eiwa 0.0.1 → 0.0.2
This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
- checksums.yaml +4 -4
- data/Gemfile.lock +4 -4
- data/README.md +2 -0
- data/eiwa.gemspec +1 -1
- data/lib/eiwa/jmdict_doc.rb +1 -1
- data/lib/eiwa/jmdict_entities.rb +77 -76
- data/lib/eiwa/version.rb +1 -1
- metadata +4 -4
checksums.yaml
CHANGED
|
@@ -1,7 +1,7 @@
|
|
|
1
1
|
---
|
|
2
2
|
SHA256:
|
|
3
|
-
metadata.gz:
|
|
4
|
-
data.tar.gz:
|
|
3
|
+
metadata.gz: c5888e4802408cc8efdb55ddadfb560ec38d10971fee95b20dd53af2f31f487c
|
|
4
|
+
data.tar.gz: 93b7101b430ee123a905065f87e5d8ba336e1a18f145c7801daeb2b9b9a5ba72
|
|
5
5
|
SHA512:
|
|
6
|
-
metadata.gz:
|
|
7
|
-
data.tar.gz:
|
|
6
|
+
metadata.gz: 64faccd9958b9c359fcd7a7ff40de013bd1060bbf9a59735be4d52c824dd3e6e77abbb81ab42ae037b9175990dd15baa33722503b47d8aaebbb037a8a303f965
|
|
7
|
+
data.tar.gz: 209efe931acfa8563ea1819f4e9b5a7d07a996d1545458b139de14f82971b27f47e0da726ed5151b3922aa05f6c5809bd3196269e0f6d254cb87e801c81ffe6e
|
data/Gemfile.lock
CHANGED
|
@@ -1,7 +1,7 @@
|
|
|
1
1
|
PATH
|
|
2
2
|
remote: .
|
|
3
3
|
specs:
|
|
4
|
-
eiwa (0.0.
|
|
4
|
+
eiwa (0.0.2)
|
|
5
5
|
nokogiri
|
|
6
6
|
|
|
7
7
|
GEM
|
|
@@ -13,7 +13,7 @@ GEM
|
|
|
13
13
|
method_source (0.9.2)
|
|
14
14
|
mini_portile2 (2.4.0)
|
|
15
15
|
minitest (5.11.3)
|
|
16
|
-
nokogiri (1.10.
|
|
16
|
+
nokogiri (1.10.9)
|
|
17
17
|
mini_portile2 (~> 2.4.0)
|
|
18
18
|
parallel (1.17.0)
|
|
19
19
|
parser (2.6.4.1)
|
|
@@ -22,7 +22,7 @@ GEM
|
|
|
22
22
|
coderay (~> 1.1.0)
|
|
23
23
|
method_source (~> 0.9.0)
|
|
24
24
|
rainbow (3.0.0)
|
|
25
|
-
rake (
|
|
25
|
+
rake (13.0.1)
|
|
26
26
|
rubocop (0.72.0)
|
|
27
27
|
jaro_winkler (~> 1.5.1)
|
|
28
28
|
parallel (~> 1.10)
|
|
@@ -46,7 +46,7 @@ DEPENDENCIES
|
|
|
46
46
|
eiwa!
|
|
47
47
|
minitest (~> 5.0)
|
|
48
48
|
pry
|
|
49
|
-
rake (~>
|
|
49
|
+
rake (~> 13.0)
|
|
50
50
|
standard
|
|
51
51
|
|
|
52
52
|
BUNDLED WITH
|
data/README.md
CHANGED
|
@@ -74,4 +74,6 @@ entries = Eiwa.parse_file("path/to/some.xml", type: :jmdict_e)
|
|
|
74
74
|
Note that for the abridged Japanese-English dictionary, this will consume about
|
|
75
75
|
500MB of RAM.
|
|
76
76
|
|
|
77
|
+
### The entry object model
|
|
77
78
|
|
|
79
|
+
I haven't documented the [Entry](https://github.com/searls/eiwa/blob/master/lib/eiwa/tag/entry.rb) type or its child types yet, but they should be pretty easy to piece together by inspecting the output and [checking the source listings](https://github.com/searls/eiwa/blob/master/lib/eiwa/tag).
