tinycus 1.0.5 → 1.0.7

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA256:
-  metadata.gz: 90849e6c1df95e50e8f6ba6818d78a8bf6352a82fc9048313a11acd7d1624bf2
-  data.tar.gz: 5b962ac97c0d25777cf527b3a61d0c0e7ebb12c2321b690eca724907bd692de7
+  metadata.gz: 83d0a3e5d3f764901829858fc5a1b9160577e46f5e277504291c21225c2a38a0
+  data.tar.gz: da3dece01b93b157e1e36d7e07b27ead0f55dab621101af4444e716672028323
 SHA512:
-  metadata.gz: 6af58dc8e3928c473e96416b20fee0893aba4b9abf3838f9689e5b95326eb74c80986a6fa09d47c3ca8240b00d2c17609d2b1adbdc74e86ac98055889ca9aff7
-  data.tar.gz: a5e5a4cad6fbfae7fe6588864cc3607eec471b0c0db6e8adfc3657c44826b63696cac92c410423ef62fe96f3e618e0a476e8b89d764404115a68fc03dc7a8336
+  metadata.gz: 16fb025b30abf1d6650aa2a76108117a88e30e5d5b524fc8aae83aca1531b62a2b4baab4da9167b673636cae413b8225fc236772185c9c1063bc4a31bc1321c3
+  data.tar.gz: e0cc0963f112a25e782875ecb179e035a0dabb1a6d20b2e3b59d0b2c5584bfb0bbe1d9b41c275dac631b7c520822a1e34b1c1340184a64f27bda52d403c3dd4b

data/tinycus.rb

@@ -4,45 +4,51 @@ require 'json'
 
 module Tinycus
 
-  # The four "alpha_" functions work on Greek and English, also most Latin characters; see comments in Tr.get_greek_collation_tr.
+=begin rdoc
+The four "alpha_" functions work on Greek and English, also most Latin characters; see comments in Tr.get_greek_collation_tr.
+=end
 
+  # Sort a list of strings in alphabetical order.
   def Tinycus.alpha_sort(l,n:false)
     return Tinycus.sort(l,Tinycus.alpha_collation,n:false)
   end
 
+  # Tests two strings for equality in alphabetical order, returns a boolean.
   def Tinycus.alpha_equal(a,b,n:true)
     return (Tinycus.alpha_compare(a,b,n:n)==0)
   end
 
+  # An alphabetical <=> function.
   def Tinycus.alpha_compare(a,b,n:true)
     # return (Tinycus::Tr.remove_accents_from_greek(a,n:n).downcase <=> Tinycus::Tr.remove_accents_from_greek(b,n:n).downcase)
     collation_tr = Tinycus.alpha_collation
     return collation_tr.apply(a,n:n) <=> collation_tr.apply(b,n:n)
   end
 
-  def Tinycus.alpha_collation
+  def Tinycus.alpha_collation() # :nodoc:
     return Tinycus::Tr.get_greek_collation_tr
   end
 
+  # synonym of Tinycus.contains_vowel, for readability when using it on a single character
   def Tinycus.is_vowel(c)
-    # synonym of Tinycus.contains_vowel, for readability when using it on a single character
     return Tinycus.contains_vowel(c)
   end
 
+  # Works for Greek and Latin; considers y to be a vowel; doesn't handle stuff like Welsh w.
   def Tinycus.contains_vowel(s)
-    # works for Greek and Latin; considers y to be a vowel; doesn't handle stuff like Welsh w
     if Tinycus::Tr.remove_accents_from_greek(s).downcase.match?(/[αειουηωaeiouyæ]/) then return true else return false end
   end
 
   # ---
 
+  # Like alpha_sort, but more general, using any Tr object to define a collation order.
   def Tinycus.sort(l,collation_tr,n:false)
     p = l.clone # This does work on a list of strings: ruby -e "a=['p','q']; b=a.clone; b[1]='x'; print a"
     p.sort_by! { |x| collation_tr.apply(x,n:n) } # ruby's sort_by! only applies the block once to each element, to form an index
     return p
   end
 
