sanscript 0.1.0 → 0.2.0

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA1:
-  metadata.gz: d94fb9c3290ec64af941b806bb8cd78f0b66b442
-  data.tar.gz: 3e298a3b363a89081fbf603247f2b49a71144b8a
+  metadata.gz: 4e9133bd05cf7deb7c03cec7e4c9d46bac3eabaf
+  data.tar.gz: e36b9dfe888ee1070d674d72a766130527ce603b
 SHA512:
-  metadata.gz: 8ed2a31fa2f140f4e0085638996cbf31693735d07348fb367505fa104a06a1d22834f18ac3cbc0696079ff0503729b0b64192c006ccce1945ad3de5737d8aef3
-  data.tar.gz: fe535ef6247b9d91ab23643566a2d9e86d7c144cdccf66f3ce6f8b18ba49830813c8d9c20e27c10fcb536c480cee25d504119bf2ba7f04d2676b49f59d4135d4
+  metadata.gz: 651f1f92402d0b8507ffeda3df3d93ca9c16316f37f0903cb4034efda2a1e5d92df9c9f40810e34f43466de84e15c22d771f73404248811846b6249f5eefa4ba
+  data.tar.gz: c79cc7021d15d791663faf6d3895bea4e80981c5f6061e780d4d38e057df3f6b1e89ff4eab5fcf8298a889f15d2c9289cf49074a7206fbd2ecb4c75444a32c7b

data/lib/sanscript/benchmark.rb

@@ -1,53 +1,79 @@
 # frozen_string_literal: true
 
 require "sanscript/refinements"
-require "benchmark"
+begin
+  require "benchmark/ips"
+rescue LoadError
+  module Benchmark
+    def self.ips(*)
+      raise NotImplementedError, "You must install the `benchmark-ips` gem first."
+    end
+  end
+end
 
 module Sanscript
   using Refinements
   module Benchmark
     module_function
 
-    def detection!
-      n = 100_000
+    def detect!
       iast_string = "nānāśāstrasubhāṣitāmṛtarasaiḥ śrotrotsavaṃ kurvatāṃ yeṣāṃ yānti dināni paṇḍitajanavyāyāmakhinnātmanām teṣāṃ janma ca jīvitaṃ ca sukṛtaṃ tair eva bhūr bhūṣitā śeṣaih kiṃ paśuvad vivekarahitair bhūbhārabhūtair naraḥ"
       deva_string = "नानाशास्त्रसुभाषितामृतरसैः श्रोत्रोत्सवं कुर्वतां येषां यान्ति दिनानि पण्डितजनव्यायामखिन्नात्मनाम् तेषां जन्म च जीवितं च सुकृतं तैर् एव भूर् भूषिता शेषैह् किं पशुवद् विवेकरहितैर् भूभारभूतैर् नरः"
+      slp1_string = "nAnASAstrasuBAzitAmftarasEH SrotrotsavaM kurvatAM yezAM yAnti dinAni paRqitajanavyAyAmaKinnAtmanAm tezAM janma ca jIvitaM ca sukftaM tEr eva BUr BUzitA SezEh kiM paSuvad vivekarahitEr BUBAraBUtEr naraH"
+      hk_string = "nAnAzAstrasubhASitAmRtarasaiH zrotrotsavaM kurvatAM yeSAM yAnti dinAni paNDitajanavyAyAmakhinnAtmanAm teSAM janma ca jIvitaM ca sukRtaM tair eva bhUr bhUSitA zeSaih kiM pazuvad vivekarahitair bhUbhArabhUtair naraH"
+      malayalam_string = "അ ആ ഇ ഈ ഉ ഊ ഋ ൠ ഌ ൡ എ ഏ ഐ ഒ ഓ ഔ"
 
