sekka 1.5.9 → 1.6.0

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA1:
-  metadata.gz: ecfc09c0a0ece949a937bece16b5c2d2aa3412a2
-  data.tar.gz: 4a73557deb029dbf3043bbafcff62f596b66e442
+  metadata.gz: 059e9182b4e6dc1dc9cd996f1840f7ba6f819a9b
+  data.tar.gz: 7d7dc6eeee190b860c2d3c43d61e140672069d6a
 SHA512:
-  metadata.gz: ad755240d92827c790a4b8ab6ec50f849a6db7b94dcf2150370879a4482c6eddf09e9bbef18cc6b4dcc514146f54572bddee33da886034172ab192ec7d87ec8e
-  data.tar.gz: 2c2413b680b46816ca12862db2cf71c370115a36fa055bf4b6ba134d85ebb5b7f8f89ebd49282324403161a657fac42835af832aa4f81b82185bce012140bb68
+  metadata.gz: da64491bc34442de5b0abe7798cdfa3afc797a64bae0650fd0ab0f97897273db9777fe68bf9665ae0811d00634c4faaeb0563b00efc3688023f771a580af2019
+  data.tar.gz: 6b38924ce6b313babe7461d9df811d447902a234f5a4cf1346d87fb7f214f518c97b77a8c4b4e712cd103af57c9b7dbe4d7134da05e0fb0044217d7f50626914

data/.travis.yml

@@ -1,7 +1,6 @@
 language: ruby
 
 rvm:
- - 1.9.2
  - 1.9.3
  - 2.0.0
  - 2.1.3

data/Rakefile

@@ -23,17 +23,13 @@ require 'rake'
 require "bundler/gem_tasks"
 require 'jeweler2'
 
-# generate `normal' and `azik' dictionary
-# (generateTypes = [ "N", "A" ])
-generateTypes = [ "N" ]
-
 
 task :default => [:test] do
 end
 
 task :compile do
   # generate version.rb
-  dictVersion = "1.4.2"
+  dictVersion = "1.5.0"
   vh = Jeweler::VersionHelper.new "."
   open( "./lib/sekka/sekkaversion.rb", "w" ) {|f|
     f.puts(   "class SekkaVersion" )
@@ -112,82 +108,51 @@ task :test do
     files << "./test/henkan-main.nnd  pure"
   end
   files.each {|filename|
-    sh  sprintf( "ruby -I ./lib -S nendo -d %s", filename )
+    sh  sprintf( "ruby -I ./lib -S nendo -I ./lib -d %s", filename )
   }
   sh "cat test.record"
 end
 
 task :bench do
-  sh "time ruby -I ./lib /usr/local/bin/nendo ./test/approximate-bench.nnd"
-  sh "time ruby -I ./lib /usr/local/bin/nendo ./test/henkan-bench.nnd"
+  sh "time nendo -I ./lib ./test/approximate-bench.nnd"
+  sh "time nendo -I ./lib ./test/henkan-bench.nnd"
 end
 
-task :alljisyo  => [ :alljisyoS, :alljisyoL ]
-task :alljisyoS => [ :jisyoS, :loadS, :dumpS, :md5 ]
-task :alljisyoL => [ :jisyoL, :loadL, :dumpL, :md5 ]
+task :alljisyo => [ :jisyo, :load, :dump, :md5 ]
 
 task :md5 do
-  sh "md5sum ./data/SEKKA-JISYO.SMALL.N.tsv > ./data/SEKKA-JISYO.SMALL.N.md5"
-  sh "md5sum ./data/SEKKA-JISYO.LARGE.N.tsv > ./data/SEKKA-JISYO.LARGE.N.md5"
+  sh "md5sum ./data/SEKKA-JISYO.N.tsv > ./data/SEKKA-JISYO.N.md5"
 end
 