|
data/eiwa.gemspec
CHANGED
|
@@ -21,7 +21,7 @@ Gem::Specification.new do |spec|
|
|
|
21
21
|
|
|
22
22
|
spec.add_dependency "nokogiri"
|
|
23
23
|
spec.add_development_dependency "bundler", "~> 1.17"
|
|
24
|
-
spec.add_development_dependency "rake", "~>
|
|
24
|
+
spec.add_development_dependency "rake", "~> 13.0"
|
|
25
25
|
spec.add_development_dependency "minitest", "~> 5.0"
|
|
26
26
|
spec.add_development_dependency "standard"
|
|
27
27
|
spec.add_development_dependency "pry"
|
data/lib/eiwa/jmdict_doc.rb
CHANGED
data/lib/eiwa/jmdict_entities.rb
CHANGED
|
@@ -1,39 +1,57 @@
|
|
|
1
1
|
module Eiwa
|
|
2
2
|
JMDICT_ENTITIES = {
|
|
3
|
+
"Buddh" => "Buddhist term",
|
|
3
4
|
"MA" => "martial arts term",
|
|
5
|
+
"Shinto" => "Shinto term",
|
|
4
6
|
"X" => "rude or X-rated term (not displayed in educational software)",
|
|
5
7
|
"abbr" => "abbreviation",
|
|
8
|
+
"adj-f" => "noun or verb acting prenominally",
|
|
6
9
|
"adj-i" => "adjective (keiyoushi)",
|
|
7
10
|
"adj-ix" => "adjective (keiyoushi) - yoi/ii class",
|
|
11
|
+
"adj-kari" => "`kari' adjective (archaic)",
|
|
12
|
+
"adj-ku" => "`ku' adjective (archaic)",
|
|
8
13
|
"adj-na" => "adjectival nouns or quasi-adjectives (keiyodoshi)",
|
|
14
|
+
"adj-nari" => "archaic/formal form of na-adjective",
|
|
9
15
|
"adj-no" => "nouns which may take the genitive case particle `no'",
|
|
10
16
|
"adj-pn" => "pre-noun adjectival (rentaishi)",
|
|
17
|
+
"adj-shiku" => "`shiku' adjective (archaic)",
|
|
11
18
|
"adj-t" => "`taru' adjective",
|
|
12
|
-
"adj-f" => "noun or verb acting prenominally",
|
|
13
19
|
"adv" => "adverb (fukushi)",
|
|
14
20
|
"adv-to" => "adverb taking the `to' particle",
|
|
21
|
+
"anat" => "anatomical term",
|
|
15
22
|
"arch" => "archaism",
|
|
23
|
+
"archit" => "architecture term",
|
|
24
|
+
"astron" => "astronomy, etc. term",
|
|
16
25
|
"ateji" => "ateji (phonetic) reading",
|
|
17
26
|
"aux" => "auxiliary",
|
|
18
|
-
"aux-v" => "auxiliary verb",
|
|
19
27
|
"aux-adj" => "auxiliary adjective",
|
|
20
|
-
"
|
|
28
|
+
"aux-v" => "auxiliary verb",
|
|
29
|
+
"baseb" => "baseball term",
|
|
30
|
+
"biol" => "biology term",
|
|
31
|
+
"bot" => "botany term",
|
|
32
|
+
"bus" => "business term",
|
|
21
33
|
"chem" => "chemistry term",
|
|
22
34
|
"chn" => "children's language",
|
|
23
35
|
"col" => "colloquialism",
|
|
24
36
|
"comp" => "computer terminology",
|
|
25
37
|
"conj" => "conjunction",
|
|
38
|
+
"cop" => "copula",
|
|
26
39
|
"cop-da" => "copula",
|
|
27
40
|
"ctr" => "counter",
|
|
28
41
|
"derog" => "derogatory",
|
|
29
42
|
"eK" => "exclusively kanji",
|
|
43
|
+
"econ" => "economics term",
|
|
30
44
|
"ek" => "exclusively kana",
|
|
45
|
+
"engr" => "engineering term",
|
|
31
46
|
"exp" => "expressions (phrases, clauses, etc.)",
|
|
32
47
|
"fam" => "familiar language",
|
|
33
48
|
