Extend support for irregular inflections of suru-verbs

CHANGELOG.md

@@ -16,6 +16,10 @@ app.
 - Added handling for 戶 and 內 kyūjitai.
 - Added deinflection for additional forms of -sugiru
   ([#2033](https://github.com/birchill/10ten-ja-reader/pull/2033)).
+- Added deinflection for irregular forms of `vs-s` and `vz` class suru verbs
+  ([#2038](https://github.com/birchill/10ten-ja-reader/pull/2038)).
+- Removed deinflection of `vs-c` class verbs, as they are not used in modern
+  Japanese.
 - Fixed unreasonable matches caused by duplicates in the deinflection reason chain
   ([#1966](https://github.com/birchill/10ten-ja-reader/issues/1966)).
 - Fixed sorting of deinflected results in some cases (e.g. 見とれる).

README.md

@@ -53,7 +53,7 @@
   <img src="https://raw.githubusercontent.com/birchill/10ten-ja-reader/main/docs/jou-conversion-cropped.png" alt="Screenshot showing translation of 四畳半 into 7.29 square meters" title="Area translation" width="640">
 
 - Recognition of a wide range of grammatical forms
-  (e.g. vs-c verbs like 兼した,
+  (e.g. irregular inflections of する-verbs like 罰せられる,
   irregular verbs like いらっしゃいます,
   continuous forms like 食べてた,
   ん as a negative form like 分からん、知らん,

_locales/en/messages.json

@@ -371,6 +371,9 @@
   "deinflect_imperative_negative": {
     "message": "imperative negative"
   },
+  "deinflect_irregular": {
+    "message": "irregular"
+  },
   "deinflect_ki": {
     "message": "-ki"
   },

_locales/ja/messages.json

@@ -371,6 +371,9 @@
   "deinflect_imperative_negative": {
     "message": "否定命令形"
   },
+  "deinflect_irregular": {
+    "message": "不規則"
+  },
   "deinflect_ki": {
     "message": "連体形"
   },

_locales/zh_hans/messages.json

@@ -371,6 +371,9 @@
   "deinflect_imperative_negative": {
     "message": "否定命令形"
   },
+  "deinflect_irregular": {
+    "message": "不规则"
+  },
   "deinflect_ki": {
     "message": "连体形"
   },

src/background/deinflect.test.ts

@@ -128,14 +128,62 @@ describe('deinflect', () => {
     });
   });
 
-  it('deinflects vs-c', () => {
-    const result = deinflect('兼した');
-    const match = result.find((candidate) => candidate.word === '兼す');
-    expect(match).toEqual({
-      reasonChains: [[Reason.Past]],
-      type: 18,
-      word: '兼す',
-    });
+  it('deinflects all forms of する', () => {
+    const cases = [
+      ['した', [Reason.Past]],
+      ['しよう', [Reason.Volitional]],
+      ['しない', [Reason.Negative]],
+      ['せぬ', [Reason.Negative]],
+      ['せん', [Reason.Negative]],
+      ['せず', [Reason.Zu]],
+      ['される', [Reason.Passive]],
+      ['させる', [Reason.Causative]],
+      ['しろ', [Reason.Imperative]],
+      ['せよ', [Reason.Imperative]],
+      ['すれば', [Reason.Ba]],
+      ['できる', [Reason.Potential]],
+    ];
+
+    for (const [inflected, reasons] of cases) {
+      const result = deinflect(inflected as string);
+      const match = result.find(
+        (candidate) =>
+          candidate.word == 'する' && candidate.type & WordType.SuruVerb
+      );
+      expect(match).toBeDefined();
+      expect(match!.reasonChains).toEqual([reasons]);
+    }
+  });
+
+  it('deinflects additional forms of special class suru-verbs', () => {
+    const cases = [
+      ['発する', '発せさせる', [Reason.Irregular, Reason.Causative]],
+      ['発する', '発せられる', [Reason.Irregular, Reason.PotentialOrPassive]],
+      ['発する', '発しさせる', [Reason.Irregular, Reason.Causative]],
+      ['発する', '発しられる', [Reason.Irregular, Reason.PotentialOrPassive]],
+      // 五段化
+      ['発する', '発さない', [Reason.Irregular, Reason.Negative]],
+      ['発する', '発さず', [Reason.Irregular, Reason.Zu]],
+      ['発する', '発そう', [Reason.Irregular, Reason.Volitional]],
+      ['愛する', '愛せる', [Reason.Irregular, Reason.Potential]],
+      ['愛する', '愛せば', [Reason.Irregular, Reason.Ba]],
+      ['愛する', '愛せ', [Reason.Irregular, Reason.Imperative]],
+      // ずる / vz class verbs
+      ['信ずる', '信ぜぬ', [Reason.Irregular, Reason.Negative]],
+      ['信ずる', '信ぜず', [Reason.Irregular, Reason.Zu]],
+      ['信ずる', '信ぜさせる', [Reason.Irregular, Reason.Causative]],
+      ['信ずる', '信ぜられる', [Reason.Irregular, Reason.PotentialOrPassive]],
+      ['信ずる', '信ずれば', [Reason.Irregular, Reason.Ba]],
+      ['信ずる', '信ぜよ', [Reason.Irregular, Reason.Imperative]],
+    ];
+
+    for (const [plain, inflected, reasons] of cases) {
+      const result = deinflect(inflected as string);
+      const match = result.find((candidate) => candidate.word == plain);
+      expect(match).toBeDefined();
+      expect(match!.type).toEqual(WordType.SpecialSuruVerb);
+      expect(match!.reasonChains).toEqual([reasons]);
+    }
   });
 