-task :jisyoS do
-  generateTypes.each {|x|
-    sh "ruby ./bin/sekka-jisyo convert#{x} ./data/SKK-JISYO.L.201501           >  ./data/SEKKA-JISYO.SMALL.#{x}"
-    sh "ruby ./bin/sekka-jisyo convert#{x} ./data/SKK-JISYO.L.hira-kata        >> ./data/SEKKA-JISYO.SMALL.#{x}"
-    sh "ruby ./bin/sekka-jisyo convert#{x} ./data/SKK-JISYO.hiragana-phrase    >> ./data/SEKKA-JISYO.SMALL.#{x}"
-    sh "ruby ./bin/sekka-jisyo convert#{x} ./data/SKK-JISYO.hiragana-phrase2   >> ./data/SEKKA-JISYO.SMALL.#{x}"
-    sh "ruby ./bin/sekka-jisyo convert#{x} ./data/SKK-JISYO.hiragana-phrase3   >> ./data/SEKKA-JISYO.SMALL.#{x}"
-  }
+task :jisyo do
+  sh "ruby ./bin/sekka-jisyo convertN ./data/SKK-JISYO.L.201501           >  ./data/SEKKA-JISYO.N"
+  sh "ruby ./bin/sekka-jisyo convertN ./data/SKK-JISYO.L.hira-kata        >> ./data/SEKKA-JISYO.N"
+  sh "ruby ./bin/sekka-jisyo convertN ./data/SKK-JISYO.fullname           >> ./data/SEKKA-JISYO.N"
+  sh "ruby ./bin/sekka-jisyo convertN ./data/SKK-JISYO.jinmei             >> ./data/SEKKA-JISYO.N"
+  sh "ruby ./bin/sekka-jisyo convertN ./data/SKK-JISYO.station            >> ./data/SEKKA-JISYO.N"
+  sh "ruby ./bin/sekka-jisyo convertN ./data/SKK-JISYO.hiragana-phrase    >> ./data/SEKKA-JISYO.N"
+  sh "ruby ./bin/sekka-jisyo convertN ./data/SKK-JISYO.hiragana-phrase2   >> ./data/SEKKA-JISYO.N"
+  sh "ruby ./bin/sekka-jisyo convertN ./data/SKK-JISYO.hiragana-phrase3   >> ./data/SEKKA-JISYO.N"
 end
 
-task :jisyoL do
-  generateTypes.each {|x|
-    sh "ruby ./bin/sekka-jisyo convert#{x} ./data/SKK-JISYO.L.201501           >  ./data/SEKKA-JISYO.LARGE.#{x}"
-    sh "ruby ./bin/sekka-jisyo convert#{x} ./data/SKK-JISYO.L.hira-kata        >> ./data/SEKKA-JISYO.LARGE.#{x}"
-    sh "ruby ./bin/sekka-jisyo convert#{x} ./data/SKK-JISYO.fullname           >> ./data/SEKKA-JISYO.LARGE.#{x}"
-    sh "ruby ./bin/sekka-jisyo convert#{x} ./data/SKK-JISYO.jinmei             >> ./data/SEKKA-JISYO.LARGE.#{x}"
-    sh "ruby ./bin/sekka-jisyo convert#{x} ./data/SKK-JISYO.station            >> ./data/SEKKA-JISYO.LARGE.#{x}"
-    sh "ruby ./bin/sekka-jisyo convert#{x} ./data/SKK-JISYO.hiragana-phrase    >> ./data/SEKKA-JISYO.LARGE.#{x}"
-    sh "ruby ./bin/sekka-jisyo convert#{x} ./data/SKK-JISYO.hiragana-phrase2   >> ./data/SEKKA-JISYO.LARGE.#{x}"
-    sh "ruby ./bin/sekka-jisyo convert#{x} ./data/SKK-JISYO.hiragana-phrase3   >> ./data/SEKKA-JISYO.LARGE.#{x}"
-  }
+task :load do
+  sh "ruby ./bin/sekka-jisyo load    ./data/SEKKA-JISYO.N  ./data/SEKKA-JISYO.N.tch#xmsiz=1024m"
 end
 
-task :loadS do
-  generateTypes.each {|x|
-    sh "ruby ./bin/sekka-jisyo load    ./data/SEKKA-JISYO.SMALL.#{x}  ./data/SEKKA-JISYO.SMALL.#{x}.tch#xmsiz=1024m"
-  }
-end
-
-task :loadL do
-  generateTypes.each {|x|
-    sh "ruby ./bin/sekka-jisyo load    ./data/SEKKA-JISYO.LARGE.#{x}  ./data/SEKKA-JISYO.LARGE.#{x}.tch#xmsiz=1024m"
-  }
-end
-
-task :dumpS do
-  generateTypes.each {|x|
-    sh "ruby ./bin/sekka-jisyo dump    ./data/SEKKA-JISYO.SMALL.#{x}.tch#xmsiz=1024m > ./data/SEKKA-JISYO.SMALL.#{x}.tsv"
-  }
-end
-
-task :dumpL do
-  generateTypes.each {|x|
-    sh "ruby ./bin/sekka-jisyo dump    ./data/SEKKA-JISYO.LARGE.#{x}.tch#xmsiz=1024m > ./data/SEKKA-JISYO.LARGE.#{x}.tsv"
-  }
+task :dump do
+  sh "ruby ./bin/sekka-jisyo dump    ./data/SEKKA-JISYO.N.tch#xmsiz=1024m > ./data/SEKKA-JISYO.N.tsv"
 end
 