"fem" => "female term or language",
|
|
49
|
+
"finc" => "finance term",
|
|
34
50
|
"food" => "food term",
|
|
51
|
+
"geol" => "geology, etc. term",
|
|
35
52
|
"geom" => "geometry term",
|
|
36
53
|
"gikun" => "gikun (meaning as reading) or jukujikun (special kanji reading)",
|
|
54
|
+
"hob" => "Hokkaido-ben",
|
|
37
55
|
"hon" => "honorific or respectful (sonkeigo) language",
|
|
38
56
|
"hum" => "humble (kenjougo) language",
|
|
39
57
|
"iK" => "word containing irregular kanji usage",
|
|
@@ -42,45 +60,93 @@ module Eiwa
|
|
|
42
60
|
"int" => "interjection (kandoushi)",
|
|
43
61
|
"io" => "irregular okurigana usage",
|
|
44
62
|
"iv" => "irregular verb",
|
|
63
|
+
"joc" => "jocular, humorous term",
|
|
64
|
+
"ksb" => "Kansai-ben",
|
|
65
|
+
"ktb" => "Kantou-ben",
|
|
66
|
+
"kyb" => "Kyoto-ben",
|
|
67
|
+
"kyu" => "Kyuushuu-ben",
|
|
68
|
+
"law" => "law, etc. term",
|
|
45
69
|
"ling" => "linguistics terminology",
|
|
46
70
|
"m-sl" => "manga slang",
|
|
71
|
+
"mahj" => "mahjong term",
|
|
47
72
|
"male" => "male term or language",
|
|
48
73
|
"male-sl" => "male slang",
|
|
49
74
|
"math" => "mathematics",
|
|
75
|
+
"med" => "medicine, etc. term",
|
|
50
76
|
"mil" => "military",
|
|
77
|
+
"music" => "music term",
|
|
51
78
|
"n" => "noun (common) (futsuumeishi)",
|
|
52
79
|
"n-adv" => "adverbial noun (fukushitekimeishi)",
|
|
53
|
-
"n-
|
|
80
|
+
"n-pr" => "proper noun",
|
|
54
81
|
"n-pref" => "noun, used as a prefix",
|
|
82
|
+
"n-suf" => "noun, used as a suffix",
|
|
55
83
|
"n-t" => "noun (temporal) (jisoumeishi)",
|
|
84
|
+
"nab" => "Nagano-ben",
|
|
56
85
|
"num" => "numeric",
|
|
57
86
|
"oK" => "word containing out-dated kanji",
|
|
58
87
|
"obs" => "obsolete term",
|
|
59
88
|
"obsc" => "obscure term",
|
|
60
|
-
"ok" => "out-dated or obsolete kana usage",
|
|
61
89
|
"oik" => "old or irregular kana form",
|
|
90
|
+
"ok" => "out-dated or obsolete kana usage",
|
|
62
91
|
"on-mim" => "onomatopoeic or mimetic word",
|
|
92
|
+
"osb" => "Osaka-ben",
|
|
93
|
+
"physics" => "physics terminology",
|
|
63
94
|
"pn" => "pronoun",
|
|
64
95
|
"poet" => "poetical term",
|
|
65
96
|
"pol" => "polite (teineigo) language",
|
|
66
97
|
"pref" => "prefix",
|
|
67
98
|
"proverb" => "proverb",
|
|
68
99
|
"prt" => "particle",
|
|
69
|
-
"physics" => "physics terminology",
|
|
70
100
|
"quote" => "quotation",
|
|
71
101
|
"rare" => "rare",
|
|
102
|
+
"rkb" => "Ryuukyuu-ben",
|
|
72
103
|
"sens" => "sensitive",
|
|
104
|
+
"shogi" => "shogi term",
|
|
73
105
|
"sl" => "slang",
|
|
106
|
+
"sports" => "sports term",
|
|
74
107
|
"suf" => "suffix",
|
|
108
|
+
"sumo" => "sumo term",
|
|
109
|
+
"thb" => "Touhoku-ben",
|
|
110
|
+
"tsb" => "Tosa-ben",
|
|
111
|
+
"tsug" => "Tsugaru-ben",
|
|
75
112
|
"uK" => "word usually written using kanji alone",
|
|
76
113
|
"uk" => "word usually written using kana alone",