-      ::Benchmark.bmbm(18) do |x|
+      ::Benchmark.ips do |x|
+        x.config(time: 5, warmup: 1)
+        x.report("Detect Devanagari") do
+          Sanscript::Detect.detect_script(deva_string) == :devanagari
+        end
+        x.report("Detect Malayalam") do
+          Sanscript::Detect.detect_script(malayalam_string) == :malayalam
+        end
         x.report("Detect IAST") do
-          n.times { raise unless Sanscript.detect(iast_string) == :iast }
+          Sanscript::Detect.detect_script(iast_string) == :iast
         end
-        x.report("Detect Devanagari") do
-          n.times { raise unless Sanscript.detect(deva_string) == :devanagari }
+        x.report("Detect SLP1") do
+          Sanscript::Detect.detect_script(slp1_string) == :slp1
         end
+        x.report("Detect HK") do
+          Sanscript::Detect.detect_script(hk_string) == :hk
+        end
+        x.compare!
       end
+      true
     end
 
-    def transliteration!
-      n = 5_000
+    def transliterate!
       iast_string = "nānāśāstrasubhāṣitāmṛtarasaiḥ śrotrotsavaṃ kurvatāṃ yeṣāṃ yānti dināni paṇḍitajanavyāyāmakhinnātmanām teṣāṃ janma ca jīvitaṃ ca sukṛtaṃ tair eva bhūr bhūṣitā śeṣaih kiṃ paśuvad vivekarahitair bhūbhārabhūtair naraḥ"
-
       deva_string = "नानाशास्त्रसुभाषितामृतरसैः श्रोत्रोत्सवं कुर्वतां येषां यान्ति दिनानि पण्डितजनव्यायामखिन्नात्मनाम् तेषां जन्म च जीवितं च सुकृतं तैर् एव भूर् भूषिता शेषैह् किं पशुवद् विवेकरहितैर् भूभारभूतैर् नरः"
-      ::Benchmark.bmbm(18) do |x|
-        x.report("IAST**>Devanagari") do
-          n.times { Sanscript.transliterate(iast_string, :devanagari) }
-        end
+      slp1_string = "nAnASAstrasuBAzitAmftarasEH SrotrotsavaM kurvatAM yezAM yAnti dinAni paRqitajanavyAyAmaKinnAtmanAm tezAM janma ca jIvitaM ca sukftaM tEr eva BUr BUzitA SezEh kiM paSuvad vivekarahitEr BUBAraBUtEr naraH"
+
+      ::Benchmark.ips do |x|
+        x.config(time: 5, warmup: 2)
+
         x.report("IAST==>Devanagari") do
-          n.times { Sanscript.transliterate(iast_string, :iast, :devanagari) }
-        end
-        x.report("IAST**>SLP1") do
-          n.times { Sanscript.transliterate(iast_string, :slp1) }
+          Sanscript.transliterate(iast_string, :iast, :devanagari) == deva_string
         end
         x.report("IAST==>SLP1") do
-          n.times { Sanscript.transliterate(iast_string, :iast, :slp1) }
+          Sanscript.transliterate(iast_string, :iast, :slp1) == slp1_string
+        end
+        x.report("SLP1==>Devanagari") do
+          Sanscript.transliterate(slp1_string, :slp1, :devanagari) == deva_string
+        end
+        x.report("SLP1==>IAST") do
+          Sanscript.transliterate(slp1_string, :slp1, :iast) == iast_string
         end
-        x.report("Devanagari**>SLP1") do
-          n.times { Sanscript.transliterate(deva_string, :slp1) }
+        x.report("Devanagari==>SLP1") do
+          Sanscript.transliterate(deva_string, :devanagari, :slp1) == slp1_string
         end
-        x.report("Devanagari**>IAST") do
-          n.times { Sanscript.transliterate(deva_string, :iast) }
+        x.report("Devanagari==>IAST") do
+          Sanscript.transliterate(deva_string, :devanagari, :iast) == iast_string
         end
+        x.compare!
       end
+      true
     end
   end
 end

data/lib/sanscript/detect.rb

@@ -1,4 +1,6 @@
 # frozen_string_literal: true
+# rubocop:disable Style/CaseEquality
+
 #
 # Developed from code available @ https://github.com/sanskrit/detect.js
 #
@@ -31,7 +33,7 @@ module Sanscript
     RE_ITRANS_ONLY = /ee|oo|\^[iI]|RR[iI]|L[iI]|~N|N\^|Ch|chh|JN|sh|Sh|\.a/
 