 # SKK-JISYO.hiragana-phrase はWikipediaから作られる。
 task :phrase => [ "/tmp/jawiki.txt.gz", "./data/wikipedia/jawiki.hiragana.txt" ] do
   sh "sort ./data/wikipedia/jawiki.hiragana.txt | uniq -c | sort > ./data/wikipedia/ranking.txt"
-  sh "ruby -I ./lib /usr/local/bin/nendo ./data/hiragana_phrase_in_wikipedia2.nnd ./data/wikipedia/ranking.txt > ./data/SKK-JISYO.hiragana-phrase"
+  sh "nendo -I ./lib ./data/hiragana_phrase_in_wikipedia2.nnd ./data/wikipedia/ranking.txt > ./data/SKK-JISYO.hiragana-phrase"
   sh "echo 'して //' >> ./data/SKK-JISYO.hiragana-phrase"
 end
 
 file "./data/wikipedia/jawiki.hiragana.txt" do
   sh "zcat /tmp/jawiki.txt.gz | mecab --input-buffer-size=65536 -O wakati --output=/tmp/jawiki.wakati.txt"
-  sh "ruby -I ./lib /usr/local/bin/nendo ./data/hiragana_phrase_in_wikipedia.nnd /tmp/jawiki.wakati.txt > ./data/wikipedia/jawiki.hiragana.txt"
+  sh "nendo -I ./lib ./data/hiragana_phrase_in_wikipedia.nnd /tmp/jawiki.wakati.txt > ./data/wikipedia/jawiki.hiragana.txt"
   sh "rm -f /tmp/jawiki.wakati.txt"
 end
 
@@ -203,7 +168,7 @@ end
 
 # SKK-JISYO.hiragana-phrase2 はIPADicから作られる。
 task :phrase2 => [ "./data/ipadic.all.utf8.txt" ] do
-  sh "time ruby -I ./lib /usr/local/bin/nendo ./data/hiragana_phrase_in_ipadic.nnd             ./data/ipadic.all.utf8.txt | sort | uniq > ./data/SKK-JISYO.hiragana-phrase2"
+  sh "time nendo -I ./data/hiragana_phrase_in_ipadic.nnd             ./data/ipadic.all.utf8.txt | sort | uniq > ./data/SKK-JISYO.hiragana-phrase2"
 end
 
 file "./data/ipadic.all.utf8.txt" do

data/VERSION.yml

@@ -1,4 +1,4 @@
 ---
 :major: 1
-:minor: 5
-:patch: 9
+:minor: 6
+:patch: 0

data/bin/sekka-server

@@ -16,31 +16,16 @@ DICTURL = "https://raw.githubusercontent.com/kiyoka/sekka/master/public_dict/" +
 
 PIDFILE = DICTDIR + "/pid"
 
-DICTTYPE = "N"
-if ENV.has_key?( 'SEKKA_AZIK' )
-  STDERR.printf( "Error: SEKKA_AZIK became obsolute. sekka-server use normal dictionary.\n" )
-end
-
-DICTSIZE = if ENV.has_key?( 'SEKKA_LARGE' )
-             if 1 == ENV[ 'SEKKA_LARGE' ].to_i
-               "LARGE"
-             else
-               "SMALL"
-             end
-           else
-             "SMALL"
-           end
-
 TC_OPTS = "#xmsiz=256m"
-TC_FILE = DICTDIR + "/SEKKA-JISYO.#{DICTSIZE}.#{DICTTYPE}.tch" + TC_OPTS
-TSVFILE = DICTDIR + "/SEKKA-JISYO.#{DICTSIZE}.#{DICTTYPE}.tsv"
-SUMFILE = DICTDIR + "/SEKKA-JISYO.#{DICTSIZE}.#{DICTTYPE}.md5"
+TC_FILE = DICTDIR + "/SEKKA-JISYO.N.tch" + TC_OPTS
+TSVFILE = DICTDIR + "/SEKKA-JISYO.N.tsv"
+SUMFILE = DICTDIR + "/SEKKA-JISYO.N.md5"
 