-  def Tinycus.run_tests
+  def Tinycus.run_tests # :nodoc:
     # removing accents:
     tests = [
       ['',''],
@@ -125,23 +131,13 @@ module Tinycus
       end
     }
 
+    Tinycus::Tr.run_tests()
+    Tinycus::MiscGreek.run_tests()
 
   end
 
-  # fixme:
-  #   Determine byte order and make sure we convert to native (i.e., BE rather than LE if we're on a big-endian machine).
-  @@bloater = Encoding::Converter.new('UTF-8','UTF-32LE')
-  @@shrinker = Encoding::Converter.new('UTF-32LE','UTF-8')
-
-  def Tinycus.bloat(s) # private method
-    return @@bloater.convert(s)
-  end
-  def Tinycus.shrink(s) # private method
-    return @@shrinker.convert(s)
-  end
-
   class Tinycus::MiscGreek
-    def MiscGreek.run_tests
+    def MiscGreek.run_tests # :nodoc:
       print "testing MiscGreek.add_second_accent...\n"
       [['θεμείλια','θεμείλιά'],
        ['πόλεμονδε','πόλεμόνδε'],
@@ -155,9 +151,9 @@ module Tinycus
       }
     end
 
+    # A rough approximation, for cases where we don't need perfect precision and either don't have Ifthimos's syllabification module
+    # or don't want the performance hit.
     def MiscGreek.estimate_syll_count(x)
-      # A rough approximation, for cases where we don't need perfect precision and either don't have Ransom's greek/syllab.rb
-      # or don't want the performance hit.
       x = x.downcase
       x = x.gsub(/[ϊ]/,'e') # prevent it from being misinterpreted as a diphthong after the diaresis is stripped below
       x = Tr.remove_accents_from_greek(x)
@@ -181,10 +177,10 @@ module Tinycus
       return Tr.remove_tonal_accents_from_greek(x)!=x
     end
 
-    def MiscGreek.add_second_accent(w_orig)
-      # e.g., if w is θεμείλια, returns θεμείλιά
-      w = w_orig.clone # shallow copy, works on a string; I'm not clear in why this is necessary, but it is; modification to w_orig is visible in
-                       # output of make test_misc_greek
+    # Modifies a word as would be appropriate if it was followed by an enclitic. E.g., if the input is θεμείλια, returns θεμείλιά.
+    def MiscGreek.add_second_accent(w)
+      w = w.clone # shallow copy, works on a string; I'm not clear in why this is necessary, but it is; modification to w is visible in
+                  # output of make test_misc_greek
       if w=~/(.*)δε$/ then
         stem = $1
         nsyll = MiscGreek.estimate_syll_count(stem)
@@ -202,14 +198,14 @@ module Tinycus
       return w
     end
 
+    # For a given word, try to predict every possible form it could take in a text, including
+    # both possible capitalizations, acute/grave, and multiple accents.
+    # The word w should already have been converted into a canonical dictionary form (typically a single acute accent).
+    # This is not 100% perfect, mainly because the rules for multiple accents are complicated and Tinycus doesn't include a full
+    # syllabification algorithm.
+    # I tested this as a round-trip on all multiply accented words occurring in Homer. The following three words were the only
+    # ones where it failed: κάλλίον, σταφύλῇ, ὕπὸ.
     def MiscGreek.all_cases_and_accents(w)
-      # For a given word, try to predict every possible form it could take in a text, including
-      # both possible capitalizations, acute/grave, and multiple accents.
-      # The word w should already have been converted into a canonical dictionary form (typically a single acute accent).
-      # This is not 100% perfect, mainly because the rules for multiple accents are complicated and Tinycus doesn't include a full
-      # syllabification algorithm.
-      # I tested this as a round-trip on all multiply accented words occurring in Homer. The following three words were the only
-      # ones where it failed: κάλλίον, σταφύλῇ, ὕπὸ.
       forms = [w.downcase]
       forms = forms+forms.map { |x| Tr.greek_acute_to_grave(x) }
       forms = forms+forms.map { |x| MiscGreek.add_second_accent(x) }
@@ -235,10 +231,11 @@ module Tinycus
     @@greek_acute_to_grave = nil
     @@prep_greek_to_collation_form = nil
 