     # Match on SLP1-only characters and bigrams
-    RE_SLP1_ONLY = /[fFxXEOCYwWqQPB]|kz|Nk|Ng|tT|dD|Sc|Sn|[aAiIuUfFxXeEoO]R|G[yr]|(\\W|^)G'/
+    RE_SLP1_ONLY = /[fFxXEOCYwWqQPB]|kz|N[kg]|tT|dD|S[cn]|[aAiIuUeo]R|G[yr]/
 
     # Match on Velthuis-only characters
     RE_VELTHUIS_ONLY = /\.[mhnrlntds]|"n|~s/
@@ -39,7 +41,7 @@ module Sanscript
     # Match on chars shared by ITRANS and Velthuis
     RE_ITRANS_OR_VELTHUIS_ONLY = /aa|ii|uu|~n/
 
-    # Match on characters unavailable in Harvard-Kyoto
+    # Match on characters available in Harvard-Kyoto
     RE_HARVARD_KYOTO = /[aAiIuUeoRMHkgGcjJTDNtdnpbmyrlvzSsh]/
 
     private_constant :RE_BRAHMIC_RANGE, :RE_BRAHMIC_SCRIPTS, :RE_IAST_OR_KOLKATA_ONLY,
@@ -50,24 +52,25 @@ module Sanscript
 
     def detect_script(text)
       # Brahmic schemes are all within a specific range of code points.
-      if text =~ RE_BRAHMIC_RANGE
+      if RE_BRAHMIC_RANGE === text
         RE_BRAHMIC_SCRIPTS.each do |script, regex|
-          return script if text =~ regex
+          return script if regex === text
         end
       end
 
       # Romanizations
-      if text =~ RE_IAST_OR_KOLKATA_ONLY
-        text =~ RE_KOLKATA_ONLY ? :kolkata : :iast
-      elsif text =~ RE_ITRANS_ONLY
+      if RE_IAST_OR_KOLKATA_ONLY === text
+        return :kolkata if RE_KOLKATA_ONLY === text
+        :iast
+      elsif RE_ITRANS_ONLY === text
         :itrans
-      elsif text =~ RE_SLP1_ONLY
+      elsif RE_SLP1_ONLY === text
         :slp1
-      elsif text =~ RE_VELTHUIS_ONLY
+      elsif RE_VELTHUIS_ONLY === text
         :velthuis
-      elsif text =~ RE_ITRANS_OR_VELTHUIS_ONLY
+      elsif RE_ITRANS_OR_VELTHUIS_ONLY === text
         :itrans
-      elsif text =~ RE_HARVARD_KYOTO
+      elsif RE_HARVARD_KYOTO === text
         :hk
       else
         :unknown

data/lib/sanscript/transliterate.rb

@@ -14,7 +14,8 @@ module Sanscript
   using Refinements
   module Transliterate
     class << self
-      attr_reader :defaults, :schemes, :roman_schemes, :all_alternates
+      attr_reader :defaults, :schemes, :scheme_names, :brahmic_schemes, :roman_schemes,
+                  :all_alternates
     end
 
     @defaults = {
@@ -27,12 +28,13 @@ module Sanscript
     module_function
 
     #
-    #  Return a list of available schemes.
+    #  Check whether the given scheme encodes Brahmic Sanskrit.
     #
-    #  @return      array of scheme identifiers
+    #  @param name  the scheme name
+    #  @return      boolean
     #
-    def scheme_names
-      @schemes.keys.sort!
+    def brahmic_scheme?(name)
+      @brahmic_schemes.include?(name.to_sym)
     end
 
     #
@@ -64,7 +66,12 @@ module Sanscript
     #                described above.
     #
     def add_brahmic_scheme(name, scheme)
-      @schemes[name.to_sym] = scheme.deep_dup.deep_freeze
+      name = name.to_sym
+      scheme = scheme.deep_dup
+      @schemes[name] = scheme.deep_freeze
+      @brahmic_schemes.add(name)
+      @scheme_names.add(name)
+      scheme
     end
 
     #
@@ -82,6 +89,8 @@ module Sanscript
       scheme[:vowel_marks] = scheme[:vowels][1..-1] unless scheme.key?(:vowel_marks)
       @schemes[name] = scheme.deep_freeze
       @roman_schemes.add(name)
+      @scheme_names.add(name)
+      scheme
     end
 