-GDBM_FILE = DICTDIR + "/SEKKA-JISYO.#{DICTSIZE}.#{DICTTYPE}.db"
+GDBM_FILE = DICTDIR + "/SEKKA-JISYO.N.db"
 
 
-URLURL  = DICTURL + "/SEKKA-JISYO.#{DICTSIZE}.#{DICTTYPE}.url"
-SUMURL  = DICTURL + "/SEKKA-JISYO.#{DICTSIZE}.#{DICTTYPE}.md5"
+URLURL  = DICTURL + "/SEKKA-JISYO.N.url"
+SUMURL  = DICTURL + "/SEKKA-JISYO.N.md5"
 
 MEMCACHED = "localhost:11211"                # memcahced
 

data/data/.gitignore

@@ -26,6 +26,8 @@
 /SEKKA-JISYO.LARGE.dump
 /SEKKA-JISYO.LARGE.tch
 /SEKKA-JISYO.N.SMALL
+/SEKKA-JISYO.N.md5
+/SEKKA-JISYO.N.tsv
 /SEKKA-JISYO.S.201001.dump
 /SEKKA-JISYO.S.201001.kct
 /SEKKA-JISYO.S.201001.tch

data/emacs/sekka.el

@@ -3,7 +3,7 @@
 ;; Copyright (C) 2010-2014 Kiyoka Nishiyama
 ;;
 ;; Author: Kiyoka Nishiyama <kiyoka@sumibi.org>
-;; Version: 1.5.9          ;;SEKKA-VERSION
+;; Version: 1.6.0          ;;SEKKA-VERSION
 ;; Keywords: ime, skk, japanese
 ;; Package-Requires: ((cl-lib "0.3") (concurrent "0.3.1") (popup "0.5.0"))
 ;; URL: https://github.com/kiyoka/sekka
@@ -1746,7 +1746,7 @@ point から行頭方向に同種の文字列が続く間を漢字変換しま
 (setq default-input-method "japanese-sekka")
 
 (defconst sekka-version
-  "1.5.9" ;;SEKKA-VERSION
+  "1.6.0" ;;SEKKA-VERSION
   )
 (defun sekka-version (&optional arg)
   "入力モード変更"

data/lib/sekka/roman-lib.nnd

@@ -737,10 +737,22 @@
 (define (is-katakana str)
   (if (rxmatch #/^[ア-ンァィゥェォャュョッー]+$/ str) #t #f))
 
-;; 平仮名の文字列かどうかを評価する
+;; ひらがなの文字列かどうかを評価する
 (define (is-hiragana str)
   (if (rxmatch #/^[あ-んぁぃぅぇぉゃゅょっー]+$/ str) #t #f))
 
+;; ひらがなの文字列を含むかどうかを評価する
+(define (include-hiragana str)
+  (if (rxmatch #/[あ-んぁぃぅぇぉゃゅょっー]+/ str) #t #f))
+
+;; 漢字の文字列かどうかを評価する
+(define (is-kanji str)
+  (if (rxmatch #/^[\p{Han}]+$/ str) #t #f))
+
+;; 漢字の文字列を含むかどうか評価する
+(define (include-kanji str)
+  (if (rxmatch #/[\p{Han}]/ str) #t #f))
+
 ;; 送り仮名付き平仮名文字列(例:"おこなu") かどうかを評価する
 (define (is-hiragana-and-okuri str)
   (if (rxmatch #/^[あ-んぁぃぅぇぉゃゅょっー]+[a-z]$/ str) #t #f))

data/lib/sekka/sekkaversion.rb

@@ -1,4 +1,4 @@
 class SekkaVersion
-  def  self.version()     "1.5.9"  end
-  def  self.dictVersion() "1.4.2"  end
+  def  self.version()     "1.6.0"  end
+  def  self.dictVersion() "1.5.0"  end
 end

data/public_dict/1.5.0/SEKKA-JISYO.N.md5

@@ -0,0 +1 @@
+3cbe0e13da141bda955fe8e53f378dd0  ./data/SEKKA-JISYO.N.tsv

data/public_dict/1.5.0/SEKKA-JISYO.N.url

@@ -0,0 +1 @@
+https://s3-ap-northeast-1.amazonaws.com/sekkadict/1.5.0/SEKKA-JISYO.N.tsv