|
|
77
114
|
"unc" => "unclassified",
|
|
78
|
-
"
|
|
115
|
+
"v-unspec" => "verb unspecified",
|
|
79
116
|
"v1" => "Ichidan verb",
|
|
80
117
|
"v1-s" => "Ichidan verb - kureru special class",
|
|
81
118
|
"v2a-s" => "Nidan verb with 'u' ending (archaic)",
|
|
119
|
+
"v2b-k" => "Nidan verb (upper class) with `bu' ending (archaic)",
|
|
120
|
+
"v2b-s" => "Nidan verb (lower class) with `bu' ending (archaic)",
|
|
121
|
+
"v2d-k" => "Nidan verb (upper class) with `dzu' ending (archaic)",
|
|
122
|
+
"v2d-s" => "Nidan verb (lower class) with `dzu' ending (archaic)",
|
|
123
|
+
"v2g-k" => "Nidan verb (upper class) with `gu' ending (archaic)",
|
|
124
|
+
"v2g-s" => "Nidan verb (lower class) with `gu' ending (archaic)",
|
|
125
|
+
"v2h-k" => "Nidan verb (upper class) with `hu/fu' ending (archaic)",
|
|
126
|
+
"v2h-s" => "Nidan verb (lower class) with `hu/fu' ending (archaic)",
|
|
127
|
+
"v2k-k" => "Nidan verb (upper class) with `ku' ending (archaic)",
|
|
128
|
+
"v2k-s" => "Nidan verb (lower class) with `ku' ending (archaic)",
|
|
129
|
+
"v2m-k" => "Nidan verb (upper class) with `mu' ending (archaic)",
|
|
130
|
+
"v2m-s" => "Nidan verb (lower class) with `mu' ending (archaic)",
|
|
131
|
+
"v2n-s" => "Nidan verb (lower class) with `nu' ending (archaic)",
|
|
132
|
+
"v2r-k" => "Nidan verb (upper class) with `ru' ending (archaic)",
|
|
133
|
+
"v2r-s" => "Nidan verb (lower class) with `ru' ending (archaic)",
|
|
134
|
+
"v2s-s" => "Nidan verb (lower class) with `su' ending (archaic)",
|
|
135
|
+
"v2t-k" => "Nidan verb (upper class) with `tsu' ending (archaic)",
|
|
136
|
+
"v2t-s" => "Nidan verb (lower class) with `tsu' ending (archaic)",
|
|
137
|
+
"v2w-s" => "Nidan verb (lower class) with `u' ending and `we' conjugation (archaic)",
|
|
138
|
+
"v2y-k" => "Nidan verb (upper class) with `yu' ending (archaic)",
|
|
139
|
+
"v2y-s" => "Nidan verb (lower class) with `yu' ending (archaic)",
|
|
140
|
+
"v2z-s" => "Nidan verb (lower class) with `zu' ending (archaic)",
|
|
141
|
+
"v4b" => "Yodan verb with `bu' ending (archaic)",
|
|
142
|
+
"v4g" => "Yodan verb with `gu' ending (archaic)",
|
|
82
143
|
"v4h" => "Yodan verb with `hu/fu' ending (archaic)",
|
|
144
|
+
"v4k" => "Yodan verb with `ku' ending (archaic)",
|
|
145
|
+
"v4m" => "Yodan verb with `mu' ending (archaic)",
|
|
146
|
+
"v4n" => "Yodan verb with `nu' ending (archaic)",
|
|
83
147
|
"v4r" => "Yodan verb with `ru' ending (archaic)",
|
|
148
|
+
"v4s" => "Yodan verb with `su' ending (archaic)",
|
|
149
|
+
"v4t" => "Yodan verb with `tsu' ending (archaic)",
|
|
84
150
|
"v5aru" => "Godan verb - -aru special class",
|
|
85
151
|
"v5b" => "Godan verb with `bu' ending",
|
|
86
152
|
"v5g" => "Godan verb with `gu' ending",
|
|
@@ -95,83 +161,18 @@ module Eiwa
|
|
|
95
161
|
"v5u" => "Godan verb with `u' ending",
|