   it('deinflects irregular forms of 行く', () => {
@@ -310,8 +358,8 @@ describe('deinflect', () => {
       ['歩いてる', '歩く', 2, undefined],
       ['泳いでいる', '泳ぐ', 2, undefined],
       ['泳いでる', '泳ぐ', 2, undefined],
-      ['話している', '話す', 18, undefined],
-      ['話してる', '話す', 18, undefined],
+      ['話している', '話す', 2, undefined],
+      ['話してる', '話す', 2, undefined],
       ['死んでいる', '死ぬ', 2, undefined],
       ['死んでる', '死ぬ', 2, undefined],
       ['飼っている', '飼う', 2, undefined],

src/background/deinflect.ts

@@ -37,6 +37,7 @@ export const enum Reason {
   SuruNoun,
   ZaruWoEnai,
   NegativeTe,
+  Irregular,
 }
 
 export const deinflectL10NKeys: { [key: number]: string } = {
@@ -76,6 +77,7 @@ export const deinflectL10NKeys: { [key: number]: string } = {
   [Reason.SuruNoun]: 'deinflect_suru_noun',
   [Reason.ZaruWoEnai]: 'deinflect_zaru_wo_enai',
   [Reason.NegativeTe]: 'deinflect_negative_te',
+  [Reason.Irregular]: 'deinflect_irregular',
 };
 
 const enum Type {
@@ -85,14 +87,21 @@ const enum Type {
   IAdj = 1 << 2,
   KuruVerb = 1 << 3,
   SuruVerb = 1 << 4,
-  NounVS = 1 << 5,
-  All = IchidanVerb | GodanVerb | IAdj | KuruVerb | SuruVerb | NounVS,
+  SpecialSuruVerb = 1 << 5,
+  NounVS = 1 << 6,
+  All = IchidanVerb |
+    GodanVerb |
+    IAdj |
+    KuruVerb |
+    SuruVerb |
+    SpecialSuruVerb |
+    NounVS,
   // Intermediate types
-  Initial = 1 << 6, // original word before any deinflection (from-type only)
-  TaTeStem = 1 << 7,
-  DaDeStem = 1 << 8,
-  MasuStem = 1 << 9,
-  IrrealisStem = 1 << 10,
+  Initial = 1 << 7, // original word before any deinflection (from-type only)
+  TaTeStem = 1 << 8,
+  DaDeStem = 1 << 9,
+  MasuStem = 1 << 10,
+  IrrealisStem = 1 << 11,
 }
 