     #
@@ -93,15 +102,23 @@ module Sanscript
 
     # Set up various schemes
     begin
+      # Re-add existing Brahmic schemes in order to add them to `scheme_names`
+      # and to freeze them up.
+      brahmic_scheme_names = %i[bengali devanagari gujarati gurmukhi kannada malayalam
+                                oriya tamil telugu]
+      brahmic_scheme_names.each do |name|
+        add_brahmic_scheme(name, @schemes[name])
+      end
+
       # Set up roman schemes
       kolkata = @schemes[:kolkata] = @schemes[:iast].deep_dup
-      scheme_names = %i[iast itrans hk kolkata slp1 velthuis wx]
+      roman_scheme_names = %i[iast itrans hk kolkata slp1 velthuis wx]
       kolkata[:vowels] = %w[a ā i ī u ū ṛ ṝ ḷ ḹ e ē ai o ō au]
 
       # These schemes already belong to Sanscript.schemes. But by adding
-      # them again with `addRomanScheme`, we automatically build up
+      # them again with `add_roman_scheme`, we automatically build up
       # `roman_schemes` and define a `vowel_marks` field for each one.
-      scheme_names.each do |name|
+      roman_scheme_names.each do |name|
         add_roman_scheme(name, @schemes[name])
       end
 
@@ -112,8 +129,7 @@ module Sanscript
       @all_alternates[:itrans_dravidian] = @all_alternates[:itrans]
       add_roman_scheme(:itrans_dravidian, itrans_dravidian)
 
-      # ensure deep freeze on all existing schemes and alternates
-      @schemes.each { |_, scheme| scheme.deep_freeze }
+      # ensure deep freeze on alternates
       @all_alternates.each { |_, scheme| scheme.deep_freeze }
     end
 

data/lib/sanscript/transliterate/schemes.rb

@@ -275,9 +275,15 @@ module Sanscript
       },
     }
 
-    # Set of names of schemes
+    # Set of names of Roman schemes
     @roman_schemes = Set.new
 
+    # Set of names of Brahmic schemes
+    @brahmic_schemes = Set.new
+
+    # Set of names of all schemes
+    @scheme_names = Set.new
+
     # Map of alternate encodings.
     @all_alternates = {
       itrans: {

data/lib/sanscript/version.rb

@@ -1,4 +1,4 @@
 # frozen_string_literal: true
 module Sanscript
-  VERSION = "0.1.0"
+  VERSION = "0.2.0"
 end

data/sanscript.gemspec

@@ -24,6 +24,7 @@ Gem::Specification.new do |spec|
   spec.add_development_dependency "rspec", "~> 3.5"
   spec.add_development_dependency "codeclimate-test-reporter", "~> 0.6"
   spec.add_development_dependency "pry", "~> 0.10"
+  spec.add_development_dependency "benchmark-ips", "~> 2.6"
 
   spec.add_runtime_dependency "ice_nine", "~> 0.11"
 end

metadata

@@ -1,14 +1,14 @@
 --- !ruby/object:Gem::Specification
 name: sanscript
 version: !ruby/object:Gem::Version
-  version: 0.1.0
+  version: 0.2.0
 platform: ruby
 authors:
 - Tim Bellefleur
 autorequire: 
 bindir: exe
 cert_chain: []
-date: 2016-07-13 00:00:00.000000000 Z
+date: 2016-07-14 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: bundler
@@ -80,6 +80,20 @@ dependencies:
     - - "~>"
       - !ruby/object:Gem::Version
         version: '0.10'
+- !ruby/object:Gem::Dependency
+  name: benchmark-ips
+  requirement: !ruby/object:Gem::Requirement
+    requirements:
+    - - "~>"
+      - !ruby/object:Gem::Version
+        version: '2.6'
+  type: :development
+  prerelease: false
+  version_requirements: !ruby/object:Gem::Requirement
+    requirements:
+    - - "~>"
+      - !ruby/object:Gem::Version
+        version: '2.6'
 - !ruby/object:Gem::Dependency
   name: ice_nine
   requirement: !ruby/object:Gem::Requirement