+    # Initialize a data structure that represents an action equivalent to String#tr(a,b), but faster.
+    # Including redundant characters or unchanged characters is harmless and is fixed in this constructor; it does not cause
+    # any performance hit when the object is actually used. The initializer takes linear time and memory in the size of
+    # the inputs.
     def initialize(a,b)
-      # Initialize a data structure that represents an action equivalent to string.tr(a,b), but faster.
-      # Including redudant characters is harmless and is fixed in this constructor; it does not cause
-      # any performance hit when the object is actually used.
       if a.length!=b.length then raise "lengths unequal, #{a.length} and #{b.length}" end
       @l = a.length
       @orig_tables = [a.clone,b.clone] # stash them away for testing purposes
@@ -246,7 +243,7 @@ module Tinycus
       @h = {}
       0.upto(@l-1) { |i|
         p,q = a[i],b[i]
-        @h[p] = q
+        if p!=q then @h[p] = q end
       }
       @h.freeze
       
@@ -254,9 +251,13 @@ module Tinycus
 
     attr_reader :l,:a,:b
 
+    # Takes a predefined Tr object and uses it to perform the equivalent of String#tr. Takes O(1) time in the length of
+    # the translation list.
+    # This function tends to be a bottleneck for performance in real-world applications. I tried several algorithms.
+    # See notes in comments at top of scripts/benchmark.rb.
+    # Something like Gnu gperf is theoretically superior in certain ways (e.g., avoiding the theoretical possibility of
+    # a hash collision).
     def apply(s,n:false)
-      # This function tends to be a bottleneck for performance in real-world applications. I tried several algorithms.
-      # See notes in comments at top of scripts/benchmark.rb.
 
       if n then s = s.unicode_normalize(:nfc) end # 30% performance hit, not necessary if input has already been normalized
 
@@ -271,9 +272,10 @@ module Tinycus
 
     end
 
-    def self_test(alphabet)
-      # Raises an exception if it fails. Otherwise just returns silently.
-      # If not nil, then the alphabet parameter gives a list of characters that are allowed to exist in the output.
+    # Tests a Tr object and pokes and prods it to see if it seems OK.
+    # Raises an exception if it fails. Otherwise just returns silently.
+    # If not nil, then the alphabet argument should give a list of characters that are allowed to exist in the output.
+    def self_test(alphabet) # :nodoc:
       a,b = @orig_tables
       if self.apply(a)!=b then raise "error in self_test, applying me to original a does not give original b" end
       if self.apply(b)!=b then raise "error in self_test, applying me to original b does not give original b" end
@@ -442,6 +444,7 @@ module Tinycus
       return @@prep_remove_accents_from_greek.apply(s)
     end
 
+    # Slow.
     def Tr.remove_macrons_and_breves(s)
       # This can't be implemented using my fast method, because most of these are composed characters.
       if !(s.kind_of?(String)) then return s end
@@ -500,12 +503,11 @@ a.each { |c|
 =end
 end
 
+    # Changes a macronized string to one that looks like this: ἕννυ_μι.
+    # We don't handle grave and circumflex accents, but those don't occur in dictionary headers with macrons.
+    # For an IfMows object in Ifthimos, this can be done using stringify(macronization:'underbar').
     def Tr.macronized_to_underbar_style(s)
-      # Changes a macronized string to one that looks like this: ἕννυ_μι.
       # The lists in the regexes are generated by the commented-out scripts below, and are not actually totally comprehensive. 
-      # We don't handle grave and circumflex accents, but those don't occur in dictionary headers with macrons.
-      # For an IfMows object in Ifthimos, this can be done using stringify(macronization:'underbar').
-      # ---
       x = s.clone
       x = x.gsub(/(ϊ̄)/) { "#{Tinycus::Tr.remove_macrons_and_breves($1)}_" } # iota with diaresis and macron, occurs in ἀϊκή
       # First handle letters that have both a macron and a breve, treating them as if they weren't macronized at all:
@@ -542,7 +544,7 @@ print a.join('|'),"--\n"
 =end
     end
 
-    def Tr.get_greek_collation_tr
+    def Tr.get_greek_collation_tr # :nodoc:
       if @@prep_greek_to_collation_form.nil? then
         @@prep_greek_to_collation_form = Tr.collation_form('el')
       end
@@ -557,10 +559,10 @@ print a.join('|'),"--\n"
       return @@prep_greek_to_collation_form.apply(s)
     end
 
+    # Returns a Tinycus::Tr object which can then be used to act on strings using the apply() method.
+    # The 'el' locale is a standard thing that software like ICU uses for polytonic Greek. The object constructed with this
+    # locale will also remove most accents and macrons from Latin characters, but will miss some cases like Czech, and will not handle Cyrillic.
     def Tr.remove_accents(locale)
-      # Returns a Tinycus::Tr object which can then be used to act on strings using the apply() method.
-      # The 'el' locale is a standard thing that software like ICU uses for polytonic Greek. The object constructed with this
-      # locale will also remove most accents and macrons from Latin characters, but will miss some cases like Czech, and will not handle Cyrillic.
       t = {
         "el"=>[
           "ÀÁÂÃÄÅÆÇÈÉÊËÌÍÎÏÑÒÓÔÕÖØÙÚÛÜÝàáâãäåæçèéêëìíîïñòóôõöøùúûüýÿΆΈΊΌΐάέήίΰϊϋόύώỏἀἁἂἃἄἅἆἈἉἊἌἍἎἐἑἒἓἔἕἘἙἜἝἠἡἢἣἤἥἦἧἨἩἫἬἭἮἯἰἱἲἳἴἵἶἷἸἹἼἽἾὀὁὂὃὄὅὈὉὊὋὌὍὐὑὓὔὕὖὗὙὝὠὡὢὣὤὥὦὧὨὩὫὬὭὮὯὰὲὴὶὸὺὼᾐᾑᾓᾔᾕᾖᾗᾠᾤᾦᾧᾰᾱᾳᾴᾶᾷᾸᾹῂῃῄῆῇῐῑῒῖῗῘῙῠῡῢῥῦῨῩῬῳῴῶῷῸῤᾆᾄᾂᾁᾇᾅᾃᾍᾡ",
@@ -573,11 +575,11 @@ print a.join('|'),"--\n"
       return result
     end
 
+    # Returns a Tinycus::Tr object which can then be used to act on strings using the apply() method. Gives a form that 
+    # can be alphabetized properly.
+    # The 'el' locale is a standard thing that software like ICU uses for polytonic Greek. The object constructed with this
+    # locale will also produce correct results for most Latin-script words, will miss some cases like Czech, and will not handle Cyrillic.
     def Tr.collation_form(locale)
-      # Returns a Tinycus::Tr object which can then be used to act on strings using the apply() method. Gives a form that 
-      # can be alphabetized properly.
-      # The 'el' locale is a standard thing that software like ICU uses for polytonic Greek. The object constructed with this
-      # locale will also produce correct results for most Latin-script words, will miss some cases like Czech, and will not handle Cyrillic.
       t = {
         "el"=>[
           "ÀÁÂÃÄÅÆÇÈÉÊËÌÍÎÏÑÒÓÔÕÖØÙÚÛÜÝàáâãäåæçèéêëìíîïñòóôõöøùúûüýÿΆΈΊΌΐάέήίΰϊϋόύώỏἀἁἂἃἄἅἆἈἉἊἌἍἎἐἑἒἓἔἕἘἙἜἝἠἡἢἣἤἥἦἧἨἩἫἬἭἮἯἰἱἲἳἴἵἶἷἸἹἼἽἾὀὁὂὃὄὅὈὉὊὋὌὍὐὑὓὔὕὖὗὙὝὠὡὢὣὤὥὦὧὨὩὫὬὭὮὯὰὲὴὶὸὺὼᾐᾑᾓᾔᾕᾖᾗᾠᾤᾦᾧᾰᾱᾳᾴᾶᾷᾸᾹῂῃῄῆῇῐῑῒῖῗῘῙῠῡῢῥῦῨῩῬῳῴῶῷῸῤᾆᾄᾂᾁᾇᾅᾃΑΒΓΔΕΖΗΘΙΚΛΜΝΞΟΠΡΣΤΥΦΧΨΩςᾍ",
@@ -590,15 +592,14 @@ print a.join('|'),"--\n"
       return result
     end
 
+    # E.g., Tr.add_breathing_to_character('α','rough') gives 'ἁ'.
     def Tr.add_breathing_to_character(c,what)
       plain,d = Tinycus.disassemble_greek_char(c)
       d['breathing'] = what
       return Tinycus.assemble_greek_char(plain,d)
     end
 
-
-
-    def Tr.run_tests
+    def Tr.run_tests # :nodoc:
       # to execute this, do a "make test_tr"
       ['el'].each { |locale|
         tr = Tinycus::Tr.remove_accents(locale)
@@ -613,10 +614,10 @@ print a.join('|'),"--\n"
 
   class Tinycus::Script
 
+    # Script can be 'latin', 'greek', or 'hebrew'.
+    # C can be both, lowercase, or uppercase.
+    # For scripts that don't have case, c is ignored.
     def Script.alphabet(script,c:'both')
-      # Script can be 'latin', 'greek', or 'hebrew'.
-      # C can be both, lowercase, or uppercase.
-      # For scripts that don't have case, c is ignored.
       t = {
         'latin'=>{'has_case'=>true},
         'greek'=>{'has_case'=>true},
@@ -633,7 +634,7 @@ print a.join('|'),"--\n"
       die("illegal value of c=#{c}, must be both, lowercase, or uppercase")
     end
 
-    def Script.alphabet_helper(script,include_lc_only_chars)
+    def Script.alphabet_helper(script,include_lc_only_chars) # :nodoc:
       if script=='latin'  then return 'abcdefghijklmnopqrstuvwxyz' end
       if script=='greek'  then 
         result = 'αβγδεζηθικλμνξοπρστυφχψω'
@@ -651,8 +652,9 @@ print a.join('|'),"--\n"
 
   class Tinycus::Cleanup
 
+    # Designed for external data sources that can have all kinds of nasty crap in them. Slow, thorough, silent, and brutal.
+    # Attempts to eliminate the traces of incomplete beta code conversion that are found in some Project Perseus XML files.
     def Cleanup.clean_up_grotty_greek(s,allow_latin:false,clean_perseus:true,standardize_punctuation:true)
-      # Designed for external data sources that can have all kinds of nasty crap in them. Slow, thorough, silent, and brutal.
       a = s.split(/(\s+)/) # returns a string in which even indices are words, odd indices are whitespace
       b = []
       0.upto(a.length-1) { |i|
@@ -676,8 +678,8 @@ print a.join('|'),"--\n"
       return s
     end
 
+    # Like clean_up_grotty_greek, but works on a single word.
     def Cleanup.clean_up_grotty_greek_one_word(s,allow_latin:false,clean_perseus:true,standardize_punctuation:true)
-      # This works on a single word.
       s = s.unicode_normalize(:nfc)
       s = Cleanup.clean_up_greek_combining_characters(s,allow_latin:allow_latin)
       # In Perseus's Polybius, they have bracketed text sometimes. In their system, this should probably be a separate punctuation token.
@@ -705,7 +707,7 @@ print a.join('|'),"--\n"
       return s
     end
 
-    def Cleanup.clean_up_greek_combining_characters(s,allow_latin:false)
+    def Cleanup.clean_up_greek_combining_characters(s,allow_latin:false) # :nodoc:
       combining_comma_above = [787].pack('U')
       combining_acute_accent = [769].pack('U')
       greek_koronis = [8125].pack('U')
@@ -735,8 +737,7 @@ print a.join('|'),"--\n"
       return s
     end
 
-    def Cleanup.clean_up_greek_beta_code(s)
-      # This was for when I mistakenly used old beta code version of project perseus.
+    def Cleanup.clean_up_greek_beta_code(s) # :nodoc:
       # Even with perseus 2.1, some stuff seems to come through that looks like beta code, e.g., ἀργει~ος.
       # https://github.com/PerseusDL/treebank_data/issues/30
       s = s.sub(/\((.)/) { $1.tr("αειουηω","ἁἑἱὁὑἡὡ") }
@@ -751,8 +752,8 @@ print a.join('|'),"--\n"
       return s
     end
 
+    # Works on any string, doesn't have to be a single word. Standardize elision character and middle dot/ano teleia.
     def Cleanup.standardize_greek_punctuation(s)
-      # Works on any string, doesn't have to be a single word. Standardize elision character and middle dot/ano teleia.
       # Perseus and Monro/Allen write ρ with breathing mark instead of ρ᾽ when there's elision:
       s = s.gsub(/(?<=[[:alpha:]])[ῤῥ](?![[:alpha:]])/,'ρ᾽')
       # ... Note that we do need to reinsert the breathing mark, or else we lose the info needed to do accurate lemmatization. Cf. Spelling module.
@@ -1013,26 +1014,31 @@ end
 JSON
 @@beta_code_conversion = nil
 
-  def Tinycus.beta_code_conversion_table
-    if @@beta_code_conversion.nil? then @@beta_code_conversion=JSON.parse(@@beta_code_conversion_json) end
+  def Tinycus.beta_code_conversion_table # :nodoc:
+    if @@beta_code_conversion.nil? then
+      @@beta_code_conversion=JSON.parse(@@beta_code_conversion_json)
+      @@beta_code_conversion.freeze
+    end
     return @@beta_code_conversion
   end
 
+  # Converts a unicode character to beta code. The input must be utf8/nfc.
   def Tinycus.greek_char_unicode_to_beta_code(u)
     b = Tinycus.beta_code_conversion_table()[0][u]
     if !b.nil? then return b else return u end # most failures will just be whitespace, punctuation, etc.
   end
 
+  # Converts a character from beta code to unicode.
   def Tinycus.greek_char_beta_code_to_unicode(b)
     b = Tinycus.canonicalize_char_greek_beta_code(b)
     u = Tinycus.beta_code_conversion_table()[1][b]
     if !u.nil? then return u else return b end
   end
 
+  # Breathing normally comes after accent, but sometimes you see things in the wild where it's reversed.
+  # I can't find any documentation for any preferred or canonical order. What circumstantial evidence I could find I put into the WP article.
+  # Note that the order of |+ doesn't matter, because the same letter can't have both.
   def Tinycus.canonicalize_char_greek_beta_code(b)
-    # Breathing normally comes after accent, but sometimes you see things in the wild where it's reversed.
-    # I can't find any documentation for any preferred or canonical order. What circumstantial evidence I could find I put into the WP article.
-    # Note that the order of |+ doesn't matter, because the same letter can't have both.
     ")(/\\=|+&'".chars.each { |c|
       if b=~/(.*)#{Regexp::quote(c)}(.*)/ then b = $1+$2+c end
     }
@@ -1040,6 +1046,7 @@ JSON
     return b
   end
 
+  # Converts a string from utf8/nfc to beta code.
   def Tinycus.greek_unicode_to_beta_code(u)
     u = Tinycus::Cleanup.clean_up_grotty_greek(u,allow_latin:true,clean_perseus:true,standardize_punctuation:false)
     # ... the conversion below will not work on unicode that isn't done cleanly and according to modern standards
@@ -1051,6 +1058,7 @@ JSON
     return b
   end
 
+  # Converts a string from beta code to unicode.
   def Tinycus.greek_beta_code_to_unicode(b)
     # This implementation will be kind of slow because it does regex replacements in place.
     b = b.clone
@@ -1065,13 +1073,13 @@ JSON
     return b
   end
 
+  # Returns [plain,d], where plain is a lowercase, unaccented Greek letter (α-ω, plus ς), and d is
+  # a hash with the following keys:
+  #   uppercase, diar, iota_subscript - boolean values
+  #   tonal - string value: none acute grave circumflex
+  #   breathing - string value: none smooth rough
+  # Doesn't handle macrons and breves. I have a function IfMows.disassemble_char in Ifthimos that does that.
   def Tinycus.disassemble_greek_char(c)
-    # Returns [plain,d], where plain is a lowercase, unaccented Greek letter (α-ω, plus ς), and d is
-    # a hash with the following keys:
-    #   uppercase, diar, iota_subscript - boolean values
-    #   tonal - string value: none acute grave circumflex
-    #   breathing - string value: none smooth rough
-    # Doesn't handle macrons and breves. I have a function IfMows.disassemble_char in Ifthimos that does that.
     x = Tinycus.disassemble_greek_char_binary(c)
     if x.nil? then return nil end
     plain,decor = x
@@ -1092,9 +1100,8 @@ JSON
     return [plain,d]
   end
 
+  # The inverse of Tinycus.disassemble_greek_char.
   def Tinycus.assemble_greek_char(plain,d)
-    # The inverse of Tinycus.disassemble_greek_char.
-    # Doesn't handle macrons and breves. I have a function IfMows.assemble_char in Ifthimos that does that.
     b = 0
     b |=        0b1 if d['uppercase']
     b |=       0b10 if d['diar']
@@ -1108,7 +1115,7 @@ JSON
     return Tinycus.assemble_greek_char_hex(x)
   end
 
-  def Tinycus.disassemble_greek_char_binary(c)
+  def Tinycus.disassemble_greek_char_binary(c) # :nodoc:
     # Returns [plain,b], where plain is a lowercase, unaccented Greek letter (α-ω, plus ς), and b is
     # an integer containing a set of flags encoded in binary, as follows:
     #  b |=        0b1 if d['uppercase']
@@ -1122,7 +1129,7 @@ JSON
     return @@disassemble_greek_char_hash[c]
   end
 
-  def Tinycus.assemble_greek_char_hex(x)
+  def Tinycus.assemble_greek_char_hex(x) # :nodoc:
     # The inverse of the map in Tinycus.disassemble_greek_char_binary.
     # Accepts an input such as 'α08', where the second and third characters are the hex representation of of
     # the set of flags described in the comments in the forward map.

metadata

@@ -1,7 +1,7 @@
 --- !ruby/object:Gem::Specification
 name: tinycus
 version: !ruby/object:Gem::Version
-  version: 1.0.5
+  version: 1.0.7
 platform: ruby
 authors:
 - Benjamin Crowell
@@ -16,7 +16,8 @@ description: "This is a ruby library to do some string functions efficiently tha
 email: 
 executables: []
 extensions: []
-extra_rdoc_files: []
+extra_rdoc_files:
+- README.md
 files:
 - LICENSE
 - README.md
@@ -24,9 +25,16 @@ files:
 homepage: https://bitbucket.org/ben-crowell/tinycus
 licenses:
 - GPL-3.0-only
-metadata: {}
+metadata:
+  contact_uri: http://lightandmatter.com/area4author.html
+  homepage_uri: https://bitbucket.org/ben-crowell/tinycus
+  source_code_uri: https://bitbucket.org/ben-crowell/tinycus
 post_install_message: 
-rdoc_options: []
+rdoc_options:
+- "--exclude"
+- "(bad_combining|generating)"
+- "--main"
+- README.md
 require_paths:
 - lib
 required_ruby_version: !ruby/object:Gem::Requirement