 export { Type as WordType };
@@ -169,6 +178,12 @@ const deinflectRuleData: Array<
   ['ざるえぬ', '', Type.IAdj, Type.IrrealisStem, [Reason.ZaruWoEnai]],
   ['ざる得ぬ', '', Type.IAdj, Type.IrrealisStem, [Reason.ZaruWoEnai]],
   ['しないで', 'する', Type.Initial, Type.SuruVerb, [Reason.NegativeTe]],
+  ['しさせる', 'する', Type.IchidanVerb, Type.SpecialSuruVerb, [Reason.Irregular, Reason.Causative]],
+  ['しられる', 'する', Type.IchidanVerb, Type.SpecialSuruVerb, [Reason.Irregular, Reason.PotentialOrPassive]],
+  ['せさせる', 'する', Type.IchidanVerb, Type.SpecialSuruVerb, [Reason.Irregular, Reason.Causative]],
+  ['せられる', 'する', Type.IchidanVerb, Type.SpecialSuruVerb, [Reason.Irregular, Reason.PotentialOrPassive]],
+  ['ぜさせる', 'ずる', Type.IchidanVerb, Type.SpecialSuruVerb, [Reason.Irregular, Reason.Causative]],
+  ['ぜられる', 'ずる', Type.IchidanVerb, Type.SpecialSuruVerb, [Reason.Irregular, Reason.PotentialOrPassive]],
   ['たゆたう', 'たゆたう', Type.TaTeStem, Type.GodanVerb, []],
   ['たゆとう', 'たゆとう', Type.TaTeStem, Type.GodanVerb, []],
   ['のたまう', 'のたまう', Type.TaTeStem, Type.GodanVerb, []],
@@ -193,6 +208,7 @@ const deinflectRuleData: Array<
   ['御座い', '御座る', Type.MasuStem, Type.GodanVerb, [Reason.MasuStem]],
   ['させる', 'る', Type.IchidanVerb, Type.IchidanVerb | Type.KuruVerb, [Reason.Causative]],
   ['させる', 'する', Type.IchidanVerb, Type.SuruVerb, [Reason.Causative]],
+  ['さない', 'する', Type.IAdj, Type.SpecialSuruVerb, [Reason.Irregular, Reason.Negative]],
   ['される', '', Type.IchidanVerb, Type.IrrealisStem, [Reason.CausativePassive]],
   ['される', 'する', Type.IchidanVerb, Type.SuruVerb, [Reason.Passive]],
   ['しない', 'する', Type.IAdj, Type.SuruVerb, [Reason.Negative]],
@@ -202,6 +218,7 @@ const deinflectRuleData: Array<
   ['すぎる', '', Type.IchidanVerb, Type.MasuStem, [Reason.Sugiru]],
   ['過ぎる', 'い', Type.IchidanVerb, Type.IAdj, [Reason.Sugiru]],
   ['過ぎる', '', Type.IchidanVerb, Type.MasuStem, [Reason.Sugiru]],
+  ['ずれば', 'ずる', Type.Initial, Type.SpecialSuruVerb, [Reason.Irregular, Reason.Ba]],
   ['たまう', 'たまう', Type.TaTeStem, Type.GodanVerb, []],
   ['たもう', 'たもう', Type.TaTeStem, Type.GodanVerb, []],
   ['揺蕩う', '揺蕩う', Type.TaTeStem, Type.GodanVerb, []],
@@ -239,8 +256,8 @@ const deinflectRuleData: Array<
   ['こい', 'くる', Type.Initial, Type.KuruVerb, [Reason.Imperative]],
   ['こう', 'く', Type.Initial, Type.GodanVerb, [Reason.Volitional]],
   ['ごう', 'ぐ', Type.Initial, Type.GodanVerb, [Reason.Volitional]],
-  ['しろ', 'す', Type.Initial, Type.SuruVerb, [Reason.Imperative]],
   ['しろ', 'する', Type.Initial, Type.SuruVerb, [Reason.Imperative]],
+  ['さず', 'する', Type.Initial, Type.SpecialSuruVerb, [Reason.Irregular, Reason.Zu]],
   ['すぎ', 'い', Type.Initial, Type.IAdj, [Reason.Sugiru]],
   ['すぎ', '', Type.Initial, Type.MasuStem, [Reason.Sugiru]],
   ['過ぎ', 'い', Type.Initial, Type.IAdj, [Reason.Sugiru]],
@@ -249,17 +266,19 @@ const deinflectRuleData: Array<
   ['せず', 'する', Type.Initial, Type.SuruVerb, [Reason.Zu]],
   ['せぬ', 'する', Type.Initial, Type.SuruVerb, [Reason.Negative]],
   ['せん', 'する', Type.Initial, Type.SuruVerb, [Reason.Negative]],
-  ['せず', 'す', Type.Initial, Type.SuruVerb, [Reason.Zu]],
-  ['せぬ', 'す', Type.Initial, Type.SuruVerb, [Reason.Negative]],
-  ['せん', 'す', Type.Initial, Type.SuruVerb, [Reason.Negative]],
-  ['せば', 'す', Type.Initial, Type.GodanVerb | Type.SuruVerb, [Reason.Ba]],
+  ['せば', 'す', Type.Initial, Type.GodanVerb, [Reason.Ba]],
+  ['せば', 'する', Type.Initial, Type.SpecialSuruVerb, [Reason.Irregular, Reason.Ba]],
+  ['せる', 'する', Type.IchidanVerb, Type.SpecialSuruVerb, [Reason.Irregular, Reason.Potential]],
   ['せよ', 'する', Type.Initial, Type.SuruVerb, [Reason.Imperative]],
-  ['せよ', 'す', Type.Initial, Type.SuruVerb, [Reason.Imperative]],
   ['せる', 'す', Type.IchidanVerb, Type.GodanVerb, [Reason.Potential]],
   ['せる', '', Type.IchidanVerb, Type.IrrealisStem, [Reason.Causative]],
+  ['ぜず', 'ずる', Type.Initial, Type.SpecialSuruVerb, [Reason.Irregular, Reason.Zu]],
+  ['ぜぬ', 'ずる', Type.Initial, Type.SpecialSuruVerb, [Reason.Irregular, Reason.Negative]],
+  ['ぜよ', 'ずる', Type.Initial, Type.SpecialSuruVerb, [Reason.Irregular, Reason.Imperative]],
   ['そう', '', Type.Initial, Type.MasuStem, [Reason.Sou]],
   ['そう', 'い', Type.Initial, Type.IAdj, [Reason.Sou]],
   ['そう', 'す', Type.Initial, Type.GodanVerb, [Reason.Volitional]],
+  ['そう', 'する', Type.Initial, Type.SpecialSuruVerb, [Reason.Irregular, Reason.Volitional]],
   ['たい', '', Type.IAdj, Type.MasuStem, [Reason.Tai]],
   ['たら', '', Type.Initial, Type.TaTeStem, [Reason.Tara]],
   ['だら', '', Type.Initial, Type.DaDeStem, [Reason.Tara]],
@@ -334,10 +353,11 @@ const deinflectRuleData: Array<
   ['さ', 'す', Type.IrrealisStem, Type.GodanVerb, []],
   ['し', 'す', Type.MasuStem, Type.GodanVerb, [Reason.MasuStem]],
   ['し', 'する', Type.MasuStem, Type.SuruVerb, [Reason.MasuStem]],
-  ['し', 'す', Type.TaTeStem, Type.GodanVerb | Type.SuruVerb, []],
+  ['し', 'す', Type.TaTeStem, Type.GodanVerb, []],
   ['し', 'する', Type.TaTeStem, Type.SuruVerb, []],
   ['ず', '', Type.Initial, Type.IrrealisStem, [Reason.Zu]],
   ['せ', 'す', Type.Initial, Type.GodanVerb, [Reason.Imperative]],
+  ['せ', 'する', Type.Initial, Type.SpecialSuruVerb, [Reason.Irregular, Reason.Imperative]],
   ['た', 'つ', Type.IrrealisStem, Type.GodanVerb, []],
   ['た', '', Type.Initial, Type.TaTeStem, [Reason.Past]],
   ['だ', '', Type.Initial, Type.DaDeStem, [Reason.Past]],
@@ -514,9 +534,21 @@ export function deinflect(word: string): CandidateWord[] {
           continue;
         }
 
-        // Continue if the rule introduces a duplicate in the reason chain,
-        // as it wouldn't make sense grammatically.
+        // Verify that adding the rule won't lead to duplicates or an unreasonable
+        // sequencing of reasons in the reason chain.
         const ruleReasons = new Set(rule.reasons);
+
+        // Avoid matches such as 'potential < potential or passive' or
+        // 'potential < causative', which are grammatically incorrect.
+        // This is an issue when handling irregular forms like 罰せられる,
+        // where the structure could be misinterpreted as
+        // 'irregular < potential < potential or passive' instead of the correct
+        // 'irregular < potential or passive'.
+        if (ruleReasons.has(Reason.Potential)) {
+          ruleReasons.add(Reason.PotentialOrPassive);
+          ruleReasons.add(Reason.Causative);
+        }
+
         if (thisCandidate.reasonChains.flat().some((r) => ruleReasons.has(r))) {
           continue;
         }

src/background/word-search.ts

@@ -286,7 +286,14 @@ function entryMatchesType(entry: DictionaryWordResult, type: number): boolean {
 
   if (
     type & WordType.SuruVerb &&
-    hasMatchingSense((pos) => pos.startsWith('vs-'))
+    hasMatchingSense((pos) => pos === 'vs-i' || pos === 'vs-s')
+  ) {
+    return true;
+  }
+
+  if (
+    type & WordType.SpecialSuruVerb &&
+    hasMatchingSense((pos) => pos === 'vs-s' || pos === 'vz')
   ) {
     return true;
   }