data/sekka.gemspec

@@ -19,7 +19,7 @@ Gem::Specification.new do |spec|
   spec.require_paths = ["lib"]
   spec.add_dependency "eventmachine", "~> 1.0"
   spec.add_dependency "memcache-client", "~> 1.8"
-  spec.add_dependency "nendo", "= 0.7.1"
+  spec.add_dependency "nendo", "= 0.7.2"
   spec.add_dependency "distributed-trie", "= 0.8.0"
   spec.add_dependency "rack",  "~> 1.5"
   spec.add_dependency "ruby-progressbar", "~> 1.4"

data/test/roman-lib.nnd

@@ -251,6 +251,74 @@
        #f
        (is-hiragana-and-okuri "123"))
 
+;;-------------------------------------------------------------------
+(test-section "other judgement functions")
+
+(test* "include hiragana 1"
+       #t
+       (include-hiragana "123あ456"))
+
+(test* "include hiragana 2"
+       #f
+       (include-hiragana "123A456"))
+
+(test* "include hiragana 3"
+       #f
+       (include-hiragana "漢字"))
+
+(test* "include hiragana 4"
+       #f
+       (include-hiragana "カタカナ"))
+
+(test* "is kanji 1"
+       #t
+       (is-kanji "漢字"))
+
+(test* "is kanji 2"
+       #t
+       (is-kanji "薔薇"))
+
+(test* "is kanji 3"
+       #f
+       (is-kanji "感じ"))
+
+(test* "is kanji 4"
+       #f
+       (is-kanji "ひらがな"))
+
+(test* "is kanji 5"
+       #f
+       (is-kanji "ABCDE"))
+
+(test* "is kanji 6"
+       #f
+       (is-kanji "¢"))
+
+(test* "is kanji 7"
+       #f
+       (is-kanji "з"))
+
+(test* "include kanji 1"
+       #t
+       (include-kanji "感じ"))
+
+(test* "include kanji 2"
+       #t
+       (include-kanji "ABC漢字DEF"))
+
+(test* "include kanji 3"
+       #f
+       (include-kanji "ABCDEF"))
+
+(test* "include kanji 4"
+       #f
+       (include-kanji "ひらがな"))
+
+(test* "include kanji 5"
+       #f
+       (include-kanji "カタカナ"))
+
+
 ;;-------------------------------------------------------------------
 (test-section "drop okurigana functions")
 (test* "drop okri 1"

data/test/sekka-dump-out-1.txt

@@ -1868,4 +1868,4 @@ M:みr	/見/観;(字義:みわたす)/視;(字義:じっとみる)/診;(diagnose
 M:わけ	/訳/和気;地名/理由;当て字/分け/分/譯;「訳」の旧字/戯奴
 M:わたしh	/私
 M:を	/小/緒/雄/男/牡/尾/悪/汚/惡;「悪」の旧字(人名用漢字)
-SEKKA:VERSION	1.4.2
+SEKKA:VERSION	1.5.0

metadata

@@ -1,14 +1,14 @@
 --- !ruby/object:Gem::Specification
 name: sekka
 version: !ruby/object:Gem::Version
-  version: 1.5.9
+  version: 1.6.0
 platform: ruby
 authors:
 - Kiyoka Nishiyama
 autorequire: 
 bindir: bin
 cert_chain: []
-date: 2015-02-05 00:00:00.000000000 Z
+date: 2015-02-28 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: eventmachine
@@ -44,14 +44,14 @@ dependencies:
     requirements:
     - - '='
       - !ruby/object:Gem::Version
-        version: 0.7.1
+        version: 0.7.2
   type: :runtime
   prerelease: false
   version_requirements: !ruby/object:Gem::Requirement
     requirements:
     - - '='
       - !ruby/object:Gem::Version
-        version: 0.7.1
+        version: 0.7.2
 - !ruby/object:Gem::Dependency
   name: distributed-trie
   requirement: !ruby/object:Gem::Requirement
@@ -229,6 +229,8 @@ files:
 - public_dict/1.4.2/SEKKA-JISYO.LARGE.N.url
 - public_dict/1.4.2/SEKKA-JISYO.SMALL.N.md5
 - public_dict/1.4.2/SEKKA-JISYO.SMALL.N.url
+- public_dict/1.5.0/SEKKA-JISYO.N.md5
+- public_dict/1.5.0/SEKKA-JISYO.N.url
 - script/sekkaserver.debian
 - sekka.gemspec
 - test/.gitignore