|
96
162
|
"v5u-s" => "Godan verb with `u' ending (special class)",
|
|
97
163
|
"v5uru" => "Godan verb - Uru old class verb (old form of Eru)",
|
|
98
|
-
"vz" => "Ichidan verb - zuru verb (alternative form of -jiru verbs)",
|
|
99
164
|
"vi" => "intransitive verb",
|
|
100
165
|
"vk" => "Kuru verb - special class",
|
|
101
166
|
"vn" => "irregular nu verb",
|
|
102
167
|
"vr" => "irregular ru verb, plain form ends with -ri",
|
|
103
168
|
"vs" => "noun or participle which takes the aux. verb suru",
|
|
104
169
|
"vs-c" => "su verb - precursor to the modern suru",
|
|
105
|
-
"vs-s" => "suru verb - special class",
|
|
106
170
|
"vs-i" => "suru verb - included",
|
|
107
|
-
"
|
|
108
|
-
"osb" => "Osaka-ben",
|
|
109
|
-
"ksb" => "Kansai-ben",
|
|
110
|
-
"ktb" => "Kantou-ben",
|
|
111
|
-
"tsb" => "Tosa-ben",
|
|
112
|
-
"thb" => "Touhoku-ben",
|
|
113
|
-
"tsug" => "Tsugaru-ben",
|
|
114
|
-
"kyu" => "Kyuushuu-ben",
|
|
115
|
-
"rkb" => "Ryuukyuu-ben",
|
|
116
|
-
"nab" => "Nagano-ben",
|
|
117
|
-
"hob" => "Hokkaido-ben",
|
|
171
|
+
"vs-s" => "suru verb - special class",
|
|
118
172
|
"vt" => "transitive verb",
|
|
119
173
|
"vulg" => "vulgar expression or word",
|
|
120
|
-
"
|
|
121
|
-
"
|
|
122
|
-
"
|
|
123
|
-
"adj-nari" => "archaic/formal form of na-adjective",
|
|
124
|
-
"n-pr" => "proper noun",
|
|
125
|
-
"v-unspec" => "verb unspecified",
|
|
126
|
-
"v4k" => "Yodan verb with `ku' ending (archaic)",
|
|
127
|
-
"v4g" => "Yodan verb with `gu' ending (archaic)",
|
|
128
|
-
"v4s" => "Yodan verb with `su' ending (archaic)",
|
|
129
|
-
"v4t" => "Yodan verb with `tsu' ending (archaic)",
|
|
130
|
-
"v4n" => "Yodan verb with `nu' ending (archaic)",
|
|
131
|
-
"v4b" => "Yodan verb with `bu' ending (archaic)",
|
|
132
|
-
"v4m" => "Yodan verb with `mu' ending (archaic)",
|
|
133
|
-
"v2k-k" => "Nidan verb (upper class) with `ku' ending (archaic)",
|
|
134
|
-
"v2g-k" => "Nidan verb (upper class) with `gu' ending (archaic)",
|
|
135
|
-
"v2t-k" => "Nidan verb (upper class) with `tsu' ending (archaic)",
|
|
136
|
-
"v2d-k" => "Nidan verb (upper class) with `dzu' ending (archaic)",
|
|
137
|
-
"v2h-k" => "Nidan verb (upper class) with `hu/fu' ending (archaic)",
|
|
138
|
-
"v2b-k" => "Nidan verb (upper class) with `bu' ending (archaic)",
|
|
139
|
-
"v2m-k" => "Nidan verb (upper class) with `mu' ending (archaic)",
|
|
140
|
-
"v2y-k" => "Nidan verb (upper class) with `yu' ending (archaic)",
|
|
141
|
-
"v2r-k" => "Nidan verb (upper class) with `ru' ending (archaic)",
|
|
142
|
-
"v2k-s" => "Nidan verb (lower class) with `ku' ending (archaic)",
|
|
143
|
-
"v2g-s" => "Nidan verb (lower class) with `gu' ending (archaic)",
|
|
144
|
-
"v2s-s" => "Nidan verb (lower class) with `su' ending (archaic)",
|
|
145
|
-
"v2z-s" => "Nidan verb (lower class) with `zu' ending (archaic)",
|
|
146
|
-
