_bytes('foo', 2)
+    end
+    it 'should equal #digest21 with 3 byte' do
+      PearsonHashing.digest32('foo').should == \
+        PearsonHashing.digest_with_x_bytes('foo', 3)
     end
   end
+  
 end

metadata

@@ -1,7 +1,7 @@
 --- !ruby/object:Gem::Specification
 name: pearson-hashing
 version: !ruby/object:Gem::Version
-  version: 0.0.3
+  version: 0.0.4
   prerelease: 
 platform: ruby
 authors:
@@ -9,7 +9,7 @@ authors:
 autorequire: 
 bindir: bin
 cert_chain: []
-date: 2012-01-17 00:00:00.000000000Z
+date: 2012-01-18 00:00:00.000000000Z
 dependencies: []
 description: pearson hasing provides "fast hashing of variable-length text strings
 email: