Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Deinflect Json overhaul #433

Merged
merged 18 commits into from
Feb 4, 2024
Merged

Deinflect Json overhaul #433

Show file tree
Hide file tree
Changes from 13 commits
Commits
Show all changes
18 commits
Select commit Hold shift + click to select a range
c6f43e8
deinflect overhaul template
Casheeew Dec 23, 2023
eb357c8
remove duplicates with -e rule
Casheeew Dec 23, 2023
72dfd56
deinflection updates
Casheeew Dec 27, 2023
8b9700b
add neba deinflections
Casheeew Dec 27, 2023
4f0fb83
kansaiben updates
Casheeew Dec 28, 2023
de71406
remove redundancy
Casheeew Dec 29, 2023
f58d427
Revert "remove redundancy"
Casheeew Dec 29, 2023
6c81cc7
merge upstream
Casheeew Feb 2, 2024
0c9c1ec
update deinflect.json
Casheeew Feb 2, 2024
f0a8dd5
move deinflect.json to japanese-transforms.json
Casheeew Feb 2, 2024
4294fb1
remove trailing spaces
Casheeew Feb 2, 2024
48bb95a
wip add tests
Casheeew Feb 2, 2024
ab2b5e1
update deinflect.json, add tests
Casheeew Feb 2, 2024
f1a9321
run 'npm run test-code-write'
Casheeew Feb 3, 2024
4adc8ab
rename to kansai-ben
Casheeew Feb 3, 2024
bbda8b8
update tests
Casheeew Feb 3, 2024
cb99fc1
run 'npm run test-code-write' again
Casheeew Feb 3, 2024
df75839
add -tara deinflection
Casheeew Feb 3, 2024
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
206 changes: 204 additions & 2 deletions ext/data/language/japanese-transforms.json
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -410,6 +410,69 @@
{"suffixIn": "來ぬ", "suffixOut": "來る", "conditionsIn": [], "conditionsOut": ["vk"]}
]
},
{
"name": "-mu",
"rules": [
{"suffixIn": "む", "suffixOut": "る", "conditionsIn": [], "conditionsOut": ["v1"]},
{"suffixIn": "かむ", "suffixOut": "く", "conditionsIn": [], "conditionsOut": ["v5"]},
{"suffixIn": "がむ", "suffixOut": "ぐ", "conditionsIn": [], "conditionsOut": ["v5"]},
{"suffixIn": "さむ", "suffixOut": "す", "conditionsIn": [], "conditionsOut": ["v5"]},
{"suffixIn": "たむ", "suffixOut": "つ", "conditionsIn": [], "conditionsOut": ["v5"]},
{"suffixIn": "なむ", "suffixOut": "ぬ", "conditionsIn": [], "conditionsOut": ["v5"]},
{"suffixIn": "ばむ", "suffixOut": "ぶ", "conditionsIn": [], "conditionsOut": ["v5"]},
{"suffixIn": "まむ", "suffixOut": "む", "conditionsIn": [], "conditionsOut": ["v5"]},
{"suffixIn": "らむ", "suffixOut": "る", "conditionsIn": [], "conditionsOut": ["v5"]},
{"suffixIn": "わむ", "suffixOut": "う", "conditionsIn": [], "conditionsOut": ["v5"]},
{"suffixIn": "ぜむ", "suffixOut": "ずる", "conditionsIn": [], "conditionsOut": ["vz"]},
{"suffixIn": "せむ", "suffixOut": "する", "conditionsIn": [], "conditionsOut": ["vs"]},
{"suffixIn": "為む", "suffixOut": "為る", "conditionsIn": [], "conditionsOut": ["vs"]},
{"suffixIn": "こむ", "suffixOut": "くる", "conditionsIn": [], "conditionsOut": ["vk"]},
{"suffixIn": "来む", "suffixOut": "来る", "conditionsIn": [], "conditionsOut": ["vk"]},
{"suffixIn": "來む", "suffixOut": "來る", "conditionsIn": [], "conditionsOut": ["vk"]}
]
},
{
"name": "-zaru",
"rules": [
{"suffixIn": "ざる", "suffixOut": "る", "conditionsIn": [], "conditionsOut": ["v1"]},
{"suffixIn": "かざる", "suffixOut": "く", "conditionsIn": [], "conditionsOut": ["v5"]},
{"suffixIn": "がざる", "suffixOut": "ぐ", "conditionsIn": [], "conditionsOut": ["v5"]},
{"suffixIn": "さざる", "suffixOut": "す", "conditionsIn": [], "conditionsOut": ["v5"]},
{"suffixIn": "たざる", "suffixOut": "つ", "conditionsIn": [], "conditionsOut": ["v5"]},
{"suffixIn": "なざる", "suffixOut": "ぬ", "conditionsIn": [], "conditionsOut": ["v5"]},
{"suffixIn": "ばざる", "suffixOut": "ぶ", "conditionsIn": [], "conditionsOut": ["v5"]},
{"suffixIn": "まざる", "suffixOut": "む", "conditionsIn": [], "conditionsOut": ["v5"]},
{"suffixIn": "らざる", "suffixOut": "る", "conditionsIn": [], "conditionsOut": ["v5"]},
{"suffixIn": "わざる", "suffixOut": "う", "conditionsIn": [], "conditionsOut": ["v5"]},
{"suffixIn": "ぜざる", "suffixOut": "ずる", "conditionsIn": [], "conditionsOut": ["vz"]},
{"suffixIn": "せざる", "suffixOut": "する", "conditionsIn": [], "conditionsOut": ["vs"]},
{"suffixIn": "為ざる", "suffixOut": "為る", "conditionsIn": [], "conditionsOut": ["vs"]},
{"suffixIn": "こざる", "suffixOut": "くる", "conditionsIn": [], "conditionsOut": ["vk"]},
{"suffixIn": "来ざる", "suffixOut": "来る", "conditionsIn": [], "conditionsOut": ["vk"]},
{"suffixIn": "來ざる", "suffixOut": "來る", "conditionsIn": [], "conditionsOut": ["vk"]}
]
},
{
"name": "-neba",
"rules": [
{"suffixIn": "ねば", "suffixOut": "る", "conditionsIn": [], "conditionsOut": ["v1"]},
{"suffixIn": "かねば", "suffixOut": "く", "conditionsIn": [], "conditionsOut": ["v5"]},
{"suffixIn": "がねば", "suffixOut": "ぐ", "conditionsIn": [], "conditionsOut": ["v5"]},
{"suffixIn": "さねば", "suffixOut": "す", "conditionsIn": [], "conditionsOut": ["v5"]},
{"suffixIn": "たねば", "suffixOut": "つ", "conditionsIn": [], "conditionsOut": ["v5"]},
{"suffixIn": "なねば", "suffixOut": "ぬ", "conditionsIn": [], "conditionsOut": ["v5"]},
{"suffixIn": "ばねば", "suffixOut": "ぶ", "conditionsIn": [], "conditionsOut": ["v5"]},
{"suffixIn": "まねば", "suffixOut": "む", "conditionsIn": [], "conditionsOut": ["v5"]},
{"suffixIn": "らねば", "suffixOut": "る", "conditionsIn": [], "conditionsOut": ["v5"]},
{"suffixIn": "わねば", "suffixOut": "う", "conditionsIn": [], "conditionsOut": ["v5"]},
{"suffixIn": "ぜねば", "suffixOut": "ずる", "conditionsIn": [], "conditionsOut": ["vz"]},
{"suffixIn": "せねば", "suffixOut": "する", "conditionsIn": [], "conditionsOut": ["vs"]},
{"suffixIn": "為ねば", "suffixOut": "為る", "conditionsIn": [], "conditionsOut": ["vs"]},
{"suffixIn": "こねば", "suffixOut": "くる", "conditionsIn": [], "conditionsOut": ["vk"]},
{"suffixIn": "来ねば", "suffixOut": "来る", "conditionsIn": [], "conditionsOut": ["vk"]},
{"suffixIn": "來ねば", "suffixOut": "來る", "conditionsIn": [], "conditionsOut": ["vk"]}
]
},
{
"name": "adv",
"rules": [
Expand Down Expand Up @@ -844,8 +907,147 @@
{"suffixIn": "ちぇえ", "suffixOut": "ちゃい", "conditionsIn": [], "conditionsOut": ["adj-i"]},
{"suffixIn": "でえ", "suffixOut": "どい", "conditionsIn": [], "conditionsOut": ["adj-i"]},
{"suffixIn": "れえ", "suffixOut": "れい", "conditionsIn": [], "conditionsOut": ["adj-i"]},
{"suffixIn": "べえ", "suffixOut": "ばい", "conditionsIn": [], "conditionsOut": ["adj-i"]}
{"suffixIn": "べえ", "suffixOut": "ばい", "conditionsIn": [], "conditionsOut": ["adj-i"]},
{"suffixIn": "てえ", "suffixOut": "たい", "conditionsIn": [], "conditionsOut": ["adj-i"]},
{"suffixIn": "ねぇ", "suffixOut": "ない", "conditionsIn": [], "conditionsOut": ["adj-i"]},
{"suffixIn": "めぇ", "suffixOut": "むい", "conditionsIn": [], "conditionsOut": ["adj-i"]},
{"suffixIn": "みぃ", "suffixOut": "むい", "conditionsIn": [], "conditionsOut": ["adj-i"]},
{"suffixIn": "ちぃ", "suffixOut": "つい", "conditionsIn": [], "conditionsOut": ["adj-i"]},
{"suffixIn": "せぇ", "suffixOut": "すい", "conditionsIn": [], "conditionsOut": ["adj-i"]},
{"suffixIn": "けぇ", "suffixOut": "かい", "conditionsIn": [], "conditionsOut": ["adj-i"]},
{"suffixIn": "げぇ", "suffixOut": "がい", "conditionsIn": [], "conditionsOut": ["adj-i"]},
{"suffixIn": "げぇ", "suffixOut": "ごい", "conditionsIn": [], "conditionsOut": ["adj-i"]},
{"suffixIn": "せぇ", "suffixOut": "さい", "conditionsIn": [], "conditionsOut": ["adj-i"]},
{"suffixIn": "めぇ", "suffixOut": "まい", "conditionsIn": [], "conditionsOut": ["adj-i"]},
{"suffixIn": "ぜぇ", "suffixOut": "ずい", "conditionsIn": [], "conditionsOut": ["adj-i"]},
{"suffixIn": "っぜぇ", "suffixOut": "ずい", "conditionsIn": [], "conditionsOut": ["adj-i"]},
{"suffixIn": "れぇ", "suffixOut": "らい", "conditionsIn": [], "conditionsOut": ["adj-i"]},
{"suffixIn": "でぇ", "suffixOut": "どい", "conditionsIn": [], "conditionsOut": ["adj-i"]},
{"suffixIn": "れぇ", "suffixOut": "れい", "conditionsIn": [], "conditionsOut": ["adj-i"]},
{"suffixIn": "べぇ", "suffixOut": "ばい", "conditionsIn": [], "conditionsOut": ["adj-i"]},
{"suffixIn": "てぇ", "suffixOut": "たい", "conditionsIn": [], "conditionsOut": ["adj-i"]}
]
},
{
"name": "slang",
"rules": [
{"suffixIn": "てぇてぇ", "suffixOut": "とうとい", "conditionsIn": [], "conditionsOut": ["adj-i"]},
{"suffixIn": "てぇてぇ", "suffixOut": "尊い", "conditionsIn": [], "conditionsOut": ["adj-i"]},
{"suffixIn": "おなしゃす", "suffixOut": "おねがいします", "conditionsIn": [], "conditionsOut": ["v5"]},
{"suffixIn": "おなしゃす", "suffixOut": "お願いします", "conditionsIn": [], "conditionsOut": ["v5"]},
{"suffixIn": "あざす", "suffixOut": "ありがとうございます", "conditionsIn": [], "conditionsOut": ["v5"]},
{"suffixIn": "さーせん", "suffixOut": "すみません", "conditionsIn": [], "conditionsOut": []},
{"suffixIn": "神ってる", "suffixOut": "神がかっている", "conditionsIn": [], "conditionsOut": ["v1p"]},
{"suffixIn": "じわる", "suffixOut": "じわじわ来る", "conditionsIn": [], "conditionsOut": ["vk"]},
{"suffixIn": "おさしみ", "suffixOut": "おやすみ", "conditionsIn": [], "conditionsOut": []},
{"suffixIn": "おやさい", "suffixOut": "おやすみ", "conditionsIn": [], "conditionsOut": []}
]
},
{
"name": "ksb",
"description": "Negative form of kansai-ben verbs",
"rules": [
{"suffixIn": "へん", "suffixOut": "ない", "conditionsIn": [], "conditionsOut": ["adj-i"]},
{"suffixIn": "ひん", "suffixOut": "ない", "conditionsIn": [], "conditionsOut": ["adj-i"]},
{"suffixIn": "せえへん", "suffixOut": "しない", "conditionsIn": [], "conditionsOut": ["adj-i"]},
{"suffixIn": "へんかった", "suffixOut": "なかった", "conditionsIn": [], "conditionsOut": []},
{"suffixIn": "ひんかった", "suffixOut": "なかった", "conditionsIn": [], "conditionsOut": []},
{"suffixIn": "うてへん", "suffixOut": "ってない", "conditionsIn": [], "conditionsOut": ["adj-i"]}
]
},
{
"name": "ksb",
"description": "-ku stem of kansai-ben adjectives",
"rules": [
{"suffixIn": "う", "suffixOut": "く", "conditionsIn": [], "conditionsOut": []},
{"suffixIn": "こう", "suffixOut": "かく", "conditionsIn": [], "conditionsOut": []},
{"suffixIn": "ごう", "suffixOut": "がく", "conditionsIn": [], "conditionsOut": []},
{"suffixIn": "そう", "suffixOut": "さく", "conditionsIn": [], "conditionsOut": []},
{"suffixIn": "とう", "suffixOut": "たく", "conditionsIn": [], "conditionsOut": []},
{"suffixIn": "のう", "suffixOut": "なく", "conditionsIn": [], "conditionsOut": []},
{"suffixIn": "ぼう", "suffixOut": "ばく", "conditionsIn": [], "conditionsOut": []},
{"suffixIn": "もう", "suffixOut": "まく", "conditionsIn": [], "conditionsOut": []},
{"suffixIn": "ろう", "suffixOut": "らく", "conditionsIn": [], "conditionsOut": []},
{"suffixIn": "よう", "suffixOut": "よく", "conditionsIn": [], "conditionsOut": []},
{"suffixIn": "しゅう", "suffixOut": "しく", "conditionsIn": [], "conditionsOut": []}
]
},
{
"name": "ksb",
"description": "-te form of kansai-ben adjectives",
"rules": [
{"suffixIn": "うて", "suffixOut": "くて", "conditionsIn": [], "conditionsOut": []},
{"suffixIn": "こうて", "suffixOut": "かくて", "conditionsIn": [], "conditionsOut": []},
{"suffixIn": "ごうて", "suffixOut": "がくて", "conditionsIn": [], "conditionsOut": []},
{"suffixIn": "そうて", "suffixOut": "さくて", "conditionsIn": [], "conditionsOut": []},
{"suffixIn": "とうて", "suffixOut": "たくて", "conditionsIn": [], "conditionsOut": []},
{"suffixIn": "のうて", "suffixOut": "なくて", "conditionsIn": [], "conditionsOut": []},
{"suffixIn": "ぼうて", "suffixOut": "ばくて", "conditionsIn": [], "conditionsOut": []},
{"suffixIn": "もうて", "suffixOut": "まくて", "conditionsIn": [], "conditionsOut": []},
{"suffixIn": "ろうて", "suffixOut": "らくて", "conditionsIn": [], "conditionsOut": []},
{"suffixIn": "ようて", "suffixOut": "よくて", "conditionsIn": [], "conditionsOut": []},
{"suffixIn": "しゅうて", "suffixOut": "しくて", "conditionsIn": [], "conditionsOut": []}
]
},
{
"name": "ksb",
"description": "Negative form of kansai-ben adjectives",
"rules": [
{"suffixIn": "うない", "suffixOut": "くない", "conditionsIn": ["adj-i"], "conditionsOut": ["adj-i"]},
{"suffixIn": "こうない", "suffixOut": "かくない", "conditionsIn": ["adj-i"], "conditionsOut": ["adj-i"]},
{"suffixIn": "ごうない", "suffixOut": "がくない", "conditionsIn": ["adj-i"], "conditionsOut": ["adj-i"]},
{"suffixIn": "そうない", "suffixOut": "さくない", "conditionsIn": ["adj-i"], "conditionsOut": ["adj-i"]},
{"suffixIn": "とうない", "suffixOut": "たくない", "conditionsIn": ["adj-i"], "conditionsOut": ["adj-i"]},
{"suffixIn": "のうない", "suffixOut": "なくない", "conditionsIn": ["adj-i"], "conditionsOut": ["adj-i"]},
{"suffixIn": "ぼうない", "suffixOut": "ばくない", "conditionsIn": ["adj-i"], "conditionsOut": ["adj-i"]},
{"suffixIn": "もうない", "suffixOut": "まくない", "conditionsIn": ["adj-i"], "conditionsOut": ["adj-i"]},
{"suffixIn": "ろうない", "suffixOut": "らくない", "conditionsIn": ["adj-i"], "conditionsOut": ["adj-i"]},
{"suffixIn": "ようない", "suffixOut": "よくない", "conditionsIn": ["adj-i"], "conditionsOut": ["adj-i"]},
{"suffixIn": "しゅうない", "suffixOut": "しくない", "conditionsIn": ["adj-i"], "conditionsOut": ["adj-i"]}
]
},
{
"name": "ksb",
"description": "-te form of kansai-ben verbs",
Casheeew marked this conversation as resolved.
Show resolved Hide resolved
"rules": [
{"suffixIn": "うて", "suffixOut": "って", "conditionsIn": [], "conditionsOut": []},
{"suffixIn": "おうて", "suffixOut": "あって", "conditionsIn": [], "conditionsOut": []},
{"suffixIn": "こうて", "suffixOut": "かって", "conditionsIn": [], "conditionsOut": []},
{"suffixIn": "ごうて", "suffixOut": "がって", "conditionsIn": [], "conditionsOut": []},
{"suffixIn": "そうて", "suffixOut": "さって", "conditionsIn": [], "conditionsOut": []},
{"suffixIn": "ぞうて", "suffixOut": "ざって", "conditionsIn": [], "conditionsOut": []},
{"suffixIn": "とうて", "suffixOut": "たって", "conditionsIn": [], "conditionsOut": []},
{"suffixIn": "どうて", "suffixOut": "だって", "conditionsIn": [], "conditionsOut": []},
{"suffixIn": "のうて", "suffixOut": "なって", "conditionsIn": [], "conditionsOut": []},
{"suffixIn": "ほうて", "suffixOut": "はって", "conditionsIn": [], "conditionsOut": []},
{"suffixIn": "ぼうて", "suffixOut": "ばって", "conditionsIn": [], "conditionsOut": []},
{"suffixIn": "もうて", "suffixOut": "まって", "conditionsIn": [], "conditionsOut": []},
{"suffixIn": "ろうて", "suffixOut": "らって", "conditionsIn": [], "conditionsOut": []},
{"suffixIn": "ようて", "suffixOut": "やって", "conditionsIn": [], "conditionsOut": []},
{"suffixIn": "ゆうて", "suffixOut": "いって", "conditionsIn": [], "conditionsOut": []}
]
},
{
"name": "ksb",
"description": "past form of kansai-ben terms",
"rules": [
{"suffixIn": "うた", "suffixOut": "った", "conditionsIn": [], "conditionsOut": []},
{"suffixIn": "おうた", "suffixOut": "あった", "conditionsIn": [], "conditionsOut": []},
{"suffixIn": "こうた", "suffixOut": "かった", "conditionsIn": [], "conditionsOut": []},
{"suffixIn": "ごうた", "suffixOut": "がった", "conditionsIn": [], "conditionsOut": []},
{"suffixIn": "そうた", "suffixOut": "さった", "conditionsIn": [], "conditionsOut": []},
{"suffixIn": "ぞうた", "suffixOut": "ざった", "conditionsIn": [], "conditionsOut": []},
{"suffixIn": "とうた", "suffixOut": "たった", "conditionsIn": [], "conditionsOut": []},
{"suffixIn": "どうた", "suffixOut": "だった", "conditionsIn": [], "conditionsOut": []},
{"suffixIn": "のうた", "suffixOut": "なった", "conditionsIn": [], "conditionsOut": []},
{"suffixIn": "ほうた", "suffixOut": "はった", "conditionsIn": [], "conditionsOut": []},
{"suffixIn": "ぼうた", "suffixOut": "ばった", "conditionsIn": [], "conditionsOut": []},
{"suffixIn": "もうた", "suffixOut": "まった", "conditionsIn": [], "conditionsOut": []},
{"suffixIn": "ろうた", "suffixOut": "らった", "conditionsIn": [], "conditionsOut": []},
{"suffixIn": "ようた", "suffixOut": "やった", "conditionsIn": [], "conditionsOut": []},
{"suffixIn": "ゆうた", "suffixOut": "いった", "conditionsIn": [], "conditionsOut": []}
]
}
]
}
}

Loading
Loading