fix(rule): 〜かもしれないをより厳密にチェックするように

fix #3
textlint-ja · Sep 6, 2017 · 019b870 · 019b870
1 parent 6428752
commit 019b870
Show file tree

Hide file tree

Showing 5 changed files with 2,901 additions and 25 deletions.
diff --git a/package.json b/package.json
@@ -38,6 +38,7 @@
   },
   "dependencies": {
     "kuromojin": "^1.3.1",
-    "morpheme-match": "^1.0.1"
+    "morpheme-match": "^1.0.1",
+    "morpheme-match-all": "^1.1.0"
   }
 }
diff --git a/src/dict.js b/src/dict.js
@@ -1,6 +1,7 @@
 module.exports = [
     {
         "message": `弱い表現: "かも" が使われています。`,
+        // https://azu.github.io/morpheme-match/?text=問題がある(かも。)
         "tokens": [
             {
                 "surface_form": "かも",
@@ -13,9 +14,52 @@ module.exports = [
                 "basic_form": "かも",
                 "reading": "カモ",
                 "pronunciation": "カモ"
+            },
+            {
+                "surface_form": "。",
+                "pos": "記号",
+                "pos_detail_1": "句点",
+                "pos_detail_2": "*",
+                "pos_detail_3": "*",
+                "conjugated_type": "*",
+                "conjugated_form": "*",
+                "basic_form": "。",
+                "reading": "。",
+                "pronunciation": "。"
             }
         ]
     },
+    {
+        "message": `弱い表現: "かも" が使われています。`,
+        // https://azu.github.io/morpheme-match/?text=私は弱い(かもしれ)ない
+        "tokens": [
+            {
+                "surface_form": "かも",
+                "pos": "助詞",
+                "pos_detail_1": "副助詞",
+                "pos_detail_2": "*",
+                "pos_detail_3": "*",
+                "conjugated_type": "*",
+                "conjugated_form": "*",
+                "basic_form": "かも",
+                "reading": "カモ",
+                "pronunciation": "カモ"
+            },
+            {
+                "surface_form": "しれ",
+                "pos": "動詞",
+                "pos_detail_1": "自立",
+                "pos_detail_2": "*",
+                "pos_detail_3": "*",
+                "conjugated_type": "一段",
+                "conjugated_form": ["連用形", "未然形"],
+                "basic_form": "しれる",
+                "reading": "シレ",
+                "pronunciation": "シレ"
+            }
+        ]
+
+    },
     {
         "message": `弱い表現: "思う" が使われています。`,
         "tokens": [

diff --git a/src/textlint-rule-ja-no-weak-phrase.js b/src/textlint-rule-ja-no-weak-phrase.js
@@ -1,31 +1,29 @@
 // LICENSE : MIT
 "use strict";
 const tokenize = require("kuromojin").tokenize;
-import dictionaryList from "./dict";
-const createTokenMatcher = require("morpheme-match");
-module.exports = function (context) {
-    const {Syntax, RuleError, report, getSource} = context;
-    const matcherList = dictionaryList.map(dict => {
-        return {
-            matcher: createTokenMatcher(dict["tokens"]),
-            message: dict["message"]
-        };
-    });
+const dictionaryList = require("./dict");
+const createTokenMatcher = require("morpheme-match-all");
+const matchAll = createTokenMatcher(dictionaryList);
+module.exports = function(context) {
+    const { Syntax, RuleError, report, getSource } = context;
     return {
-        [Syntax.Str](node){
+        [Syntax.Str](node) {
             const text = getSource(node);
             return tokenize(text).then(currentTokens => {
-                currentTokens.forEach(token => {
-                    matcherList.forEach(({matcher, message}) => {
-                        const {match, tokens} = matcher(token);
-                        if (!match) {
-                            return;
-                        }
-                        const firstToken = tokens[0];
-                        report(node, new RuleError(message, {
-                            index: firstToken.word_position - 1
-                        }));
-                    });
+                /**
+                 * @type {MatchResult[]}
+                 */
+                const matchResults = matchAll(currentTokens);
+                matchResults.forEach(matchResult => {
+                    const firstToken = matchResult.tokens[0];
+                    const lastToken = matchResult.tokens[matchResult.tokens.length - 1];
+                    const firstWordIndex = Math.max(firstToken.word_position - 1, 0);
+                    const lastWorkIndex = Math.max(lastToken.word_position - 1, 0);
+                    // replace $1
+                    const message = matchResult.dict.message;
+                    report(node, new RuleError(message, {
+                        index: firstWordIndex
+                    }));
                 });
             });
         }

diff --git a/test/textlint-rule-ja-no-weak-phrase-test.js b/test/textlint-rule-ja-no-weak-phrase-test.js
@@ -12,12 +12,22 @@ tester.run("textlint-rule-ja-no-weak-phrase", rule, {
     invalid: [
         // single match
         {
-            text: "弱いかもしれない",
+            text: "問題があるかも。",
             errors: [
                 {
                     message: `弱い表現: "かも" が使われています。`,
                     line: 1,
-                    column: 3
+                    column: 6
+                }
+            ]
+        },
+        {
+            text: "私は弱いかもしれない。",
+            errors: [
+                {
+                    message: `弱い表現: "かも" が使われています。`,
+                    line: 1,
+                    column: 5
                 }
             ]
         },