"v2t-s" => "Nidan verb (lower class) with `tsu' ending (archaic)",
|
|
147
|
-
"v2d-s" => "Nidan verb (lower class) with `dzu' ending (archaic)",
|
|
148
|
-
"v2n-s" => "Nidan verb (lower class) with `nu' ending (archaic)",
|
|
149
|
-
"v2h-s" => "Nidan verb (lower class) with `hu/fu' ending (archaic)",
|
|
150
|
-
"v2b-s" => "Nidan verb (lower class) with `bu' ending (archaic)",
|
|
151
|
-
"v2m-s" => "Nidan verb (lower class) with `mu' ending (archaic)",
|
|
152
|
-
"v2y-s" => "Nidan verb (lower class) with `yu' ending (archaic)",
|
|
153
|
-
"v2r-s" => "Nidan verb (lower class) with `ru' ending (archaic)",
|
|
154
|
-
"v2w-s" => "Nidan verb (lower class) with `u' ending and `we' conjugation (archaic)",
|
|
155
|
-
"archit" => "architecture term",
|
|
156
|
-
"astron" => "astronomy, etc. term",
|
|
157
|
-
"baseb" => "baseball term",
|
|
158
|
-
"biol" => "biology term",
|
|
159
|
-
"bot" => "botany term",
|
|
160
|
-
"bus" => "business term",
|
|
161
|
-
"econ" => "economics term",
|
|
162
|
-
"engr" => "engineering term",
|
|
163
|
-
"finc" => "finance term",
|
|
164
|
-
"geol" => "geology, etc. term",
|
|
165
|
-
"law" => "law, etc. term",
|
|
166
|
-
"mahj" => "mahjong term",
|
|
167
|
-
"med" => "medicine, etc. term",
|
|
168
|
-
"music" => "music term",
|
|
169
|
-
"Shinto" => "Shinto term",
|
|
170
|
-
"shogi" => "shogi term",
|
|
171
|
-
"sports" => "sports term",
|
|
172
|
-
"sumo" => "sumo term",
|
|
173
|
-
"zool" => "zoology term",
|
|
174
|
-
"joc" => "jocular, humorous term",
|
|
175
|
-
"anat" => "anatomical term",
|
|
174
|
+
"vz" => "Ichidan verb - zuru verb (alternative form of -jiru verbs)",
|
|
175
|
+
"yoji" => "yojijukugo",
|
|
176
|
+
"zool" => "zoology term"
|
|
176
177
|
}
|
|
177
178
|
end
|
data/lib/eiwa/version.rb
CHANGED
metadata
CHANGED
|
@@ -1,14 +1,14 @@
|
|
|
1
1
|
--- !ruby/object:Gem::Specification
|
|
2
2
|
name: eiwa
|
|
3
3
|
version: !ruby/object:Gem::Version
|
|
4
|
-
version: 0.0.
|
|
4
|
+
version: 0.0.2
|
|
5
5
|
platform: ruby
|
|
6
6
|
authors:
|
|
7
7
|
- Justin Searls
|
|
8
8
|
autorequire:
|
|
9
9
|
bindir: exe
|
|
10
10
|
cert_chain: []
|
|
11
|
-
date:
|
|
11
|
+
date: 2020-03-09 00:00:00.000000000 Z
|
|
12
12
|
dependencies:
|
|
13
13
|
- !ruby/object:Gem::Dependency
|
|
14
14
|
name: nokogiri
|
|
@@ -44,14 +44,14 @@ dependencies:
|
|
|
44
44
|
requirements:
|
|
45
45
|
- - "~>"
|
|
46
46
|
- !ruby/object:Gem::Version
|
|
47
|
-
version: '
|
|
47
|
+
version: '13.0'
|
|
48
48
|
type: :development
|
|
49
49
|
prerelease: false
|
|
50
50
|
version_requirements: !ruby/object:Gem::Requirement
|
|
51
51
|
requirements:
|
|
52
52
|
- - "~>"
|
|
53
53
|
- !ruby/object:Gem::Version
|
|
54
|
-
version: '
|
|
54
|
+
version: '13.0'
|
|
55
55
|
- !ruby/object:Gem::Dependency
|
|
56
56
|
name: minitest
|
|
57
57
|
requirement: !ruby/object:Gem::Requirement
|